Hypothesis Generation via LLM-Automated Language Bias for ILP

TL;DR

This paper introduces a novel approach combining multi-agent LLMs and ILP to automate and improve hypothesis generation, reducing reliance on expert-crafted biases and enhancing interpretability and robustness.

Contribution

It presents a method where LLMs generate language bias for ILP, enabling automated, scalable, and explainable hypothesis construction from raw text.

Findings

Outperforms traditional ILP in diverse scenarios

Reduces reliance on predefined symbolic structures

Enhances robustness against noise

Abstract

Inductive Logic Programming (ILP) is a principled approach for generalizing regularities from data and constructing hypotheses as interpretable logic programs. However, a key limitation is its reliance on expert-crafted language bias - the predicate inventory, types, and mode declarations that delimit the search space. We propose hypothesis generation via LLM-automated language bias: multi-agent LLMs design the bias from raw text and translate descriptions into typed facts, and a robust ILP solver induces rules under a global consistency objective. This approach reduces traditional ILP's reliance on predefined symbolic structures and the noise sensitivity of LLM-only pipelines that directly generate hypotheses as text or code. Extensive experiments in diverse, challenging scenarios validate superior performance, providing a practical, explainable, and verifiable route to hypothesis generation.