// ⌖ Interpretability
Feature Attribution (Integrated Gradients / SHAP)
Measures how much each input feature contributed to a prediction.
特征归因（积分梯度 / SHAP） 2017 Foundational core
IN PLAIN WORDS
Integrated Gradients accumulates gradients along a path from a baseline to the input; SHAP fairly splits credit with Shapley values. Good for explaining tabular models and classifiers.
Used in: Credit-model explanations, medical model audits
01Original paper
Axiomatic Attribution for Deep Networks
Mukund Sundararajan, Ankur Taly, Qiqi Yan · 2017
arXiv ↗ ✓ Verified · arXiv API
Note: SHAP：arXiv:1705.07874（2017）
02Where it sits
Learn the left column first; the right column builds on this one.
Learn first
Then you can learn
- Nothing in the atlas builds on it yet
Related:Grad-CAMSparse Autoencoders (SAE)
Left: prerequisites · right: builds on this · dashed: related. Full map →
03Learning resources
Official docs, explainers, courses and reference code. All links checked.
04Where the course teaches it
For AI / data
- Verification: verified, checked 2026-10-21.
- JSON: /api/learn/algorithms.json (id:
attribution)