Optimal Hamiltonian recognition of unknown quantum dynamics
Chengkai Zhu, Shuyu He, Yu-Ao Chen, Lei Zhang, Xin Wang · 2024
Read Optimal Hamiltonian recognition of unknown quantum dynamics by Chengkai Zhu, Shuyu He, Yu-Ao Chen, Lei Zhang, Xin Wang (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-13067.txt.
Introduction
Chengkai Zhu and co-authors tackle, in a 2024 paper, the task of naming which rule drives an unknown quantum system. Each rule, a Hamiltonian, sets how the system changes over time. To name it, staff may probe the system a set number of times, then guess. The team frames the task as a blend of two known fields: tests that pick between claims, and the science of fine measures. They work with a single quantum bit whose rule is one of two or three set rules.
Their method uses two signal tools that can run a target curve no matter which rule is true. With math proofs, they show their success rate is the best that can be had, and falls as one over the probe count. Tests on a real quantum chip back the claim. A side task on linked multi-bit systems also works in number tests.
Catalog record
| CREATOR | Chengkai Zhu, Shuyu He, Yu-Ao Chen, Lei Zhang, Xin Wang |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 22 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-13067
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Optimal Hamiltonian recognition of unknown quantum dynamics Chengkai Zhu, Shuyu He, Yu-Ao Chen, Lei Zhang, Xin Wang Abstract Identifying unknown Hamiltonians from their quantum dynamics is a pivotal challenge in quantum technologies and fundamental physics. In this paper, we introduce Hamiltonian recognition, a framework that bridges quantum hypothesis testing and quantum metrology, aiming to identify the Hamiltonian governing quantum dynamics from a known set of Hamiltonians. To identify H H for an unknown qubit quantum evolution exp ( − i H θ ) \exp(-iH\theta) with unknown θ \theta , from two or three orthogonal Hamiltonians, we develop a quantum algorithm for coherent function simulation, built on two quantum signal processing (QSP) structures. It can simultaneously realize a target polynomial based on measurement results regardless of the chosen signal unitary for the QSP. Utilizing semidefinite optimization and group representation theory, we prove that our methods achieve the optimal average success probability, taken over possible Hamiltonians H H and parameters θ \theta , decays as O ( 1 / k ) O(1/k) with k k queries of the unknown unitary transformation. Furthermore, we demonstrate the validity of our protocol on a superconducting quantum processor. This work presents an efficient method to recognize Hamiltonians from limited queries of the dynamics, opening new avenues in composite channel discrimination and quantum metrology. I Introduction The evolution of a quantum state in an isolated system is governed by the Schrödinger equation, with a Hamiltonian H determining the unitary operation U_{H}(\theta)\coloneqq\exp(-iH\theta) . Learning this unitary operation can be approached through quantum process tomography Mohseni2008 ; Gutoski_2014 or by fitting the parameter \theta based on a hypothesized model H , reminiscent of assuming a linear or other model and performing regression in machine learning. In the context of quantum metrology Giovannetti_2004 ; Giovannetti_2006 ; Degen2017 , the hypothesis on H is fixed and estimating \theta for a known Hamiltonian H has been widely studied and has broad applications in various fields Huelga1997 ; Dorner2009 ; Anisimov2010 ; Schnabel_2010 ; Joo2011 ; Genoni2011 ; Thomas2011 ; Barzanjeh_2015 . The ultimate precision in parameter estimation is a central focus of quantum metrology, with the standard quantum limit scaling as 1/k and the Heisenberg limit scaling as 1/k^{2} Kitagawa1993 ; Demkowicz2014 ; Yuan2015 ; Yuan2016 ; Pang_2017 ; Liu_2024 . Moreover, multi-parameter quantum metrology where the Hamiltonian is not fully known is substantially more complex due to the incompatibility Szczykulska_2016 ; Liu_2019 ; Demkowicz_Dobrza_ski_2020 . The celebrated quantum Cramér-Rao bound helstrom1969quantum ; holevo2011probabilistic can typically not be saturated Ragy2016 . However, could the challenges of multi-parameter estimation be circumvented if we have some prior information about the Hamiltonian? For instance, consider the task of estimating parameters \alpha and \beta encoded in a Hamiltonian H=\alpha H_{0}+\beta H_{1} . Suppose we further know there is a possible multiplicative relationship between \alpha and \beta , e.g., either \alpha=a\beta or \alpha=b\beta for some known a,b . Can we efficiently identify which case actually holds and then reduce the problem to single-parameter metrology? Motivated by this question, we introduce the task of Hamiltonian recognition , which is a basic form of Hamiltonian learning. The specific problem we consider is to guess H from several given candidates by querying the unknown evolution U_{H}(\theta) multiple times for arbitrary \theta . The target is to maximize the success probability averaging over all possible H and \theta . Hamiltonian recognition differs from Hamiltonian learning Wang_2017 ; Bairey2019 ; Li2020 ; Che2021 ; wilde2022 ; Yu_2023 ; Huang2023 ; Gu_2024 or certification Wiebe2014 , where the evolution time is typically controllable. In contrast, Hamiltonian recognition aims to develop a time-independent protocol with a single-shot measurement. It can be viewed as a special case of Hamiltonian learning that focuses on identifying the Hamiltonians from a given class rather than learning the complete Hamiltonian itself. The Hamiltonian recognition is closely related to the composite hypothesis testing of quantum channels Bergh2023 , where an unknown channel is promised to be from one of two sets, and the task is to decide which set it belongs to by querying it k times. In Hamiltonian recognition, two (or more) sets of unitary operations are generated by different possible Hamiltonians with varying parameters \theta . When \theta is fixed, the corresponding binary and multiple unitary discrimination have been extensively studied (e.g., Acin2001 ; Ariano2001 ; D_rquot_Ariano_2002 ; Duan2007 ; Laing2009 ; Soeda_2021 ; Bavaresco_2021 ; Bavaresco_2022 ). It was shown that any pair of unitary operations could be perfectly discriminated with a finite number of uses Acin2001 ; Ariano2001 . Such perfect discrimination may not even require entanglement or joint quantum operations Duan2007 . However, little is known about composite unitary discrimination where the goal is to distinguish between two sets of unitary operations, potentially containing an infinite number of elements. The bottleneck arises from the diverse strategies that can be applied to unitary operations, such as parallel, sequential, and indefinite causal order Oreshkov_2012 ; Chiribella2013 ; Bavaresco_2021 , as well as the complex arising from two sets of unitaries. In this work, we develop a protocol for binary recognition of the underlying single-qubit Hamiltonian, given as an unknown oracle \exp(-iH\theta) . This protocol achieves an optimal average error 1/(2k+2) over possible Hamiltonians and parameters, where k is the number of queries to the unknown unitary. Remarkably, our protocol utilizes the technique of quantum signal processing (QSP) Low_2016 ; low_optimal_2016 to analyze protocol resolution errors. The QSP provides a systematic approach to making a quantum system transform under certain polynomial functions yu_power_2022 and is also helpful for analyzing multiple unitary discrimination Rossi2021 . As a byproduct, we reveal that for any positive integer k , there exist two sets, each containing k qubit unitaries, such that many pairs of unitaries, one from each set, are not orthogonal, yet these sets can be perfectly distinguished using only k queries without the need for entanglement. Notably, we demonstrate the optimality of our protocol by using semidefinite programming (SDP). We also show different strategies involving processes with indefinite causal order Oreshkov_2012 ; Chiribella2013 , will not offer any advantage in the recognition of Pauli X and Pauli Z . To validate our results, we implement our binary recognition protocol on a superconducting quantum processor from Tencent and observe the desired error decay. Furthermore, to tackle ternary Hamiltonian recognition, we develop a quantum algorithm, consisting of coherent uses of two QSP circuits, that can realize a target polynomial based on the measurement results regardless of which signal unitary is chosen for the QSP. This protocol achieves an optimal average error of 1/(3k+3) for odd k . II Preliminaries To address the task of Hamiltonian recognition, we begin by introducing quantum channel discrimination Pirandola_2019 ; Wilde_2020 ; Zhuang2020 ; Li2022 ; Salek2022 . Let {\cal L}({\cal H}) be the set of linear operators acting on a finite-dimensional Hilbert space {\cal H} . Quantum channels from an input system I to an output system O are completely positive and trace-preserving maps {\cal N}\mathrel{\mathop{\mathchar 58\relax}}{\cal L}({\cal H}_{I})\rightarrow{\cal L}({\cal H}_{O}) . The Choi operator of a quantum channel is defined as J^{{\cal N}}_{IO}\coloneqq\sum_{j,l=0}^{d_{I}-1}|j\rangle\!\langle l|_{I}\otimes{\cal N}_{I\to O}(|j\rangle\!\langle l|_{I}) . Particularly, the Choi operator of a unitary U is denoted as |U\rangle\hskip-2.27621pt\rangle\hskip-1.13809pt\langle\hskip-2.27621pt\langle U|