Uncovering Instabilities in Variational-Quantum Deep Q-Networks

TL;DR

This paper investigates the instabilities and reproducibility issues of variational quantum deep Q-networks (VQ-DQN) in reinforcement learning, compares simulated and real quantum hardware performance, and provides a robust implementation for future research.

Contribution

It systematically analyzes the stability of VQ-DQN algorithms, compares simulation with real quantum hardware, and offers a reproducible implementation for further studies.

Findings

VQ-DQN algorithms exhibit significant instabilities leading to policy divergence.

Experiments on actual quantum hardware reveal differences from simulated results.

No conclusive evidence that quantum approaches outperform classical methods in RL.

Abstract

Deep Reinforcement Learning (RL) has considerably advanced over the past decade. At the same time, state-of-the-art RL algorithms require a large computational budget in terms of training time to converge. Recent work has started to approach this problem through the lens of quantum computing, which promises theoretical speed-ups for several traditionally hard tasks. In this work, we examine a class of hybrid quantum-classical RL algorithms that we collectively refer to as variational quantum deep Q-networks (VQ-DQN). We show that VQ-DQN approaches are subject to instabilities that cause the learned policy to diverge, study the extent to which this afflicts reproduciblity of established results based on classical simulation, and perform systematic experiments to identify potential explanations for the observed instabilities. Additionally, and in contrast to most existing work on quantum reinforcement learning, we execute RL algorithms on an actual quantum processing unit (an IBM Quantum Device) and investigate differences in behaviour between simulated and physical quantum systems that suffer from implementation deficiencies. Our experiments show that, contrary to opposite claims in the literature, it cannot be conclusively decided if known quantum approaches, even if simulated without physical imperfections, can provide an advantage as compared to classical approaches. Finally, we provide a robust, universal and well-tested implementation of VQ-DQN as a reproducible testbed for future experiments.