Responsible AI Mitigations

Overview

  • Installation guide
  • Getting Started
  • How this library works with the Responsible AI Toolbox
  • API reference
  • Gallery
    • Tutorials
      • DataProcessing
        • Encoding Examples
        • Data Scaling
        • BasicImputer Example
        • IterativeDataImputer Example
        • KNNDataImputer Example
        • SeqFeatSelection Example
        • CatBoostSelection Example
        • Checking Correlations Between Variables - A Comprehensive Guide
        • Rebalance Class (imblearn)
        • Synthesizer Class (SDV)
      • DataBalanceAnalysis
        • Data Balance Analysis using the Adult Census Income dataset
        • Summary​
        • Analysis:
        • Python Data Balance Analysis
        • Context
        • Usage
        • Example Notebook
        • Explanations of Data Balance Measures
        • Mitigation
      • Cohort
        • Defining a Cohort
        • Managing cohorts
        • Cohort Manager - Scenarios and Examples
        • Decoupled Classifiers
      • Using scikit-learn’s Pipeline
        • Using Pipelines
    • Case Studies
      • Using the DataProcessing Module
        • Simple Example
        • Case Study 1
        • Case Study 2
        • Case Study 3
        • Case Study 1 - Multiple Runs
        • Case Study 2 - Multiple Runs
        • Case Study 3 - Multiple Runs
      • Using the DataBalanceAnalysis Module
        • End to End Data Balance and Error Mitigation
      • Using the Cohort Module
        • Cohort Case Study 1
        • Cohort Case Study 1 - Rebalancing
        • Cohort Case Study 1 - Using RAI ErrorAnalysis
        • Cohort Case Study 2
        • Cohort Case Study 3
        • Moving Cohorts from raimitigations to raiwidgets (and vice versa)
        • Decoupled Classifiers Case Study 1
        • Decoupled Classifiers Case Study 2
        • Decoupled Classifiers Case Study 3

Modules

  • DataProcessing
  • DataBalanceAnalysis
  • Cohort
  • Utils
Responsible AI Mitigations
  • Gallery
  • Edit on GitHub

Gallery

The best way to get started is to go through the notebook tutorials, which showcase the benefits and usage of each mitigation and metrics available in the Responsible AI Mitigations Library. Here you can find all notebooks grouped according to their goals.

Tutorials

DataProcessing

Encoding Examples
Data Scaling
BasicImputer Example
IterativeDataImputer Example
KNNDataImputer Example
SeqFeatSelection Example
CatBoostSelection Example
Checking Correlations Between Variables - A Comprehensive Guide
Rebalance Class (imblearn)
Synthesizer Class (SDV)

DataBalanceAnalysis

Data Balance Analysis using the Adult Census Income dataset
Python Data Balance Analysis

Cohort

Defining a Cohort
Managing cohorts
Cohort Manager - Scenarios and Examples
Decoupled Classifiers

Using scikit-learn’s Pipeline

Using Pipelines

Case Studies

Using the DataProcessing Module

Simple Example
Case Study 1
Case Study 2
Case Study 3
Case Study 1 - Multiple Runs
Case Study 2 - Multiple Runs
Case Study 3 - Multiple Runs

Using the DataBalanceAnalysis Module

End to End Data Balance and Error Mitigation

Using the Cohort Module

Cohort Case Study 1
Cohort Case Study 1 - Rebalancing
Cohort Case Study 1 - Using RAI ErrorAnalysis
Cohort Case Study 2
Cohort Case Study 3
Moving Cohorts from raimitigations to raiwidgets (and vice versa)
Decoupled Classifiers Case Study 1
Decoupled Classifiers Case Study 2
Decoupled Classifiers Case Study 3
Previous Next

© Copyright 2022, Microsoft.

Built with Sphinx using a theme provided by Read the Docs.