Xinzhao Wang

Computer Science · Peking University计算机科学 · 北京大学

Xinzhao Wang王昕兆王昕兆Xinzhao Wang

Ph.D. candidate · advised by Tongyang Li博士研究生 · 导师：Tongyang Li

I work on quantum algorithms and computational complexity, particularly quantum simulation, with an emphasis on query complexity and resource-efficient circuit implementations.我的研究方向是量子算法与计算复杂性，尤其关注量子模拟，重点研究查询复杂度及节省资源的线路实现。

I was a visiting scholar at UC Berkeley (July 2024–February 2025), hosted by Lin Lin.2024 年 7 月至 2025 年 2 月，我在加州大学伯克利分校访问，合作导师为 Lin Lin。

wangxz@stu.pku.edu.cn Google Scholar CV (PDF)简历（PDF）

Seeking postdoctoral positions in quantum algorithms and related areas starting in 2027.我正在寻找 2027 年开始的量子算法及相关方向博士后职位。

Portrait of Xinzhao Wang

Research研究工作

† Alphabetical author order · * Equal contribution† 作者按字母顺序排列 · * 同等贡献

Quantum Query and Computational Complexity量子查询复杂度与计算复杂性

I enjoy the clean, elegant ideas that arise in quantum query complexity. In particular, I study algorithms and query lower bounds for simulation, ground-state preparation, and estimation. I also study the computational complexity of estimating spectral properties of Hamiltonians.我喜欢量子查询复杂度中涌现的简洁而优美的思想。具体而言，我研究量子模拟、基态制备和估计问题的算法与查询复杂度下界，也研究估计哈密顿量谱性质的计算复杂性。

Time-Dependent Hamiltonian Simulation with Optimal Query Complexity

Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

AI useAI 使用说明

The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

We obtain optimal query complexity for Lipschitz-continuous time-dependent Hamiltonian simulation, matching the bound for time-independent Hamiltonians. We encode the full time-ordered evolution in a transducer and analyze catalyst-omission error using its private block. This analysis guides the choice of LCU coefficients for combining circuits with different transducer reuse counts, canceling leading error terms and yielding factorial error suppression.我们给出了 Lipschitz 连续含时哈密顿量模拟的最优查询复杂度，与不含时情形的界一致。我们将完整的时间有序演化编码进 transducer，并利用其 private block 分析省略催化态所引入的误差。该分析指引了不同 transducer 重用次数线路的 LCU 系数选择，通过消去主导误差项，实现阶乘级的误差衰减。

A Lipschitz-continuous Hamiltonian \(H(t)\) on \([0,T]\) with \(\|H(t)\|\le\alpha\) can be simulated to error \(\varepsilon\) using the optimal number of queries to a time-indexed block encoding:对于定义在 \([0,T]\) 上、满足 \(\|H(t)\|\le\alpha\) 的 Lipschitz 连续哈密顿量 \(H(t)\)，将模拟误差控制在 \(\varepsilon\) 内所需的最优含时间索引块编码查询次数为：

\[ q=O\!\left(\alpha T+ \frac{\log(1/\varepsilon)}{\log\!\left(e+\log(1/\varepsilon)/(\alpha T)\right)} \right). \]

Other papers其他论文

Simulation and ground-state preparation量子模拟与基态制备

  • Query-Optimal and Gate-Efficient Lindbladian Simulation

    Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

    Query-optimal Lindbladian simulation with polylogarithmic additional one- and two-qubit gate overhead per oracle query.实现查询复杂度最优的 Lindbladian 模拟，每次预言机查询的额外单、双量子比特门开销为多对数级。

  • Optimal Query Complexity for Ground-State Preparation

    Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors proposed using purifiers and transducer composition for ground-state preparation. LLMs were used to develop the detailed constructions and proofs, which the authors then verified and revised. Paper statement.作者提出利用 purifier 及 transducer 的组合性质进行基态制备，并借助大语言模型发展具体构造和证明，随后核验并修改相关论证。论文原文说明。

    Optimal expected and worst-case Hamiltonian-query bounds for ground-state preparation, with separate bounds on initial-state preparation queries.确定基态制备在期望和最坏情况下的最优哈密顿量查询复杂度，并单独分析初态制备预言机的调用次数。

Estimation估计问题

  • DQC1-Completeness of Normalized Trace Estimation

    Zhengfeng Ji, Tongyang Li, Changpeng Shao, Xinzhao Wang†, and Yuxin Zhang.

    Relates the approximate degree of a function to the complexity of normalized trace estimation: DQC1-completeness for log-local Hamiltonians and classical query lower bounds for sparse matrices.用函数的多项式近似度数刻画归一化迹估计的复杂性：在 log-local 哈密顿量模型中研究 DQC1 完全性，在稀疏矩阵访问模型中给出经典查询下界。

    FOCS 2026

  • Quantum Multi-Level Estimation of Functionals of Discrete Distributions

    Kean Chen, Minbo Gao, Tongyang Li, Qisheng Wang, and Xinzhao Wang†.

    Uses non-destructive singular-value discrimination to estimate contributions from different probability scales, giving near-optimal Tsallis entropy estimation for fixed non-integer orders above one.用非破坏性奇异值判别区分不同概率尺度，逐层估计其贡献；对于大于 1 的固定非整数阶，得到近最优的 Tsallis 熵估计算法。

    ICALP 2026

  • A Quantum Algorithm Framework for Discrete Probability Distributions with Applications to Rényi Entropy Estimation

    Xinzhao Wang, Shengyu Zhang, and Tongyang Li.

    A framework combining QSVT, variable-time amplitude estimation, and quantum annealing, with applications to Rényi entropy estimation.结合 QSVT、变时间振幅估计与量子退火，构建离散概率分布的量子算法框架，并应用于 Rényi 熵估计。

    IEEE TIT 2024

  • Quantum Non-Identical Mean Estimation

    Jiachen Hu, Tongyang Li, Xinzhao Wang†, Yecheng Xue, Chenyi Zhang, and Han Zhong.

    Studies quantum mean estimation when samples come from different distributions with the same or nearby means. Algorithms and lower bounds depend on how often each sampling oracle can be queried.研究样本来自具有相同或相近均值的不同分布时的量子均值估计，并根据各采样预言机允许重复调用的次数，给出算法与复杂度下界。

    TQC 2024

End-to-End Resource-Efficient Quantum Algorithms端到端资源高效的量子算法

Query complexity alone can leave substantial costs in oracle implementation unaccounted for. I am interested in designing quantum algorithms with end-to-end resource efficiency, from input access to circuit implementation. My work includes efficient implementations of common input oracles and algorithms for quantum simulation and QSVT that combine simple circuits with favorable asymptotic resource costs.仅考察查询复杂度，可能忽略实现预言机本身所需的大量资源。我希望从输入访问到线路实现，贯通量子算法的整个实现链条，设计端到端资源高效的算法。我的研究包括常用输入预言机的高效实现，以及兼具简洁线路与良好渐近资源复杂度的量子模拟和 QSVT 算法。

Product-Formula Methods and Error Mitigation乘积公式方法与误差缓解

For quantum simulation, product formulas use simple circuits and few ancillas, and their commutator-based error bounds exploit locality. I develop compensation and extrapolation methods for Hamiltonian and Lindbladian simulation that improve precision dependence while preserving these advantages, and study how to mitigate physical noise together with Trotter error.在量子模拟中，乘积公式的线路结构简单、辅助量子比特需求少，其基于对易子的误差界还能利用局域性。我发展用于哈密顿量与 Lindbladian 模拟的补偿和外推方法，在保持这些优势的同时改善精度依赖，并研究如何同时缓解物理噪声与 Trotter 误差。

Lindbladian Simulation with Commutator Bounds

Xinzhao Wang*, Shuo Zhou*, Xiaoyang Wang, Yi-Cong Zheng, Shengyu Zhang, and Tongyang Li.

We derive commutator-based Trotter error bounds for Lindbladian simulation, yielding \(O(\sqrt N)\) system-size scaling in the number of Trotter steps for local \(N\)-site systems. For observable estimation, Richardson extrapolation preserves this commutator scaling while making the number of steps per circuit polylogarithmic in inverse precision. To control the extrapolation remainder, we establish a finite-order BCH truncation bound in terms of doubly right-nested commutators, without assuming convergence of the full BCH series.我们建立了 Lindbladian 模拟中基于对易子的 Trotter 误差界，使局域 \(N\) 格点系统的 Trotter 步数具有 \(O(\sqrt N)\) 的系统规模依赖。在可观测量估计中，Richardson 外推保留了这一对易子标度，并使每条线路的步数仅以多对数方式依赖于精度的倒数。为控制外推余项，我们建立了基于双重右嵌套对易子的有限阶 BCH 截断界，而无须假设完整 BCH 级数收敛。

For Lindbladian generators \(\mathcal L_1,\ldots,\mathcal L_M\), let \(\Phi_q\) denote the degree-\(q\) term in the formal BCH logarithm of \(e^{\mathcal L_M}\cdots e^{\mathcal L_1}\), and let \(\alpha_{\mathrm{comm},q_0}\) be the associated weighted sum of diamond norms of doubly right-nested commutators. When \(\alpha_{\mathrm{comm},q_0}\le1\), the order-\(q_0\) truncation satisfies对于 Lindbladian 生成元 \(\mathcal L_1,\ldots,\mathcal L_M\)，令 \(\Phi_q\) 为 \(e^{\mathcal L_M}\cdots e^{\mathcal L_1}\) 的形式 BCH 对数中 \(q\) 次项，并令 \(\alpha_{\mathrm{comm},q_0}\) 为相应双重右嵌套对易子的 diamond 范数加权和。当 \(\alpha_{\mathrm{comm},q_0}\le1\) 时，\(q_0\) 阶截断满足

\[ \left\|\exp\!\left(\sum_{q=1}^{q_0}\Phi_q\right) -e^{\mathcal L_M}\cdots e^{\mathcal L_1}\right\|_{\diamond} \le e\,\alpha_{\mathrm{comm},q_0}. \]

These commutators have the form \([C_1,[C_2,\ldots,C_d]]\), where \(C_r=[\mathcal L_{j_{r,1}},[\ldots,[\mathcal L_{j_{r,q_r-1}},\mathcal L_{j_{r,q_r}}]]]\) and \(q_r\le q_0\). For local systems, this restriction controls factorial growth in the commutator bounds, allowing us to bound the truncation error without assuming convergence of the full BCH series.这里的对易子形如 \([C_1,[C_2,\ldots,C_d]]\)，其中 \(C_r=[\mathcal L_{j_{r,1}},[\ldots,[\mathcal L_{j_{r,q_r-1}},\mathcal L_{j_{r,q_r}}]]]\)，且 \(q_r\le q_0\)。对于局域系统，这一限制控制了对易子界中的阶乘增长，使我们能够在不假设完整 BCH 级数收敛的情况下控制截断误差。

Other papers其他论文
  • Trotter Error Compensation with Polylogarithmic Precision and Nested-Commutator Scaling without Ancillas

    Xinzhao Wang, Shuo Zhou, Ziruo Wang, Pei Zeng, Jinzhao Sun, Qi Zhao, Tom Gur, and Tongyang Li.

    For local Hamiltonians, high-order channel compensation improves the evolution-time and system-size scaling of a \(K\)-th-order product formula to that of order \(2K+1\), with polylogarithmic precision dependence per circuit and no ancillas.对于局域哈密顿量，高阶信道补偿使 \(K\) 阶乘积公式在演化时间和系统规模上的复杂度标度达到 \(2K+1\) 阶公式的水平；无须辅助量子比特，且每条线路对精度倒数仅有多对数依赖。

  • Joint Mitigation of Algorithmic and Physical Errors in Noisy Hamiltonian Simulation

    Shuo Zhou*, Xinzhao Wang*, Ruiqi Zhang*, Xiaoyang Wang, Yi-Cong Zheng, Tongyang Li, and Shengyu Zhang.

    Couples physical noise strength to the Trotter step size so that Richardson extrapolation cancels leading contributions from both Trotter error and physical noise along a one-dimensional path.将物理噪声强度与 Trotter 步长关联起来，通过一维 Richardson 外推同时消去 Trotter 误差和物理噪声的主导项。

  • Quantum Singular Value Transformation without Block Encodings

    Shantanav Chakraborty, Soumyabrata Hazra, Tongyang Li, Changpeng Shao, Xinzhao Wang†, and Yuxin Zhang.

    Implements QSVT with one ancilla and no multi-qubit controlled gates by approximating the Hamiltonian evolutions in generalized quantum signal processing with product formulas. Richardson extrapolation improves precision dependence while retaining near-optimal, commutator-sensitive gate complexity.用乘积公式近似广义量子信号处理中的哈密顿量演化，仅用一个辅助量子比特、无须多量子比特受控门即可实现 QSVT。Richardson 外推进一步改善精度依赖，并保留近最优且依赖于对易子的量子门复杂度。

Quantum Data Access and Circuit Implementation量子数据访问与线路实现

I use sparsity to reduce the T-gate cost of QROM, state preparation, and block encoding, and develop circuit implementations of Hamiltonian simulation and QSVT that use fewer gates and ancillas.我利用稀疏性降低 QROM、量子态制备和块编码的 \(T\) 门成本，并研究节省量子门与辅助量子比特的哈密顿量模拟和 QSVT 线路实现。

Optimal T Counts under Sparsity: from QROM to State Preparation and Block Encoding

Tongyang Li, Fengning Ou, Xinzhao Wang†, Penghui Yao, Pei Yuan, and Shengyu Zhang.

AI useAI 使用说明

In the second version, GPT-6 Astra helped develop approximate sparse QROM and extend the matching bounds to the near-dense regime. The authors verified and revised these arguments. Paper statement.第二版中，GPT-6 Astra 协助发展了近似稀疏 QROM 的构造，并将匹配界扩展至近稠密参数区间；相关论证由作者核验并修改。论文原文说明。

We prove matching \(T\)-count upper and lower bounds for sparse QROM, state preparation, and block encoding. Our upper bounds build on a multilevel hashing scheme for sparse QROM. The lower bounds hold even with adaptive measurements and classical control.我们证明了稀疏 QROM、量子态制备和块编码相匹配的 \(T\) 门数量上下界。上界构造基于稀疏 QROM 的多层哈希方案。即使允许自适应测量与经典控制，下界依然成立。

The worst-case number of \(T\) gates needed to prepare an \(n\)-qubit state with at most \(s\) nonzero computational-basis amplitudes to trace-distance error \(\varepsilon\) is将一个在计算基下至多具有 \(s\) 个非零振幅的 \(n\) 量子比特态制备至迹距离误差 \(\varepsilon\)，最坏情况下所需的 \(T\) 门数量为

\[ \Theta\!\left( \min\!\left\{s,\sqrt{s\log\frac{2^{n+1}}{s}}\right\} +\sqrt{s\log\frac1\varepsilon}+\log\frac1\varepsilon \right). \]

For a \(2^n\!\times 2^n\) matrix with entries bounded in magnitude by one and at most \(s\) nonzeros per row and column, the worst-case number of \(T\) gates for block encoding with normalization \(\Theta(s)\) and operator-norm error \(\varepsilon_{\mathrm{BE}}\) is对于元素绝对值至多为 1、每行每列至多有 \(s\) 个非零元素的 \(2^n\!\times 2^n\) 矩阵，实现归一化因子为 \(\Theta(s)\)、算子范数误差为 \(\varepsilon_{\mathrm{BE}}\) 的块编码，最坏情况下所需的 \(T\) 门数量为

\[ \Theta\!\left( \sqrt{2^ns\!\left(n+\log\frac1{\varepsilon_{\mathrm{BE}}}\right)} +\log\frac1{\varepsilon_{\mathrm{BE}}} \right). \]
Other papers其他论文
  • Gate-Efficient Implementation of Query-Optimal Time-Dependent Simulation

    Boyang Chen, Minbo Gao, Zhengfeng Ji, Tongyang Li, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

    An exact dyadic factorization of the ordered update product in the transducer’s non-oracle part reduces gate overhead while retaining optimal query complexity.对 transducer 中非预言机部分的有序更新算子乘积作精确的二分递归分解，在保持最优查询复杂度的同时降低额外量子门开销。

  • Quantum Singular Value Transformation without Block Encodings

    Shantanav Chakraborty, Soumyabrata Hazra, Tongyang Li, Changpeng Shao, Xinzhao Wang†, and Yuxin Zhang.

    Also listed under Product-Formula Methods. Its one-ancilla implementation has near-optimal, commutator-sensitive gate complexity.亦列于乘积公式方法。该方法仅需一个辅助量子比特，并具有近最优且依赖于对易子的量子门复杂度。

  • Randomized Quantum Singular Value Transformation

    Xinzhao Wang*, Yuxin Zhang*, Soumyabrata Hazra, Tongyang Li, Changpeng Shao, and Shantanav Chakraborty.

    Two randomized constructions avoid explicit block encodings: importance sampling directly randomizes QSVT, while qDRIFT approximates Hamiltonian evolutions in generalized quantum signal processing. Gate complexity is independent of the number of Hamiltonian terms and quadratic in the polynomial degree.两种随机化构造分别用重要性采样直接随机化 QSVT，或在广义量子信号处理中用 qDRIFT 近似哈密顿量演化，从而避免显式块编码。量子门复杂度不依赖哈密顿量项数，并对多项式次数呈二次依赖。

Online Learning and Equilibrium Computation在线学习与均衡计算

Matrix multiplicative weights connects online prediction with equilibrium computation. I study how its guarantees can adapt to the structure of a learning instance or to changing quantum states, and develop quantum algorithms for correlated equilibria in general-sum games.矩阵乘法权重更新将在线预测与均衡计算联系起来。我研究如何让学习保证适应具体实例的结构或随时间变化的量子态，并发展求解一般和博弈中相关均衡的量子算法。

Instance-Optimal Matrix Multiplicative Weight Update and Its Quantum Applications

Weiyuan Gong, Tongyang Li, Xinzhao Wang†, and Zhiyu Zhang.

We obtain instance-optimal regret for matrix learning from expert advice at the computational cost of standard matrix multiplicative weights. A one-sided Jensen trace inequality extends the potential-based analysis of matrix learning beyond the exponential potential used by standard MMWU.我们在保持标准矩阵乘法权重更新计算成本的同时，为矩阵专家建议学习给出了实例最优的遗憾界。单侧 Jensen 迹不等式将矩阵学习的势函数分析推广到标准 MMWU 所用的指数势函数之外。

If a twice-differentiable potential \(\Phi\) satisfies \(\Phi''(x)=\int_{\mathbb R}\mu(t)e^{-tx}\,\mathrm dt\) for all real \(x\) and some nonnegative \(\mu\), then any Hermitian matrices \(S,G\) with \(\|G\|_{\mathrm{op}}\le1\) satisfy the one-sided Jensen trace inequality若二阶可微势函数 \(\Phi\) 对任意实数 \(x\) 满足 \(\Phi''(x)=\int_{\mathbb R}\mu(t)e^{-tx}\,\mathrm dt\)，其中 \(\mu\) 非负，则对任意满足 \(\|G\|_{\mathrm{op}}\le1\) 的 Hermitian 矩阵 \(S,G\)，有如下单侧 Jensen 迹不等式

\[ \operatorname{Tr}\Phi(S+G) \le\frac12\operatorname{Tr}\!\left[(I+G)\Phi(S+I)+(I-G)\Phi(S-I)\right]. \]

Other papers其他论文

Publications论文发表

† Alphabetical author order. * Equal contribution.† 作者按字母顺序排列。* 同等贡献。

Published or accepted已发表或已接收

  1. 2026
    DQC1-completeness of normalized trace estimation for functions of log-local Hamiltonians

    Zhengfeng Ji, Tongyang Li, Changpeng Shao, Xinzhao Wang†, and Yuxin Zhang.

    FOCS 2026, to appear.

  2. 2026
    Quantum Multi-Level Estimation of Functionals of Discrete Distributions

    Kean Chen, Minbo Gao, Tongyang Li, Qisheng Wang, and Xinzhao Wang†.

    ICALP 2026, LIPIcs 374, 58:1–58:23.

  3. 2025
  4. 2025
    Efficient classical sampling from Gaussian boson sampling distributions on unweighted graphs

    Yexin Zhang*, Shuo Zhou*, Xinzhao Wang*, Ziruo Wang, Ziyi Yang, Rui Yang, Yecheng Xue, and Tongyang Li.

    Nature Communications 16, 9335.

  5. 2024
    Adaptive online learning of quantum states

    Xinyi Chen, Elad Hazan, Tongyang Li, Zhou Lu, Xinzhao Wang†, and Rui Yang.

    Quantum 8, 1471.

  6. 2024
    A quantum algorithm framework for discrete probability distributions with applications to Rényi entropy estimation

    Xinzhao Wang, Shengyu Zhang, and Tongyang Li.

    IEEE Transactions on Information Theory 70(5), 3399–3426.

  7. 2024
    Quantum non-identical mean estimation: Efficient algorithms and fundamental limits

    Jiachen Hu, Tongyang Li, Xinzhao Wang†, Yecheng Xue, Chenyi Zhang, and Han Zhong.

    TQC 2024, 9:1–9:21.

Preprints预印本 12 papers
  1. 2026
    Optimal Query Complexity for Ground-State Preparation

    Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors proposed using purifiers and transducer composition for ground-state preparation. LLMs were used to develop the detailed constructions and proofs, which the authors then verified and revised. Paper statement.作者提出利用 purifier 及 transducer 的组合性质进行基态制备，并借助大语言模型发展具体构造和证明，随后核验并修改相关论证。论文原文说明。

    arXiv:2609.35668.

  2. 2026
    Query-Optimal and Gate-Efficient Lindbladian Simulation

    Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

    arXiv:2609.18757.

  3. 2026
    Joint Mitigation of Algorithmic and Physical Errors in Noisy Hamiltonian Simulation

    Shuo Zhou*, Xinzhao Wang*, Ruiqi Zhang*, Xiaoyang Wang, Yi-Cong Zheng, Tongyang Li, and Shengyu Zhang.

    arXiv:2609.11508.

  4. 2026
    Gate-Efficient Implementation of the Query-Optimal Time-Dependent Hamiltonian Simulation

    Boyang Chen, Minbo Gao, Zhengfeng Ji, Tongyang Li, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

    arXiv:2608.30629.

  5. 2026
    Time-Dependent Hamiltonian Simulation with Optimal Query Complexity

    Boyang Chen, Minbo Gao, Xinzhao Wang†, and Shuo Zhou.

    AI useAI 使用说明

    The authors initiated the project. LLMs generated the core ideas and proof strategies, which the authors refined and verified. Paper statement.项目由作者发起。核心思路与证明策略由大语言模型生成，再由作者完善并核验。论文原文说明。

    arXiv:2608.06094.

  6. 2026
    Optimal T Counts under Sparsity: from QROM to State Preparation and Block Encoding

    Tongyang Li, Fengning Ou, Xinzhao Wang†, Penghui Yao, Pei Yuan, and Shengyu Zhang.

    AI useAI 使用说明

    In the second version, GPT-6 Astra helped develop approximate sparse QROM and extend the matching bounds to the near-dense regime. The authors verified and revised these arguments. Paper statement.第二版中，GPT-6 Astra 协助发展了近似稀疏 QROM 的构造，并将匹配界扩展至近稠密参数区间；相关论证由作者核验并修改。论文原文说明。

    arXiv:2607.28260.

  7. 2026
    Trotter error compensation with polylogarithmic precision and nested-commutator scaling without ancillas

    Xinzhao Wang, Shuo Zhou, Ziruo Wang, Pei Zeng, Jinzhao Sun, Qi Zhao, Tom Gur, and Tongyang Li.

    arXiv:2607.11856.

  8. 2026
    Lindbladian Simulation with Commutator Bounds

    Xinzhao Wang*, Shuo Zhou*, Xiaoyang Wang, Yi-Cong Zheng, Shengyu Zhang, and Tongyang Li.

    arXiv:2603.28602.

  9. 2025
    Randomized Quantum Singular Value Transformation

    Xinzhao Wang*, Yuxin Zhang*, Soumyabrata Hazra, Tongyang Li, Changpeng Shao, and Shantanav Chakraborty.

    arXiv:2510.06851.

  10. 2025
    Instance-Optimal Matrix Multiplicative Weight Update and Its Quantum Applications

    Weiyuan Gong, Tongyang Li, Xinzhao Wang†, and Zhiyu Zhang.

    arXiv:2509.08911.

  11. 2025
    Quantum singular value transformation without block encodings: Near-optimal complexity with minimal ancilla

    Shantanav Chakraborty, Soumyabrata Hazra, Tongyang Li, Changpeng Shao, Xinzhao Wang†, and Yuxin Zhang.

    arXiv:2504.02385.

  12. 2026
    Quantum-classical crossover in fault-tolerant quantum dynamics simulation

    Jinzhao Sun*, Bozhen Zhou*, Jue Xu*, et al. (31 authors)

    Show full author list完整作者名单

    Jinzhao Sun*, Bozhen Zhou*, Jue Xu*, Yuan Yao*, Zhenyu Du*, Zixu Zhang*, Yuntian Gu*, Junxiang Huang, Shuo Zhou, Ziruo Wang, Alexander Yosifov, Wenzheng Dong, Yiming Huang, Daniel Serrano, Xinzhao Wang , Tianfeng Feng, Shreyas Sadugol, Wenjun Yu, Zhou You, Dayue Qin, Xiao-Ming Zhang, Yantao Wu, Aditya Iyer, You Zhou, Tongyang Li, Ying Li, Xiongfeng Ma, Qi Zhao, Pei Zeng, Pan Zhang, and Xiao Yuan.

    arXiv:2607.16116.

Selected talks部分报告

  1. Quantum Multi-Level Estimation of Functionals of Discrete Distributions

    ICALP 2026, Royal Holloway, University of London

  2. Quantum Singular Value Transformation without Block Encodings

    Quantum Computing Seminar, University of Cambridge

  3. Quantum Singular Value Transformation without Block Encodings

    QCTiP 2026, Mathematical Institute, University of Oxford