This project employs reinforcement learning techniques to explore novel decoding strategies for quantum error correction, particularly focusing on the toric code, to address the challenge of...Show moreThis project employs reinforcement learning techniques to explore novel decoding strategies for quantum error correction, particularly focusing on the toric code, to address the challenge of maintaining stable quantum states for fault-tolerant quantum computing. Two game frameworks are established, including a novel dynamic game framework applicable to the training and measuring of RL agents and potential application in multiagent scenarios. The RL agents use Stable Baselines 3’s Proximal Policy Optimization and show to achieve Minimum Weight Perfect Matching performance on 3 × 3 toric code lattices in both the static and dynamic game frameworks.Show less
