Quantum reservoir probing: exploration of quantum many-body physics via computational performance
Over the past decade, the integration of quantum information science with quantum many-body physics has yielded new insights into quantum phenomena. In this talk, we further advance this synergy by focusing on the computational performance of quantum systems, beyond traditional quantum information metrics.
Our starting point is quantum reservoir computing (QRC) [1, 2], a quantum machine learning paradigm in which the natural dynamics of a quantum system (“quantum reservoir”) serve as a feature map for information processing. A key aspect of QRC is that training is limited to the classical post-processing stage, while the quantum reservoir itself remains fixed. Therefore, the computational performance of QRC directly reflects the characteristics of the quantum many-body system acting as the reservoir.
By inverting this relationship, we introduce quantum reservoir probing (QRP) [3,4], a novel framework to explore quantum many-body physics through its computational capabilities. We demonstrate the applicability of QRP by investigating two key phenomena: (i) information propagation, where QRP discerns distinct propagation dynamics reflecting the system’s inherent nature [3], and (ii) quantum phase transitions, where enhanced critical fluctuations leave clear signatures in the computational performance [4]. Our results highlight the versatility of QRP for probing a plethora of exotic quantum many-body phenomena.
[1] K. Fujii and K. Nakajima, Phys. Rev. Applied 8, 024030 (2017). [2] K. Kobayashi, K. Fujii, N. Yamamoto, PRX Quantum 5, 040325 (2024). [3] K. Kobayashi and Y. Motome, arXiv:2308.00898 (to appear in SciPost Phys.). [4] K. Kobayashi and Y. Motome, Nat. Commun. 16, 3871 (2025).