Reinforcement learning for optimal error correction of toric codes

TL;DR

This paper introduces a deep reinforcement learning approach to optimize error correction in toric codes, achieving near-theoretical thresholds and discovering policies akin to traditional methods without prior bias.

Contribution

The study demonstrates the effectiveness of deep reinforcement learning in designing high-threshold decoders for toric codes, matching near-optimal performance.

Findings

Achieves near 11% threshold for uncorrelated noise

Agent's policy resembles minimum weight perfect matching

Deep convolutional networks enable effective decoding

Abstract

We apply deep reinforcement learning techniques to design high threshold decoders for the toric code under uncorrelated noise. By rewarding the agent only if the decoding procedure preserves the logical states of the toric code, and using deep convolutional networks for the training phase of the agent, we observe near-optimal performance for uncorrelated noise around the theoretically optimal threshold of 11%. We observe that, by and large, the agent implements a policy similar to that of minimum weight perfect matchings even though no bias towards any policy is given a priori.