Neural Algorithmic Reasoning for Combinatorial Optimisation

TL;DR

This paper explores neural algorithmic reasoning to enhance solving combinatorial optimization problems, pre-training neural models on algorithms to outperform traditional deep learning approaches.

Contribution

It introduces a method of pre-training neural networks on relevant algorithms to improve performance on combinatorial optimization problems.

Findings

Pre-training on algorithms improves solution quality.

The approach outperforms non-algorithmically informed models.

Neural models can learn to solve complex combinatorial problems more effectively.

Abstract

Solving NP-hard/complete combinatorial problems with neural networks is a challenging research area that aims to surpass classical approximate algorithms. The long-term objective is to outperform hand-designed heuristics for NP-hard/complete problems by learning to generate superior solutions solely from training data. Current neural-based methods for solving CO problems often overlook the inherent "algorithmic" nature of the problems. In contrast, heuristics designed for CO problems, e.g. TSP, frequently leverage well-established algorithms, such as those for finding the minimum spanning tree. In this paper, we propose leveraging recent advancements in neural algorithmic reasoning to improve the learning of CO problems. Specifically, we suggest pre-training our neural model on relevant algorithms before training it on CO instances. Our results demonstrate that by using this learning setup, we achieve superior performance compared to non-algorithmically informed deep learning models.