GlanceNets: Interpretabile, Leak-proof Concept-based Models

Emanuele Marconato, Andrea Passerini, Stefano Teso

Keywords: concept-based models, disentanglement, alignment, interpretability, leakage, open-set-recongintion

TL;DR: We present GlanceNets, a new class of Concept-based models that achieve better interpretability via disentanglement integrations and prevent concept leakage with open set recognition

Abstract: There is growing interest in concept-based moels (CBMs) that combine high-performance and interpretability by acquiring and reasoning with a vocabulary of high-level concepts. A key requirement is that the concepts be interpretable. Existing CBMs tackle this desideratum using a variety of heuristics based on unclear notions of interpretability, and fail to acquire concepts with the intended semantics. We address this by providing a clear definition of interpretability in terms of alignment between the model’s representation and an underlying data generation process, and introduce GlanceNets, a new CBM that exploits techniques from disentangled representation learning and open-set recognition to achieve alignment thus improving the interpretability of the learned concepts. We show that GlanceNets, paired with concept-level supervision, achieve better alignment than state-of-the-art approaches while preventing spurious information from unintendedly leaking into the learned concepts.

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/glancenets-interpretabile-leak-proof-concept/code)