Diffeomorphic Counterfactuals with Generative Models
Journal article, 2024
Counterfactuals can explain classification decisions of neural networks in a human interpretable way. We propose a simple but effective method to generate such counterfactuals. More specifically, we perform a suitable diffeomorphic coordinate transformation and then perform gradient ascent in these coordinates to find counterfactuals which are classified with great confidence as a specified target class. We propose two methods to leverage generative models to construct such suitable coordinate systems that are either exactly or approximately diffeomorphic. We analyze the generation process theoretically using Riemannian differential geometry and validate the quality of the generated counterfactuals using various qualitative and quantitative measures.
Manifolds
Computational modeling
Geometry
Semantics
Generative Models
Task analysis
Counterfactual Explanations
Explainable Artificial Intelligence
Artificial intelligence
Data models
Data Manifold
Author
Ann Kathrin Dombrowski
Technische Universität Berlin
Jan Gerken
Chalmers, Mathematical Sciences, Algebra and geometry
Klaus Robert Muller
Technische Universität Berlin
Pan Kessel
Genentech
IEEE Transactions on Pattern Analysis and Machine Intelligence
0162-8828 (ISSN) 19393539 (eISSN)
Vol. 46 5 3257-3274 10345703Subject Categories
Computer Science
DOI
10.1109/TPAMI.2023.3339980
PubMed
38055368