AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

Concept Activation Regions: A Generalized Framework For Concept-Based Explanations

Jonathan Crabbé, Mihaela van der Schaar

NeurIPS 2022 · Probabilistic and scientific modeling › Molecules and proteins · approach: Probabilistic embeddings and relational inference · 16 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
Use TCAR scores instead of TCAV scores for global class-concept explanationsConcept-based explanations · MNISTPearson correlation with true concept proportion: .60 -> 1.00Improved
Use TCAR scores instead of TCAV scores for global class-concept explanationsConcept-based explanations · ECGPearson correlation with true concept proportion: .74 -> .99Improved
Use TCAR scores instead of TCAV scores for global class-concept explanationsConcept-based explanations · CUBPearson correlation with true concept proportion: .52 -> .86Improved
Tune kernel type, width, and error penalty with Bayesian optimization (ablation)CAR classifier · CUB Inception penultimate layerconcept classifier accuracy: (93 ± .2) % -> (93 ± .2) %No clear effect
Use CAR-based Integrated Gradients for concept density rather than vanilla Integrated Gradients for model outputFeature importance · MNIST, ECG, and CUBPearson correlation with vanilla feature importanceMixed
Use CAR-based feature importance to compare related conceptsConcept feature importance · MNISTPearson correlation between Loop and Curvature importance: r=.71Improved
Use CAR-based feature importance to compare mutually exclusive conceptsConcept feature importance · CUBPearson correlation between striped and solid back pattern importance: r=-.86Mixed
Use the full concept density rather than a sparse density using only SVC support vectors (ablation)CAR-based feature importance · Concept feature importancefeature attributionImproved
Increase concept-set size beyond N^c > 200 (ablation)CAR classifier · MNISTconcept classifier accuracyNo clear effect

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