Visual Perceptual to Conceptual First-Order Rule Learning Networks
2026-04-09 • Artificial Intelligence
Artificial IntelligenceMachine Learning
AI summaryⓘ
The authors present a method called γILP that learns logical rules directly from images without needing labeled data or predefined features. Their approach connects images to rule learning through a fully differentiable process, enabling the system to identify patterns and create rules from visual input. They tested γILP on both traditional symbolic data and various image datasets, showing it works well in different settings.
Inductive Logic ProgrammingDifferentiable LearningRule LearningExplainable AIRelational DataSymbolic ReasoningImage UnderstandingKandinsky Patterns
Authors
Kun Gao, Davide Soldà, Thomas Eiter, Katsumi Inoue
Abstract
Learning rules plays a crucial role in deep learning, particularly in explainable artificial intelligence and enhancing the reasoning capabilities of large language models. While existing rule learning methods are primarily designed for symbolic data, learning rules from image data without supporting image labels and automatically inventing predicates remains a challenge. In this paper, we tackle these inductive rule learning problems from images with a framework called γILP, which provides a fully differentiable pipeline from image constant substitution to rule structure induction. Extensive experiments demonstrate that γILP achieves strong performance not only on classical symbolic relational datasets but also on relational image data and pure image datasets, such as Kandinsky patterns.