View Source GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_ExplanationConfig (google_api_machine_learning v0.28.1)

Message holding configuration options for explaining model predictions. There are three feature attribution methods supported for TensorFlow models: integrated gradients, sampled Shapley, and XRAI. Learn more about feature attributions.

Attributes

integratedGradientsAttribution (type: GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_IntegratedGradientsAttribution.t, default: nil) - Attributes credit by computing 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
sampledShapleyAttribution (type: GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_SampledShapleyAttribution.t, default: nil) - An attribution method that approximates Shapley values for features that contribute to the label being predicted. A sampling strategy is used to approximate the value rather than considering all subsets of features.
xraiAttribution (type: GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_XraiAttribution.t, default: nil) - Attributes credit by computing the XRAI taking advantage of the model's fully differentiable structure. Refer to this paper for more details: https://arxiv.org/abs/1906.02825 Currently only implemented for models with natural image inputs.

Summary

Types

t()

Functions

decode(value, options)

Unwrap a decoded JSON object into its complex fields.

Types

t()

@type t() :: %GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_ExplanationConfig{
  integratedGradientsAttribution:
    GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_IntegratedGradientsAttribution.t()
    | nil,
  sampledShapleyAttribution:
    GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_SampledShapleyAttribution.t()
    | nil,
  xraiAttribution:
    GoogleApi.MachineLearning.V1.Model.GoogleCloudMlV1_XraiAttribution.t() | nil
}

Functions

decode(value, options)

@spec decode(struct(), keyword()) :: struct()

Unwrap a decoded JSON object into its complex fields.