Vertex AI V1 API - Class Google::Cloud::AIPlatform::V1::IntegratedGradientsAttribution (v0.25.0)

Reference documentation and code samples for the Vertex AI V1 API class Google::Cloud::AIPlatform::V1::IntegratedGradientsAttribution.

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

Inherits

  • Object

Extended By

  • Google::Protobuf::MessageExts::ClassMethods

Includes

  • Google::Protobuf::MessageExts

Methods

#blur_baseline_config

def blur_baseline_config() -> ::Google::Cloud::AIPlatform::V1::BlurBaselineConfig
Returns
  • (::Google::Cloud::AIPlatform::V1::BlurBaselineConfig) — Config for IG with 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

#blur_baseline_config=

def blur_baseline_config=(value) -> ::Google::Cloud::AIPlatform::V1::BlurBaselineConfig
Parameter
  • value (::Google::Cloud::AIPlatform::V1::BlurBaselineConfig) — Config for IG with 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

Returns
  • (::Google::Cloud::AIPlatform::V1::BlurBaselineConfig) — Config for IG with 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

#smooth_grad_config

def smooth_grad_config() -> ::Google::Cloud::AIPlatform::V1::SmoothGradConfig
Returns
  • (::Google::Cloud::AIPlatform::V1::SmoothGradConfig) — 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

#smooth_grad_config=

def smooth_grad_config=(value) -> ::Google::Cloud::AIPlatform::V1::SmoothGradConfig
Parameter
  • value (::Google::Cloud::AIPlatform::V1::SmoothGradConfig) — 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

Returns
  • (::Google::Cloud::AIPlatform::V1::SmoothGradConfig) — 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

#step_count

def step_count() -> ::Integer
Returns
  • (::Integer) — 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.

#step_count=

def step_count=(value) -> ::Integer
Parameter
  • value (::Integer) — 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.

Returns
  • (::Integer) — 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.