VisionLogic: From Neuron Activations to Causally Grounded Concept Rules for Vision Models

Chuqin Geng, Yuhe Jiang, Ziyu Zhao, Zhaoyue Wang, Haolin Ye, Anqi Xing, Li Zhang, Xujie Si

Keywords: Interpretability, Neural-symbolic, Logical rules, Convolutional neural network, Vision transformer, XAI

Abstract: While concept-based explanations improve interpretability over local attributions, they often rely on correlational signals and lack causal validation. We introduce VisionLogic, a novel neural–symbolic framework that produces faithful, hierarchical explanations as global logical rules over causally validated concepts. VisionLogic first learns activation thresholds to convert neuron activations into a reusable predicate vocabulary and induces class-level logical rules from these predicates. It then grounds predicates to visual concepts via ablation-based causal tests with iterative region refinement, ensuring that discovered concepts correspond to features that are causal for predicate activation. Across different vision architectures such as CNNs and ViTs, it produces interpretable concepts and compact rules that largely preserve the original model’s predictive performance. In our large-scale human evaluations, VisionLogic’s concept explanations significantly improve participants’ understanding of model behavior over prior concept-based methods. VisionLogic bridges neural representations and symbolic reasoning, providing more trustworthy explanations suited for safety-critical applications.

Primary Area: interpretability and explainable AI

Submission Number: 1005