AI RESEARCH

Matryoshka Concept Bottleneck Models

arXiv CS.LG

ArXi:2605.20612v1 Announce Type: new Concept Bottleneck Models (CBMs) have emerged as a prominent paradigm for interpretable deep learning, learning by grounding predictions in human-understandable concepts. However, their practical deployment is hindered by the high cost of test-time intervention, as correcting model errors typically requires human experts to manually inspect and verify a large set of predicted concepts.