mmqoc_codesign_apic

Code and datasets for the research paper:

This repository contains the code used in the preprint paper below:

Hardware Co-Designed Optimal Control for Programmable Atomic Quantum Processors via Reinforcement Learning arXiv:2504.11737 [Qian Ding, Dirk Englund]

Project Description

This project focuses on implementing hardware co-designed quantum optimal control (QOC) using Reinforcement Learning (RL) techniques on neutral atom platforms using programmable Photonic Integrated Circuits (PICs). The goal is to demonstrate robust, high-fidelity gate operations considering practical hardware with control imperfections like inter-channel crosstalk and beam leakage.

QOC Optimization Methods

The code is written in JAX and we implement three quantum control optimization algorithms:

1. Classical hybrid optimizer combining Self-Adaptive Differential Evolution (SaDE) and Adam (SADE-Adam)
2. Conventional Proximal policy optimization (PPO) based RL approach
3. End-to-End differentiable RL-based approach

Repository Structure

1. notebooks/ # Jupyter Notebooks for running experiments
2. results/ # code for plotting the results in figures
3. src/ # Source code for local and global control