This repository is a curated collection of academic papers, blog posts, and lectures on the topic of interpretability and explainability in machine learning. It serves as a comprehensive resource for researchers, engineers, and practitioners seeking to understand the concepts, methods, challenges, and applications of XAI. The collection aims to provide a structured overview of the field, highlighting key debates and research directions.

How It Works

The repository organizes literature by themes such as overviews, perspectives on human-AI interaction, evaluation criteria, adversarial attacks on explanations, and specific application areas like NLP and computer vision. It references seminal works and recent advancements, categorizing papers by their contribution to understanding the "what," "how," and "why" of explainable AI, including discussions on the limitations of current methods and the importance of human-centered evaluation.

Highlighted Details

• Extensive coverage of seminal works and recent research in XAI, including papers from top-tier conferences (NeurIPS, ICML, ACL, etc.).
• Detailed discussion on the limitations and potential pitfalls of various interpretability methods, such as attention mechanisms and saliency maps.
• Exploration of user-centric explanations, recourse, and the implications of regulations like GDPR on explainable AI.
• Categorization of literature into overviews, perspectives, evaluation criteria, adversarial examples, and specific application areas.

Maintenance & Community

This is a personal collection of literature, with no explicit mention of active maintenance or community engagement channels.

Licensing & Compatibility

The repository itself is not software and does not have a license. The linked academic papers are subject to their respective publication licenses and copyright.

Limitations & Caveats

This repository is a literature compilation and does not provide any code, tools, or implementations for XAI methods. Its value is purely informational, requiring users to seek out and implement the discussed techniques themselves.