By fixing CLIP-derived language similarity maps as the coefficient matrix in non-negative matrix factorization, this method produces named, faithful concept explanations for frozen image classifiers.
Towards automatic concept-based explanations.Ad- vances in neural information processing systems, 32
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2representative citing papers
H-Sets detects higher-order feature interactions in image classifiers via Hessian-guided pair merging and attributes them with IDG-Vis to generate more interpretable saliency maps than existing marginal or coarse methods.
citing papers explorer
-
Naming the Concepts Classifiers Rely On: Language-Anchored Decomposition for Faithful Explanation
By fixing CLIP-derived language similarity maps as the coefficient matrix in non-negative matrix factorization, this method produces named, faithful concept explanations for frozen image classifiers.
-
H-Sets: Hessian-Guided Discovery of Set-Level Feature Interactions in Image Classifiers
H-Sets detects higher-order feature interactions in image classifiers via Hessian-guided pair merging and attributes them with IDG-Vis to generate more interpretable saliency maps than existing marginal or coarse methods.