ULLER: A Unified Language for Learning and Reasoning
Emile van Krieken, Samy Badreddine, Robin Manhaeve, Eleonora Giunchiglia
TL;DR
ULLER addresses the lack of reproducibility and interoperability in neuro-symbolic AI by introducing a unified, first-order language with a dedicated statement syntax that decouples knowledge representation from NeSy system semantics. It formalizes the syntax and supports multiple semantics—classical, probabilistic, fuzzy, and sampling—through an interpretation-driven framework that treats neural components as conditional distributions. The work defines learning with parameterised interpretations and gradient-estimation strategies, enabling end-to-end training within the NeSy pipeline. By enabling frictionless sharing of knowledge and cross-system evaluation, ULLER aims to reduce barriers to entry and foster standardized benchmarks across diverse NeSy systems.
Abstract
The field of neuro-symbolic artificial intelligence (NeSy), which combines learning and reasoning, has recently experienced significant growth. There now are a wide variety of NeSy frameworks, each with its own specific language for expressing background knowledge and how to relate it to neural networks. This heterogeneity hinders accessibility for newcomers and makes comparing different NeSy frameworks challenging. We propose a unified language for NeSy, which we call ULLER, a Unified Language for LEarning and Reasoning. ULLER encompasses a wide variety of settings, while ensuring that knowledge described in it can be used in existing NeSy systems. ULLER has a neuro-symbolic first-order syntax for which we provide example semantics including classical, fuzzy, and probabilistic logics. We believe ULLER is a first step towards making NeSy research more accessible and comparable, paving the way for libraries that streamline training and evaluation across a multitude of semantics, knowledge bases, and NeSy systems.