Quantum Reinforcement Learning by Adaptive Non-local Observables

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Hsin-Yi, Chen, Samuel Yen-Chi, Tseng, Huan-Hsin, Yoo, Shinjae
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916864883425280
author Lin, Hsin-Yi
Chen, Samuel Yen-Chi
Tseng, Huan-Hsin
Yoo, Shinjae
author_facet Lin, Hsin-Yi
Chen, Samuel Yen-Chi
Tseng, Huan-Hsin
Yoo, Shinjae
contents Hybrid quantum-classical frameworks leverage quantum computing for machine learning; however, variational quantum circuits (VQCs) are limited by the need for local measurements. We introduce an adaptive non-local observable (ANO) paradigm within VQCs for quantum reinforcement learning (QRL), jointly optimizing circuit parameters and multi-qubit measurements. The ANO-VQC architecture serves as the function approximator in Deep Q-Network (DQN) and Asynchronous Advantage Actor-Critic (A3C) algorithms. On multiple benchmark tasks, ANO-VQC agents outperform baseline VQCs. Ablation studies reveal that adaptive measurements enhance the function space without increasing circuit depth. Our results demonstrate that adaptive multi-qubit observables can enable practical quantum advantages in reinforcement learning.
format Preprint
id arxiv_https___arxiv_org_abs_2507_19629
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Quantum Reinforcement Learning by Adaptive Non-local Observables
Lin, Hsin-Yi
Chen, Samuel Yen-Chi
Tseng, Huan-Hsin
Yoo, Shinjae
Quantum Physics
Artificial Intelligence
Machine Learning
Hybrid quantum-classical frameworks leverage quantum computing for machine learning; however, variational quantum circuits (VQCs) are limited by the need for local measurements. We introduce an adaptive non-local observable (ANO) paradigm within VQCs for quantum reinforcement learning (QRL), jointly optimizing circuit parameters and multi-qubit measurements. The ANO-VQC architecture serves as the function approximator in Deep Q-Network (DQN) and Asynchronous Advantage Actor-Critic (A3C) algorithms. On multiple benchmark tasks, ANO-VQC agents outperform baseline VQCs. Ablation studies reveal that adaptive measurements enhance the function space without increasing circuit depth. Our results demonstrate that adaptive multi-qubit observables can enable practical quantum advantages in reinforcement learning.
title Quantum Reinforcement Learning by Adaptive Non-local Observables
topic Quantum Physics
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2507.19629