Timezone: »

GlanceNets: Interpretable, Leak-proof Concept-based Models
Emanuele Marconato · Andrea Passerini · Stefano Teso

Fri Dec 09 04:40 AM -- 04:50 AM (PST) @
Event URL: https://openreview.net/forum?id=Ynwqj04Elc »

There is growing interest in concept-based models (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 causal 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.

Author Information

Emanuele Marconato (University of Trento, via Calepina 14 38122 Trento (TN) Italy VAT IT00340520220)
Andrea Passerini (University of Trento VAT IT00340520220)
Stefano Teso (University of Trento)

More from the Same Authors