Learning Explicit Circuit Representations for Quantum States from Local Measurements

Manwen Liao, Yan Zhu, Weitian Zhang, Yuxiang Yang

Keywords: quantum state learning, quantum circuit construction, local measurements, reinforcement learning

Abstract: Characterizing quantum states is essential for advancing many quantum technologies. Recently, deep neural networks have been applied to learn quantum states by generating implicit representations that map them into classical vectors. Despite their success in predicting state properties, these representations remain a black box, lacking insights into strategies for experimental reconstruction. In this work, we aim to open this black box by developing explicit representations of quantum states through the generation of preparation circuits using a reinforcement learning agent with a local fidelity reward function. Relying solely on measurement data from a few neighboring qubits, our agent accurately recovers properties of target states. Specifically, we design a quantum measurement feature aggregation block which is used to extract global features of quantum states from local measurement data. We also provide a theoretical guarantee for the proposed local fidelity reward function. Extensive experiments demonstrate the effectiveness of our approach in learning various quantum states of up to 100 qubits, including those generated by Instantaneous Quantum Polynomial circuits, evolved by Ising Hamiltonians, and many-body ground states. The learned circuit representations can be further applied to Hamiltonian learning as a downstream task utilizing a simple linear model.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Submission Number: 9753