Reinforcement Learning for Scalable Logic Optimization with Graph Neural Networks

TL;DR

This paper introduces a reinforcement learning approach combined with graph neural networks and a new node embedding technique to improve logic optimization, achieving comparable results to traditional methods on small circuits and significantly better on larger graphs.

Contribution

It presents a novel scalable method integrating graph neural networks and reinforcement learning for logic optimization, outperforming existing heuristics on large circuits.

Findings

Achieves similar size reduction as ABC on small circuits

Outperforms ABC by 1.5-1.75x on larger graphs

Introduces a scalable node embedding method

Abstract

Logic optimization is an NP-hard problem commonly approached through hand-engineered heuristics. We propose to combine graph convolutional networks with reinforcement learning and a novel, scalable node embedding method to learn which local transforms should be applied to the logic graph. We show that this method achieves a similar size reduction as ABC on smaller circuits and outperforms it by 1.5-1.75x on larger random graphs.