ANDRE: An Attention-based Neuro-symbolic Differentiable Rule Extractor

TL;DR

ANDRE introduces an attention-based neuro-symbolic ILP framework that learns interpretable first-order rules from noisy and probabilistic data, improving stability and rule extraction accuracy.

Contribution

It proposes a fully differentiable, attention-driven approach to ILP that overcomes limitations of existing symbolic and neuro-symbolic methods in uncertain settings.

Findings

Achieves competitive or superior predictive performance on ILP benchmarks.

Robustly recovers symbolic rules under noisy supervision.

Outperforms existing differentiable ILP methods in rule quality and stability.

Abstract

Inductive Logic Programming (ILP) aims to learn interpretable first-order rules from data, but existing symbolic and neuro-symbolic approaches struggle to scale to noisy and probabilistic settings. Classical ILP relies on discrete combinatorial rule search and is brittle under uncertainty, while differentiable ILP methods typically depend on predefined rule templates or inaccurate fuzzy operators that suffer from vanishing gradients or poor approximation of logical structure when reasoning over probabilistic predicate valuations. This paper proposes an Attention-based Neuro-symbolic Differentiable Rule Extractor (ANDRE), a novel ILP framework that learns first-order logic programs by optimizing over a continuous rule space with attention-based logical operators. ANDRE replaces both rule templates and logical operators with fully differentiable, attention-driven conjunction and disjunction operators that approximate logical min-max semantics, enabling accurate, stable, and interpretable reasoning over probabilistic data. By softly selecting, negating, or excluding predicates within each rule, ANDRE supports flexible rule induction while preserving symbolic structure. Extensive experiments on classical ILP benchmarks, large-scale knowledge bases, and synthetic datasets with probabilistic predicates and noisy supervision demonstrate that ANDRE achieves competitive or superior predictive performance while reliably recovering correct symbolic rules under uncertainty. In particular, ANDRE remains robust to moderate label noise, substantially outperforming existing differentiable ILP methods in both rule extraction quality and stability.