SLDR-DL: A Framework for SLD-Resolution with Deep Learning

TL;DR

This paper presents SLDR-DL, a novel framework combining deep learning with SLD-resolution to enhance logical reasoning by learning from past resolution successes.

Contribution

It introduces a new deep learning-based approach to guide SLD-resolution, integrating neural networks with logical rule reasoning in Prolog.

Findings

Neural networks learn from successful resolution processes

Framework enables logical rule reasoning with deep learning guidance

Implementation demonstrates practical integration of neural networks and logic programming

Abstract

This paper introduces an SLD-resolution technique based on deep learning. This technique enables neural networks to learn from old and successful resolution processes and to use learnt experiences to guide new resolution processes. An implementation of this technique is named SLDR-DL. It includes a Prolog library of deep feedforward neural networks and some essential functions of resolution. In the SLDR-DL framework, users can define logical rules in the form of definite clauses and teach neural networks to use the rules in reasoning processes.