Graph Concept Bottleneck Models
This work addresses the problem of improving interpretability and intervention effectiveness in deep learning models for image classification, representing an incremental advancement over existing CBMs.
The paper tackles the limitation of Concept Bottleneck Models (CBMs) ignoring relationships among concepts by proposing GraphCBMs, which incorporate latent concept graphs to model concept correlations. Results on real-world image classification tasks show superior performance and enhanced interpretability through concept structure information.
Concept Bottleneck Models (CBMs) provide explicit interpretations for deep neural networks through concepts and allow intervention with concepts to adjust final predictions. Existing CBMs assume concepts are conditionally independent given labels and isolated from each other, ignoring the hidden relationships among concepts. However, the set of concepts in CBMs often has an intrinsic structure where concepts are generally correlated: changing one concept will inherently impact its related concepts. To mitigate this limitation, we propose GraphCBMs: a new variant of CBM that facilitates concept relationships by constructing latent concept graphs, which can be combined with CBMs to enhance model performance while retaining their interpretability. Our experiment results on real-world image classification tasks demonstrate Graph CBMs offer the following benefits: (1) superior in image classification tasks while providing more concept structure information for interpretability; (2) able to utilize latent concept graphs for more effective interventions; and (3) robust in performance across different training and architecture settings.