As AI models grow larger, the demand for accountability and interpretability has become increasingly critical for understanding their decision-making processes. Concept Bottleneck Models (CBMs) have gained attention for enhancing interpretability by mapping inputs to intermediate concepts before making final predictions. However, CBMs often suffer from information leakage, where additional input data, not captured by the concepts, is used to improve task performance, complicating the interpretation of downstream predictions. In this paper, we introduce a novel approach for training both joint and sequential CBMs that allows us to identify and control leakage using decision trees. Our method quantifies leakage by comparing the decision paths of hard CBMs with their soft, leaky counterparts. Specifically, we show that soft leaky CBMs extend the decision paths of hard CBMs, particularly in cases where concept information is incomplete. Using this insight, we develop a technique to better inspect and manage leakage, isolating the subsets of data most affected by this. Through synthetic and real-world experiments, we demonstrate that controlling leakage in this way not only improves task accuracy but also yields more informative and transparent explanations.
Julearn: an easy-to-use library for leakage-free evaluation and inspection of ML models
Institute of Neuroscience and Medicine (INM-7: Brain and Behaviour), Research Centre Jülich（于利希研究中心神经科学与医学研究所（INM-7：脑与行为））；Institute of Systems Neuroscience, Heinrich Heine University Düsseldorf（杜塞尔多夫海因里希·海涅大学系统神经科学研究所）；Max Planck School of Cognition（马克斯·普朗克认知科学学院）
TRACE: A Concept Bottleneck Model for Longitudinal 3D Glioblastoma Response Assessment
Department of Biomedical and Healthcare Data Engineering, Faculty of Engineering, Cairo University（生物医学与健康数据工程系，工程学院，开罗大学）；German Research Center for Artificial Intelligence (DFKI)（德国人工智能研究中心（DFKI））；University of Oldenburg（奥尔登堡大学）
