Abstract: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.
| Comments: | TMLR March 2026 |
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2508.14255 [cs.LG] |
| (or arXiv:2508.14255v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2508.14255 arXiv-issued DOI via DataCite |
From: Haotian Xu [view email]
[v1]
Tue, 19 Aug 2025 20:23:18 UTC (14,765 KB)
[v2]
Thu, 30 Apr 2026 21:23:39 UTC (14,763 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。