analiza aiqbench
This research addresses a critical flaw in explainable AI, demonstrating that Concept Bottleneck Models can develop spurious shortcuts where high task accuracy masks unreliable symbol detection. The authors propose a reliability-aware training strategy using shared detectors and penalization to ensure explanations remain faithful even when concept supervision is reduced. For practitioners building interpretable systems, this work provides essential metrics for validating whether model decisions truly rely on intended human-interpretable concepts rather than hidden biases.
Assessing Reliability of Symbol Detection in Concept Bottleneck Models
arXiv:2606.16535v1 Announce Type: new Abstract: Concept Bottleneck Models (CBMs) are a relevant tool for explainable Artificial Intelligence because they make their predictions through human-interpretable symbols. However, high task accuracy does not guarantee that these symbols are detected faithfully: jointly trained CBMs may encode task-specific shortcuts in the bottleneck, making their explanations unreliable. In this paper, we study concept-detection reliability by swapping independently trained concept detectors and classification heads that share the same symbolic vocabulary. We use the resulting performance degradation, concept-level metrics, and symbol-wise uncertainty estimates to identify concepts that are especially prone to spurious firing. Finally, we propose a reliability-aware training strategy in which a shared concept detector is optimized with multiple classification heads and penalized for relying on globally or instance-wise unreliable symbols. On CUB-200-2011 with full concept supervision, detectors and heads are almost freely interchangeable (swap drop below one accuracy point, relative retention above $99\%$, and no concept detected below chance), whereas on a controlled synthetic task we show that, as the concept-supervision weight is reduced, models keep near-perfect task accuracy while swapped accuracy and agreement with the ground-truth concepts collapse to chance. Our reliability-aware training substantially mitigates this leakage, roughly doubling swap accuracy in the leaky regime.
Abstract:Concept Bottleneck Models (CBMs) are a relevant tool for explainable Artificial Intelligence because they make their predictions through human-interpretable symbols. However, high task accuracy does not guarantee that these symbols are detected faithfully: jointly trained CBMs may encode task-specific shortcuts in the bottleneck, making their explanations unreliable. In this paper, we study concept-detection reliability by swapping independently trained concept detectors and classification heads that share the same symbolic vocabulary. We use the resulting performance degradation, concept-level metrics, and symbol-wise uncertainty estimates to identify concepts that are especially prone to spurious firing. Finally, we propose a reliability-aware training strategy in which a shared concept detector is optimized with multiple classification heads and penalized for relying on globally or instance-wise unreliable symbols. On CUB-200-2011 with full concept supervision, detectors and heads are almost freely interchangeable (swap drop below one accuracy point, relative retention above $99\%$, and no concept detected below chance), whereas on a controlled synthetic task we show that, as the concept-supervision weight is reduced, models keep near-perfect task accuracy while swapped accuracy and agreement with the ground-truth concepts collapse to chance. Our reliability-aware training substantially mitigates this leakage, roughly doubling swap accuracy in the leaky regime.
Machine Learning (cs.LG); Computer Vision and Pattern Recognition (cs.CV); Symbolic Computation (cs.SC)
arXiv:2606.16535 [cs.LG]
(or arXiv:2606.16535v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2606.16535 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 299 cuvinte
Subiecte conexe
When History Lies: Evaluating and Improving Tool Use under Misleading Multi-Turn Histories
arXiv · 7 aug. 2026
Online Continual Learning in Edge-Cloud Applications
arXiv · 27 oct. 2026
Estimating Uncertainty from Reasoning: A Large-Scale Study of Multi- and Crosslingual MCQA Performance in LLMs
arXiv · 7 sept. 2026
From Context-Aware to Conflict-Aware: Generalizing Contrastive Decoding for Knowledge Conflict in LLMs
arXiv · 10 iun. 2026