AI RESEARCH

Concepts in Motion: Temporal Concept Bottleneck Model for Interpretable Video Classification

arXiv CS.CV

ArXi:2509.20899v3 Announce Type: replace Concept Bottleneck Models (CBMs) enable interpretable image classification by structuring predictions around human-understandable concepts, but extending this paradigm to video remains challenging due to the difficulty of extracting concepts and modeling them over time. In this paper, we