Class: Google::Cloud::AIPlatform::V1::StudySpec

Inherits:

Object

• Object
• Google::Cloud::AIPlatform::V1::StudySpec

Extended by:

Protobuf::MessageExts::ClassMethods

Includes:

Protobuf::MessageExts

Defined in:

proto_docs/google/cloud/aiplatform/v1/study.rb

Overview

Represents specification of a Study.

Defined Under Namespace

Modules: Algorithm, MeasurementSelectionType, ObservationNoise Classes: ConvexAutomatedStoppingSpec, DecayCurveAutomatedStoppingSpec, MedianAutomatedStoppingSpec, MetricSpec, ParameterSpec, StudyStoppingConfig

Instance Attribute Summary

• #algorithm ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::Algorithm

  The search algorithm specified for the Study.

• #convex_automated_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::ConvexAutomatedStoppingSpec

  The automated early stopping spec using convex stopping rule.

• #decay_curve_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::DecayCurveAutomatedStoppingSpec

  The automated early stopping spec using decay curve rule.

• #measurement_selection_type ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::MeasurementSelectionType

  Describe which measurement selection type will be used.

• #median_automated_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::MedianAutomatedStoppingSpec

  The automated early stopping spec using median rule.

• #metrics ⇒ ::Array<::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec>

  Required.

• #observation_noise ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::ObservationNoise

  The observation noise level of the study.

• #parameters ⇒ ::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec>

  Required.

• #study_stopping_config ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::StudyStoppingConfig

  Conditions for automated stopping of a Study.

Instance Attribute Details

#algorithm ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::Algorithm

Returns The search algorithm specified for the Study.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::Algorithm) —

  The search algorithm specified for the Study.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#convex_automated_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::ConvexAutomatedStoppingSpec

Returns The automated early stopping spec using convex stopping rule.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::ConvexAutomatedStoppingSpec) —

  The automated early stopping spec using convex stopping rule.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#decay_curve_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::DecayCurveAutomatedStoppingSpec

Returns The automated early stopping spec using decay curve rule.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::DecayCurveAutomatedStoppingSpec) —

  The automated early stopping spec using decay curve rule.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#measurement_selection_type ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::MeasurementSelectionType

Returns Describe which measurement selection type will be used.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::MeasurementSelectionType) —

  Describe which measurement selection type will be used

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#median_automated_stopping_spec ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::MedianAutomatedStoppingSpec

Returns The automated early stopping spec using median rule.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::MedianAutomatedStoppingSpec) —

  The automated early stopping spec using median rule.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#metrics ⇒ ::Array<::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec>

Returns Required. Metric specs for the Study.

Returns:

• (::Array<::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec>) —

  Required. Metric specs for the Study.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#observation_noise ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::ObservationNoise

Returns The observation noise level of the study. Currently only supported by the Vertex AI Vizier service. Not supported by HyperparameterTuningJob or TrainingPipeline.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::ObservationNoise) —

  The observation noise level of the study. Currently only supported by the Vertex AI Vizier service. Not supported by HyperparameterTuningJob or TrainingPipeline.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#parameters ⇒ ::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec>

Returns Required. The set of parameters to tune.

Returns:

• (::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec>) —

  Required. The set of parameters to tune.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end

#study_stopping_config ⇒ ::Google::Cloud::AIPlatform::V1::StudySpec::StudyStoppingConfig

Returns Conditions for automated stopping of a Study. Enable automated stopping by configuring at least one condition.

Returns:

• (::Google::Cloud::AIPlatform::V1::StudySpec::StudyStoppingConfig) —

  Conditions for automated stopping of a Study. Enable automated stopping by configuring at least one condition.

# File 'proto_docs/google/cloud/aiplatform/v1/study.rb', line 250

class StudySpec
  include ::Google::Protobuf::MessageExts
  extend ::Google::Protobuf::MessageExts::ClassMethods

  # Represents a metric to optimize.
  # @!attribute [rw] metric_id
  #   @return [::String]
  #     Required. The ID of the metric. Must not contain whitespaces and must be
  #     unique amongst all MetricSpecs.
  # @!attribute [rw] goal
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::GoalType]
  #     Required. The optimization goal of the metric.
  # @!attribute [rw] safety_config
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::MetricSpec::SafetyMetricConfig]
  #     Used for safe search. In the case, the metric will be a safety
  #     metric. You must provide a separate metric for objective metric.
  class MetricSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Used in safe optimization to specify threshold levels and risk tolerance.
    # @!attribute [rw] safety_threshold
    #   @return [::Float]
    #     Safety threshold (boundary value between safe and unsafe). NOTE that if
    #     you leave SafetyMetricConfig unset, a default value of 0 will be used.
    # @!attribute [rw] desired_min_safe_trials_fraction
    #   @return [::Float]
    #     Desired minimum fraction of safe trials (over total number of trials)
    #     that should be targeted by the algorithm at any time during the
    #     study (best effort). This should be between 0.0 and 1.0 and a value of
    #     0.0 means that there is no minimum and an algorithm proceeds without
    #     targeting any specific fraction. A value of 1.0 means that the
    #     algorithm attempts to only Suggest safe Trials.
    class SafetyMetricConfig
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # The available types of optimization goals.
    module GoalType
      # Goal Type will default to maximize.
      GOAL_TYPE_UNSPECIFIED = 0

      # Maximize the goal metric.
      MAXIMIZE = 1

      # Minimize the goal metric.
      MINIMIZE = 2
    end
  end

  # Represents a single parameter to optimize.
  # @!attribute [rw] double_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DoubleValueSpec]
  #     The value spec for a 'DOUBLE' parameter.
  # @!attribute [rw] integer_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::IntegerValueSpec]
  #     The value spec for an 'INTEGER' parameter.
  # @!attribute [rw] categorical_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::CategoricalValueSpec]
  #     The value spec for a 'CATEGORICAL' parameter.
  # @!attribute [rw] discrete_value_spec
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::DiscreteValueSpec]
  #     The value spec for a 'DISCRETE' parameter.
  # @!attribute [rw] parameter_id
  #   @return [::String]
  #     Required. The ID of the parameter. Must not contain whitespaces and must
  #     be unique amongst all ParameterSpecs.
  # @!attribute [rw] scale_type
  #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ScaleType]
  #     How the parameter should be scaled.
  #     Leave unset for `CATEGORICAL` parameters.
  # @!attribute [rw] conditional_parameter_specs
  #   @return [::Array<::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec>]
  #     A conditional parameter node is active if the parameter's value matches
  #     the conditional node's parent_value_condition.
  #
  #     If two items in conditional_parameter_specs have the same name, they
  #     must have disjoint parent_value_condition.
  class ParameterSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods

    # Value specification for a parameter in `DOUBLE` type.
    # @!attribute [rw] min_value
    #   @return [::Float]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Float]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DOUBLE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DoubleValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `INTEGER` type.
    # @!attribute [rw] min_value
    #   @return [::Integer]
    #     Required. Inclusive minimum value of the parameter.
    # @!attribute [rw] max_value
    #   @return [::Integer]
    #     Required. Inclusive maximum value of the parameter.
    # @!attribute [rw] default_value
    #   @return [::Integer]
    #     A default value for an `INTEGER` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class IntegerValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `CATEGORICAL` type.
    # @!attribute [rw] values
    #   @return [::Array<::String>]
    #     Required. The list of possible categories.
    # @!attribute [rw] default_value
    #   @return [::String]
    #     A default value for a `CATEGORICAL` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class CategoricalValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Value specification for a parameter in `DISCRETE` type.
    # @!attribute [rw] values
    #   @return [::Array<::Float>]
    #     Required. A list of possible values.
    #     The list should be in increasing order and at least 1e-10 apart.
    #     For instance, this parameter might have possible settings of 1.5, 2.5,
    #     and 4.0. This list should not contain more than 1,000 values.
    # @!attribute [rw] default_value
    #   @return [::Float]
    #     A default value for a `DISCRETE` parameter that is assumed to be a
    #     relatively good starting point.  Unset value signals that there is no
    #     offered starting point.  It automatically rounds to the
    #     nearest feasible discrete point.
    #
    #     Currently only supported by the Vertex AI Vizier service. Not supported
    #     by HyperparameterTuningJob or TrainingPipeline.
    class DiscreteValueSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods
    end

    # Represents a parameter spec with condition from its parent parameter.
    # @!attribute [rw] parent_discrete_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::DiscreteValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `DISCRETE` type.
    # @!attribute [rw] parent_int_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::IntValueCondition]
    #     The spec for matching values from a parent parameter of `INTEGER`
    #     type.
    # @!attribute [rw] parent_categorical_values
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec::ConditionalParameterSpec::CategoricalValueCondition]
    #     The spec for matching values from a parent parameter of
    #     `CATEGORICAL` type.
    # @!attribute [rw] parameter_spec
    #   @return [::Google::Cloud::AIPlatform::V1::StudySpec::ParameterSpec]
    #     Required. The spec for a conditional parameter.
    class ConditionalParameterSpec
      include ::Google::Protobuf::MessageExts
      extend ::Google::Protobuf::MessageExts::ClassMethods

      # Represents the spec to match discrete values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Float>]
      #     Required. Matches values of the parent parameter of 'DISCRETE' type.
      #     All values must exist in `discrete_value_spec` of parent parameter.
      #
      #     The Epsilon of the value matching is 1e-10.
      class DiscreteValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match integer values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::Integer>]
      #     Required. Matches values of the parent parameter of 'INTEGER' type.
      #     All values must lie in `integer_value_spec` of parent parameter.
      class IntValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end

      # Represents the spec to match categorical values from parent parameter.
      # @!attribute [rw] values
      #   @return [::Array<::String>]
      #     Required. Matches values of the parent parameter of 'CATEGORICAL'
      #     type. All values must exist in `categorical_value_spec` of parent
      #     parameter.
      class CategoricalValueCondition
        include ::Google::Protobuf::MessageExts
        extend ::Google::Protobuf::MessageExts::ClassMethods
      end
    end

    # The type of scaling that should be applied to this parameter.
    module ScaleType
      # By default, no scaling is applied.
      SCALE_TYPE_UNSPECIFIED = 0

      # Scales the feasible space to (0, 1) linearly.
      UNIT_LINEAR_SCALE = 1

      # Scales the feasible space logarithmically to (0, 1). The entire
      # feasible space must be strictly positive.
      UNIT_LOG_SCALE = 2

      # Scales the feasible space "reverse" logarithmically to (0, 1). The
      # result is that values close to the top of the feasible space are spread
      # out more than points near the bottom. The entire feasible space must be
      # strictly positive.
      UNIT_REVERSE_LOG_SCALE = 3
    end
  end

  # The decay curve automated stopping rule builds a Gaussian Process
  # Regressor to predict the final objective value of a Trial based on the
  # already completed Trials and the intermediate measurements of the current
  # Trial. Early stopping is requested for the current Trial if there is very
  # low probability to exceed the optimal value found so far.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}
  #     is used as the x-axis of each Trials Decay Curve. Otherwise,
  #     {::Google::Cloud::AIPlatform::V1::Measurement#step_count Measurement.step_count}
  #     will be used as the x-axis.
  class DecayCurveAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The median automated stopping rule stops a pending Trial if the Trial's
  # best objective_value is strictly below the median 'performance' of all
  # completed Trials reported up to the Trial's last measurement.
  # Currently, 'performance' refers to the running average of the objective
  # values reported by the Trial in each measurement.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     True if median automated stopping rule applies on
  #     {::Google::Cloud::AIPlatform::V1::Measurement#elapsed_duration Measurement.elapsed_duration}.
  #     It means that elapsed_duration field of latest measurement of current
  #     Trial is used to compute median objective value for each completed
  #     Trials.
  class MedianAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # Configuration for ConvexAutomatedStoppingSpec.
  # When there are enough completed trials (configured by
  # min_measurement_count), for pending trials with enough measurements and
  # steps, the policy first computes an overestimate of the objective value at
  # max_num_steps according to the slope of the incomplete objective value
  # curve. No prediction can be made if the curve is completely flat. If the
  # overestimation is worse than the best objective value of the completed
  # trials, this pending trial will be early-stopped, but a last measurement
  # will be added to the pending trial with max_num_steps and predicted
  # objective value from the autoregression model.
  # @!attribute [rw] max_step_count
  #   @return [::Integer]
  #     Steps used in predicting the final objective for early stopped trials. In
  #     general, it's set to be the same as the defined steps in training /
  #     tuning. If not defined, it will learn it from the completed trials. When
  #     use_steps is false, this field is set to the maximum elapsed seconds.
  # @!attribute [rw] min_step_count
  #   @return [::Integer]
  #     Minimum number of steps for a trial to complete. Trials which do not have
  #     a measurement with step_count > min_step_count won't be considered for
  #     early stopping. It's ok to set it to 0, and a trial can be early stopped
  #     at any stage. By default, min_step_count is set to be one-tenth of the
  #     max_step_count.
  #     When use_elapsed_duration is true, this field is set to the minimum
  #     elapsed seconds.
  # @!attribute [rw] min_measurement_count
  #   @return [::Integer]
  #     The minimal number of measurements in a Trial.  Early-stopping checks
  #     will not trigger if less than min_measurement_count+1 completed trials or
  #     pending trials with less than min_measurement_count measurements. If not
  #     defined, the default value is 5.
  # @!attribute [rw] learning_rate_parameter_name
  #   @return [::String]
  #     The hyper-parameter name used in the tuning job that stands for learning
  #     rate. Leave it blank if learning rate is not in a parameter in tuning.
  #     The learning_rate is used to estimate the objective value of the ongoing
  #     trial.
  # @!attribute [rw] use_elapsed_duration
  #   @return [::Boolean]
  #     This bool determines whether or not the rule is applied based on
  #     elapsed_secs or steps. If use_elapsed_duration==false, the early stopping
  #     decision is made according to the predicted objective values according to
  #     the target steps. If use_elapsed_duration==true, elapsed_secs is used
  #     instead of steps. Also, in this case, the parameters max_num_steps and
  #     min_num_steps are overloaded to contain max_elapsed_seconds and
  #     min_elapsed_seconds.
  # @!attribute [rw] update_all_stopped_trials
  #   @return [::Boolean]
  #     ConvexAutomatedStoppingSpec by default only updates the trials that needs
  #     to be early stopped using a newly trained auto-regressive model. When
  #     this flag is set to True, all stopped trials from the beginning are
  #     potentially updated in terms of their `final_measurement`. Also, note
  #     that the training logic of autoregressive models is different in this
  #     case. Enabling this option has shown better results and this may be the
  #     default option in the future.
  class ConvexAutomatedStoppingSpec
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The configuration (stopping conditions) for automated stopping of a Study.
  # Conditions include trial budgets, time budgets, and convergence detection.
  # @!attribute [rw] should_stop_asap
  #   @return [::Google::Protobuf::BoolValue]
  #     If true, a Study enters STOPPING_ASAP whenever it would normally enters
  #     STOPPING state.
  #
  #     The bottom line is: set to true if you want to interrupt on-going
  #     evaluations of Trials as soon as the study stopping condition is met.
  #     (Please see Study.State documentation for the source of truth).
  # @!attribute [rw] minimum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     Each "stopping rule" in this proto specifies an "if" condition. Before
  #     Vizier would generate a new suggestion, it first checks each specified
  #     stopping rule, from top to bottom in this list.
  #     Note that the first few rules (e.g. minimum_runtime_constraint,
  #     min_num_trials) will prevent other stopping rules from being evaluated
  #     until they are met. For example, setting `min_num_trials=5` and
  #     `always_stop_after= 1 hour` means that the Study will ONLY stop after it
  #     has 5 COMPLETED trials, even if more than an hour has passed since its
  #     creation. It follows the first applicable rule (whose "if" condition is
  #     satisfied) to make a stopping decision. If none of the specified rules
  #     are applicable, then Vizier decides that the study should not stop.
  #     If Vizier decides that the study should stop, the study enters
  #     STOPPING state (or STOPPING_ASAP if should_stop_asap = true).
  #     IMPORTANT: The automatic study state transition happens precisely as
  #     described above; that is, deleting trials or updating StudyConfig NEVER
  #     automatically moves the study state back to ACTIVE. If you want to
  #     _resume_ a Study that was stopped, 1) change the stopping conditions if
  #     necessary, 2) activate the study, and then 3) ask for suggestions.
  #     If the specified time or duration has not passed, do not stop the
  #     study.
  # @!attribute [rw] maximum_runtime_constraint
  #   @return [::Google::Cloud::AIPlatform::V1::StudyTimeConstraint]
  #     If the specified time or duration has passed, stop the study.
  # @!attribute [rw] min_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are fewer than this many COMPLETED trials, do not stop the
  #     study.
  # @!attribute [rw] max_num_trials
  #   @return [::Google::Protobuf::Int32Value]
  #     If there are more than this many trials, stop the study.
  # @!attribute [rw] max_num_trials_no_progress
  #   @return [::Google::Protobuf::Int32Value]
  #     If the objective value has not improved for this many consecutive
  #     trials, stop the study.
  #
  #     WARNING: Effective only for single-objective studies.
  # @!attribute [rw] max_duration_no_progress
  #   @return [::Google::Protobuf::Duration]
  #     If the objective value has not improved for this much time, stop the
  #     study.
  #
  #     WARNING: Effective only for single-objective studies.
  class StudyStoppingConfig
    include ::Google::Protobuf::MessageExts
    extend ::Google::Protobuf::MessageExts::ClassMethods
  end

  # The available search algorithms for the Study.
  module Algorithm
    # The default algorithm used by Vertex AI for [hyperparameter
    # tuning](https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview)
    # and [Vertex AI Vizier](https://cloud.google.com/vertex-ai/docs/vizier).
    ALGORITHM_UNSPECIFIED = 0

    # Simple grid search within the feasible space. To use grid search,
    # all parameters must be `INTEGER`, `CATEGORICAL`, or `DISCRETE`.
    GRID_SEARCH = 2

    # Simple random search within the feasible space.
    RANDOM_SEARCH = 3
  end

  # Describes the noise level of the repeated observations.
  #
  # "Noisy" means that the repeated observations with the same Trial parameters
  # may lead to different metric evaluations.
  module ObservationNoise
    # The default noise level chosen by Vertex AI.
    OBSERVATION_NOISE_UNSPECIFIED = 0

    # Vertex AI assumes that the objective function is (nearly)
    # perfectly reproducible, and will never repeat the same Trial
    # parameters.
    LOW = 1

    # Vertex AI will estimate the amount of noise in metric
    # evaluations, it may repeat the same Trial parameters more than once.
    HIGH = 2
  end

  # This indicates which measurement to use if/when the service automatically
  # selects the final measurement from previously reported intermediate
  # measurements. Choose this based on two considerations:
  #  A) Do you expect your measurements to monotonically improve?
  #     If so, choose LAST_MEASUREMENT. On the other hand, if you're in a
  #     situation where your system can "over-train" and you expect the
  #     performance to get better for a while but then start declining,
  #     choose BEST_MEASUREMENT.
  #  B) Are your measurements significantly noisy and/or irreproducible?
  #     If so, BEST_MEASUREMENT will tend to be over-optimistic, and it
  #     may be better to choose LAST_MEASUREMENT.
  #  If both or neither of (A) and (B) apply, it doesn't matter which
  #  selection type is chosen.
  module MeasurementSelectionType
    # Will be treated as LAST_MEASUREMENT.
    MEASUREMENT_SELECTION_TYPE_UNSPECIFIED = 0

    # Use the last measurement reported.
    LAST_MEASUREMENT = 1

    # Use the best measurement reported.
    BEST_MEASUREMENT = 2
  end
end