AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Explaining Explanations: Axiomatic Feature Interactions for Deep Networks

Joseph D. Janizek, Pascal Sturmfels, Su-In Lee

JMLR 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Self-explaining and visual explanation networks · 96 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace ReLU with SoftPlus (beta = 10) at explanation time (ablation)Integrated Hessians · ReLU networksinteraction detectabilityImproved
integrate Hessian along a path from baseline instead of using input Hessian (ablation)input Hessian · XOR functioninteraction detectionImproved
Integrated Hessians versus existing interaction methodsIntegrated Hessians · Remove and Retrain, simulated interactionserror increase speedImproved
Integrated Hessians versus monte-carlo Shapley Interaction IndexIntegrated Hessians · Remove and Retrain, additional interaction typeserror increase speedNo clear effect
increase number of input features (5, 50, 500)Integrated Hessians · computation time, 5-layer softplus networkrun timeMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy