Class: Google::Apis::AiplatformV1::GoogleCloudAiplatformV1IntegratedGradientsAttribution

Inherits:
Object
  • Object
show all
Includes:
Core::Hashable, Core::JsonObjectSupport
Defined in:
lib/google/apis/aiplatform_v1/classes.rb,
lib/google/apis/aiplatform_v1/representations.rb,
lib/google/apis/aiplatform_v1/representations.rb

Overview

An attribution method that computes the Aumann-Shapley value taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1703.01365

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(**args) ⇒ GoogleCloudAiplatformV1IntegratedGradientsAttribution

Returns a new instance of GoogleCloudAiplatformV1IntegratedGradientsAttribution.



7986
7987
7988
 
# File 'lib/google/apis/aiplatform_v1/classes.rb', line 7986

def initialize(**args)
   update!(**args)
end

Instance Attribute Details

#blur_baseline_config ⇒ Google::Apis::AiplatformV1::GoogleCloudAiplatformV1BlurBaselineConfig

Config for blur baseline. When enabled, a linear path from the maximally blurred image to the input image is created. Using a blurred baseline instead of zero (black image) is motivated by the BlurIG approach explained here: https://arxiv.org/abs/2004.03383 Corresponds to the JSON property blurBaselineConfig

Returns:



7968
7969
7970
 
# File 'lib/google/apis/aiplatform_v1/classes.rb', line 7968

def blur_baseline_config
  @blur_baseline_config
end

#smooth_grad_config ⇒ Google::Apis::AiplatformV1::GoogleCloudAiplatformV1SmoothGradConfig

Config for SmoothGrad approximation of gradients. When enabled, the gradients are approximated by averaging the gradients from noisy samples in the vicinity of the inputs. Adding noise can help improve the computed gradients. Refer to this paper for more details: https://arxiv.org/pdf/1706.03825.pdf Corresponds to the JSON property smoothGradConfig

Returns:



7976
7977
7978
 
# File 'lib/google/apis/aiplatform_v1/classes.rb', line 7976

def smooth_grad_config
  @smooth_grad_config
end

#step_count ⇒ Fixnum

Required. The number of steps for approximating the path integral. A good value to start is 50 and gradually increase until the sum to diff property is within the desired error range. Valid range of its value is [1, 100], inclusively. Corresponds to the JSON property stepCount

Returns:

  • (Fixnum) 


7984
7985
7986
 
# File 'lib/google/apis/aiplatform_v1/classes.rb', line 7984

def step_count
  @step_count
end

Instance Method Details

#update!(**args) ⇒ Object

Update properties of this object



7991
7992
7993
7994
7995
 
# File 'lib/google/apis/aiplatform_v1/classes.rb', line 7991

def update!(**args)
  @blur_baseline_config = args[:blur_baseline_config] if args.key?(:blur_baseline_config)
  @smooth_grad_config = args[:smooth_grad_config] if args.key?(:smooth_grad_config)
  @step_count = args[:step_count] if args.key?(:step_count)
end