Knowledge gathering: XAI

[youyinnn]

1. read the black-white book of XAI
2. read 3-4 front tier papers about XAI maybe

[youyinnn]

Reading log

• [ ] The book: Interpretable Machine Learning

Main reading

• [x] Paper: AI Explainability Whitepaper

  Feature attributions

  • [ ] "Why Should I Trust You?": Explaining the Predictions of Any Classifier: LIME
  • [ ] A Unified Approach to Interpreting Model Predictions: SHAP
  • [ ] Anchors: High-Precision Model-Agnostic Explanations: Anchors, they have comparision with LIME
  • [ ] The many Shapley values for model explanation: A thorough discussion of applying Shapley in XAI
  • [x] Axiomatic Attribution for Deep Networks: Integrated Gradients

  Concept extraction

  • [ ] Towards Automatic Concept-based Explanations : Concept extraction

Survey/Gathering reading

• [ ] Illuminating the Black Box: Interpreting Deep Neural Network Models for Psychiatric Research

Side reading

• [x] Paper: Testing Framework for Black-box AI Models: from IBM Research AI Team
• [x] Website: Blackbox Testing for Machine Learning Models: listed some testing approach
• [x] Website: Black-box Testing: from softwaretesinghelp.com
• [ ] Explanation in Artificial Intelligence: Insights from the Social Sciences
• [x] Paper: Why Are We Using Black Box Models in AI When We Don’t Need To? A Lesson From an Explainable AI Competition
• [ ] Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead

[youyinnn]

About the s2search:

1. Building a Better Search Engine for Semantic Scholar
2. LightGBM

[youyinnn]

A Survey of Methods for Explaining Black Box Models

[youyinnn]

Visualizing Variable Importance and Variable Interaction Effects in Machine Learning Models

[youyinnn]

PDP and H-statistic: Discovering Interaction Effects in Ensemble Models

[youyinnn]

Introduction to Machine Learning Model Interpretation

[youyinnn]

PDP and H-statistic: Discovering Interaction Effects in Ensemble Models

From the paper: https://arxiv.org/pdf/1805.04755.pdf

It has a remark for H-statistic: