Alerta

基于量子淬火动力学特征增强的分子性质预测方法和装置

Resumen de: CN122930252A

本申请公开一种基于量子淬火动力学特征增强的分子性质预测方法，包括：基于分子的结构化描述符，通过归一化、映射获得量子多体哈密顿量的参数，所述参数用于编码分子的结构特征；基于所述量子多体哈密顿量的参数、预设的初始驱动哈密顿量、调度函数通过量子处理器按时间顺序执行含时演化，在非绝热条件下将系统从初始直积态驱动至纠缠态，基于所述纠缠态通过局域测量提取单体期望值与两体关联函数值并拼接提取数值化的量子增强特征向量；基于量子增强特征向量、对应的分子性质标签通过经典机器学习模型在监督学习框架下训练并优化泛化性能，输出分子电化学性质的预测值。

混合神经网络训练优化方法

Resumen de: CN122927439A

本申请公开了一种混合神经网络训练优化方法，包括：获取当前批次对应的训练状态信息向量，并将经典特征向量与训练状态信息向量输入自适应映射控制子网络，输出编码增益权重向量。根据全局可训练缩放系数、编码增益权重向量及经典特征向量的对应分量，计算每个量子比特对应的旋转角度，并基于旋转角度通过单比特旋转门编码为初始量子态。将初始量子态输入参数化量子电路进行量子态演化。对演化后的量子态进行测量得到量子输出特征向量，并将量子输出特征向量与经典特征向量进行融合后，输入经典预测网络，输出任务预测结果。根据预先构建的联合损失函数执行端到端优化训练，对经典类参数与量子类参数采用对应更新策略进行迭代优化，直至模型收敛。

范畴论驱动的公式库态射复合及多维时间感知装置

Resumen de: CN122925781A

本发明属硬件加速技术领域。针对现有公式推导依赖软件遍历延迟不可控、时间演化数据难以硬件级纳秒检索的问题，提供一种范畴论驱动的公式库态射复合及多维时间感知装置及方法。其核心架构由五个协同模块组成：公式范畴存储单元、态射复合引擎、多维时间感知张量存储单元、时间演化追踪电路、验证接口，通过并行相等比较器实现推导路径的硬件级拼接，并以三维寻址张量存储实现纳秒级时间切片查询。基于同一架构派生三个应用层：实时公式库检索、量子门推导与等价性验证、因果与量子态演化追踪。本发明的算法特征与硬件结构存在特定技术关联，属专利法第2条第2款规定的技术方案，不依赖神经网络或大模型调用。

基于量子增强卷积长短期记忆网络的雷达回波降雨预测方法

Resumen de: CN122922923A

本发明公开了一种基于量子增强卷积长短期记忆网络的雷达回波降雨预测方法，包括如下步骤：1）采集天气雷达反射率数据，对所述数据进行解析、质量控制、坐标转换和归一化处理，并采用滑动窗口构建雷达回波输入序列和预测目标序列；2）将输入序列输入卷积长短期记忆网络，提取雷达回波的时间演变特征和空间结构特征；3）对卷积长短期记忆网络的中间特征图进行空间分块和线性降维，并采用角度嵌入方式将降维后的特征编码至量子比特；4）利用强纠缠量子线路和数据重上传方式对量子态进行变分变换，通过泡利Z测量和泡利ZZ关联测量获得量子测量特征；5）将量子测量特征进行升维和空间恢复，并以残差方式注入卷积长短期记忆网络的特征传递路径；6）采用均方误差和量子参数多样性损失对模型进行联合训练，以递归方式输出未来多帧雷达回波预测结果。本发明能够在有限量子比特条件下实现高维雷达特征的量子编码和关联特征提取，减轻强回波区域的强度衰减、边界平滑和局部结构信息损失，提高雷达回波降雨预测的准确性和训练稳定性。

一种LED量子阱层优化方法及系统

Resumen de: CN122927169A

本发明公开了一种LED量子阱层优化方法及系统，方法包括获取MOCVD多量子阱生长的历史数据集，并对所述历史数据集进行预处理；将预处理后的历史数据输入预构建的深度学习模型，进行监督训练，生成芯片性能预测模型；设定目标芯片性能指标，使用候选MOCVD参数在所述芯片性能预测模型的搜索空间内进行迭代搜索，输出最优MOCVD生长参数组合。本发明强制模型遵守基本物理规律，避免了纯数据驱动模型的反物理预测，将物理先验作为正则化项，显著降低了模型对大数据量的依赖，适合半导体研发阶段数据稀缺的场景。

基于quandle的密码学

Resumen de: EP4819531A1

A method for secure computation includes receiving a specification of an input variable including one or more first elements of a quandle and an arithmetic computation to be performed over the input variable. The input variable is encrypted using a first operation of the quandle performed over the input variable together with an encoding variable including one or more non-unity second elements of the quandle. The arithmetic computation is applied to the encrypted input variable to produce an encrypted output variable. The encrypted output variable is decrypted using the encoding variable and a second operation of the quandle to obtain a result of the arithmetic computation over the variable.

拓扑量子比特阵列的缺陷容忍设计与冗余结构

Resumen de: CN122927572A

本发明公开了一种拓扑量子比特阵列的缺陷容忍设计与冗余结构。所述结构包括主量子比特阵列、冗余量子比特阵列、缺陷检测单元、重映射控制单元和状态标记寄存器。所述方法通过周期性测量所有量子比特的投影权重和投影相位识别故障量子比特；从冗余比特池中选择可用冗余比特；通过相位并行调控模块将故障比特的投影角配置转移到冗余比特；更新耦合拓扑将冗余比特接入计算网络；更新状态标记寄存器。本发明实现了量子比特阵列的低开销缺陷容忍，故障检测周期不大于1微秒，重映射完成时间不大于100纳秒，可广泛应用于大规模拓扑量子计算阵列的制造缺陷容忍和运行故障冗余管理领域。

面向量子最大割问题的热启动硬件迁移分析与控制方法及系统

Resumen de: CN122927569A

本发明公开面向量子最大割问题的热启动硬件迁移分析与控制方法及系统，从最大割问题的基准集中选取若干量子比特图实例并求解，得到半定规划向量；根据半定规划向量进行主成分分析映射，得到球角度，以此构建热启动初始态，并设定混合器，通过混合器进行量子近似优化算法的参数优化，以此生成量子近似优化算法框架；将量子近似优化算法框架迁移至真实硬件场景，在电路硬件搜索组合优化问题的近似解；根据组合优化问题搜索开销，调整组合优化问题搜索参数，以此得到最终近似解。通过直接将连续的SDP向量几何映射到Bloch球坐标，得到一个确定性的热启动初始态及其匹配的混合器，适用于低预算、浅层和低噪声场景，提高硬件设计的效率和稳定性。

基于张量强化学习与逆向解纠缠的量子态制备方法和系统

Resumen de: CN122927563A

本发明公开了一种基于张量强化学习与逆向解纠缠的量子态制备方法和系统，涉及量子计算与人工智能交叉技术领域，本发明将目标量子态转换为矩阵乘积态，通过在当前时间步根据当前矩阵乘积态与可训练张量网络参数生成低维特征向量，并利用系统全局冯·诺伊曼纠缠熵的下降量作为奖励信号，引导张量强化学习智能体选择解纠缠双比特量子门的插入位置，使得解纠缠门序列完全由目标量子态内禀的纠缠分布驱动，自适应地生成最简电路，减少了含噪声中等规模量子硬件上的门噪声积累。本发明使用矩阵乘积态表示量子态，并通过可训练张量网络参数与当前矩阵乘积态生成低维特征向量，将计算复杂度从指数级降为多项式级别。

用于操纵量子比特的通量线

Resumen de: WO2025191211A1

There is provided mitigating Purcell loss at qubit frequency. A flux line for manipulating at least one qubit, comprises: a first segment (1302) for connecting to an input terminal for receiving a control signal; and a second segment (1304, 1604) for coupling the control signal to a Superconducting Quantum Interference Device (SQUID) operatively connected to the at least one qubit, the second segment is arranged consecutive with respect to the first segment (1302), and the first segment and the second segment are configured to have an impedance mismatch at a qubit frequency of the at least one qubit.

校准装置、超导量子计算机和实时校准方法

Resumen de: CN122927571A

本申请公开了一种用于超导量子计算机的校准装置，超导量子计算机至少包括测控系统和超导量子芯片，校准装置包括信号分配器，其输入端与测控系统中测控Z线路的输出端连接，用于将测控Z线路生成的Z控制信号分配为一路Z操控信号和一路参考测量信号。采样计算模块，其输入端与信号分配器的输出端连接，用于根据参考测量信号进行误差分析及校准计算，得到校准数据。主控模块，与采样计算模块及测控Z线路分别连接，用于根据校准数据生成校准指令，并将校准指令下发至测控Z线路，以对Z控制信号进行实时校准。这样，无需中断正常操控实验即可完成漂移检测与校准补偿。

ALGORITHMIC FAULT-TOLERANCE FOR LOW OVERHEAD QUANTUM COMPUTING

Resumen de: AU2025360923A1

Error-corrected quantum computation using transversal gates and correlated decoding is provided. A first and second logical qubit is encoded into physical qubits according to a quantum error correcting code. Based on the quantum error correcting code, a bipartite decoding graph is constructed corresponding to the first and the second logical qubits, the bipartite decoding graph comprising a plurality of detector nodes and a plurality of error nodes, each error node corresponding to an error mechanism. A transversal gate is applied to the first and the second logical qubits. Syndrome measurement of the first and the second logical qubits is performed. For each of the plurality of detector nodes affected by the corresponding error mechanism of one of the plurality of error nodes, an edge is generated on the bipartite decoding graph therebetween. A physical error configuration is determined from the bipartite decoding graph.

CONTINUOUS RELOADING OF ATOMIC QUBITS INCLUDING TWO OPTICAL CONVEYOR BELTS

Resumen de: AU2025280050A1

Devices and methods for reloading atomic qubits into a quantum computer via two optical conveyor belts are provided.

OPTICAL TRANSITION LIGHT SHIFT GATES ON GROUND STATE HYPERFINE QUBITS

Resumen de: WO2026217127A2

An entangling gate is performed on quantum objects by causing mapping manipulation signals to be incident on the quantum objects to transfer a population of a first qubit state to a first-half gate state; causing first-half gate manipulation signals to be incident on the quantum objects to couple the first-half gate state with a first-half selected qubit state to perform a first portion of an entangling operation; causing mapping manipulation signals to be incident on the quantum objects to transfer a population of the first-half gate state to the qubit sub-space and a population of a second qubit state to a second-half gate state; and causing second-half gate manipulation signals to be incident on the quantum objects to couple the second-half gate state with a second-half selected qubit state to perform a second portion of the entangling operation. A leakage detection operation may be performed mid- or post-gate.

QUANTUM BITS BASED ON HYBRID ORGANIC INORGANIC PEROVSKITES

Resumen de: WO2026217109A2

A quantum bit system for use in quantum computing systems. The quantum bit system (100) can include a hybrid organic-inorganic perovskite (HO IP) (110). The quantum bit system can also include an optical pumping source (120) having a fixed circular polarization, wherein the optical pumping source is configured to induce dynamic nuclear spin polarization in the HOIP through electron-nuclear hyperfine interaction to generate an internal magnetic field on one or more electrons that modulates electron spins. The quantum bit system can include a control light source (130) associated with the optical pumping source, wherein the control light source is circularly polarized and modulates the internal magnetic field. The quantum bit system can also include an optical probe (140) associated with the control light source, wherein the optical probe measures changes in the internal magnetic field via detected changes in the electron spins.

SCALABLE TRAINING OF IMPROVED DETECTOR ERROR MODELS FOR DECODERS OF QUANTUM ERROR CORRECTION CODES

Resumen de: WO2026217118A2

One example aspect of the present disclosure is directed to a method for operating a quantum computing system (QCS). The QCS has a set of qubits. The method includes accessing detection event data that includes stabilizers measurements for a set of stabilizers of a quantum error correction (QEC) code implemented, via. the QCS, on the set of qubits. A trained detector error model (DEM) for a decoder of the QEC code is generated. The trained detector DEM is generated by iteratively updating a set of candidate DEMs based on the detection event data and an optimization algorithm that includes an objective. A. quantum algorithm is performed, via. the QCS, based on the QEC code with the decoder configured to use the DEM.

CHIP-BASED QUANTUM PROCESSOR AND ASSOCIATED METHODS

Resumen de: WO2026217130A2

A chip-based quantum processor uses homonuclear molecules with non-zero nuclear spins. Each of these molecules serves as a qubit. To fabricate the quantum processor, a thin film of semiconductor atoms is deposited on a semiconductor substrate to form a protective under layer and over layer. A Group 15 or semiconductor molecular source forms a beam of dimer, triatomic, or tetratomic molecules that is deposited between the under layer and the over layer. Alternatively, fast Group 15 or semiconductor molecular ions are slowed and compressed in an ion funnel, or confined in an ion trap, and then deposited on top of the under layer at low energy. The quantum processor may be annealed to crystalize the deposited under layer and over layer. The deposited molecular qubits may be interrogated with electrodes, optical fields, microwave fields, or a combination thereof.

GENERATIVE LEARNING WITH SHALLOW QUANTUM CIRCUITS

Resumen de: WO2026217123A2

One example aspect of the present disclosure is directed to a method for generative quantum learning. The method includes generating a set of classical-output objects based on a classical generative model and a set of classical-input objects. The classical generative model provides a first mapping that is from the set of classical-input objects to the set of classical-output objects. A set of quantum-input objects is generated based on an encoding schema and the set of classical-input objects. The encoding schema includes a classical-to-quantum encoding process. A set of quantum-output objects is generated based on the encoding schema and the set of classical-output objects. A unitary operator is generated based on a learning algorithm. The unitary operator at least partially provides a second mapping that is from the set of quantum-input objects to the set of quantum-output objects.

QUANTUM HOLOGRAPHIC SIGNATURE WITH QUANTUM FOURIER SIGNATURE TRANSFORMER AND ENCODING HARMONIZER

Resumen de: US20260320592A1

Apparatus and methods for securing a transaction. The apparatus may include a holographic signature pad. The holographic signature pad may include an array of sensors. The sensors may be configured to sense a first quantum tunneling pattern resulting from interaction between a pen and the signature pad. The pad may be configured to create a holographic interference pattern based on the first quantum tunneling pattern. The apparatus may map the digitized interference pattern onto first qubits in a quantum register. The apparatus may be configured to interact with the quantum computer to apply a quantum Fourier transform algorithm to the first qubits to produce a quantum phase-space interference matrix. The apparatus may be configured to form a key-value pair that includes a key corresponding to a quantum encrypted fingerprint matrix and a value corresponding to the quantum phase-space interference matrix.

DIFFERENTIABLE SHOT NOISE STATE VECTOR SIMULATION

Resumen de: US20260320493A1

0000 The present disclosure introduces a differentiable framework for modeling quantum shot noise in state vector simulations used for training parameterized quantum circuits. Conventional approaches rely on multinomial sampling of measurement outcomes, which is inherently discrete and non-differentiable, blocking gradient-based optimization. To address this, the present disclosure replaces multinomial sampling with a continuous approximation using a multivariate Gaussian distribution that captures the statistical properties of shot noise. By applying a reparameterization technique, stochastic noise is expressed as a deterministic function of circuit parameters and auxiliary random variables, enabling gradients to propagate through the noise model. The method integrates noise directly into the quantum state vector, followed by normalization to maintain physical validity. This approach preserves differentiability while incorporating realistic noise effects, allowing efficient end-to-end training using standard optimization methods. The result is a scalable, noise-aware simulation framework that improves the fidelity and practicality of quantum machine learning development.

QUANTUM-CLASSICAL DECOMPOSITION OF THE UNIT COMMITMENT PROBLEM

Resumen de: US20260320464A1

A system and method are provided for solving a mixed-integer optimization program having both binary and continuous variables in a power system. An exemplary method includes: obtaining a formulation of the mixed-integer program (MIP) comprising a set of binary decision variables representing on and off activation states for generation units, and a set of continuous decision variables representing operational levels associated with the generation units; determining a subset of candidate binary variable assignments by applying a quantum approach; for each candidate binary assignment, solving a continuous-variable subproblem on a classical processor to determine corresponding values for continuous decision variables that minimize an objective cost subject to one or more constraints using a classical optimization routine; evaluating a total cost for each combination of binary assignments and corresponding continuous-variable solution; and selecting a combination of binary and continuous decision variable values that yields a solution to the MIP.

MODELS FOR DETERMINING LOGICAL ERROR RATES FOR QUANTUM ERROR CORRECTION CODES OPERATING ON QUBIT ARRAYS WITH NON-UNIFORM PHYSICAL ERROR RATES

Resumen de: WO2026217078A1

One example aspect of the present disclosure is directed to a method for operating a quantum computing system (QCS). The method includes accessing a set of physical error rates. Each physical error rate of the set of physical error rates corresponds to a physical error rate of a qubit of the set of qubits. The set of physical error rates is non-uniform. A predicted logical error rate for a quantum error correction (QEC) code is determined. The QEC code corresponds to the set of qubits. Determining the predicted logical error rate is based on the set of physical error rates and a logical error rate model for the QEC code. The QCS is operated in accordance with the predicted logical error rate for the QEC code.

DOUBLE-SINGLE-CLICK PROTOCOL FOR REMOTE STATE PREPARATION WITH A WEAK COHERENT PULSE SOURCE

Resumen de: WO2026216997A1

The present disclosure presents a double-single-click protocol for use in a quantum network, nodes of such quantum network, and related methods. The quantum network includes client nodes, server nodes, and measurement nodes, wherein at least one client node is operatively coupled via a first quantum communication link with at least one measurement node, and wherein the at least one server node is operatively coupled via a second quantum communication link with at least one measurement node, wherein the at least one server node and the at least one client node are configured to transmit quantum states toward the measurement node, and wherein the at least one server node and the at least one client node are configured to receive heralding signals from the measurement node, wherein the measurement node is configured to transmit the heralding signals upon using a double-single-click protocol.

SINGLE-CLICK PROTOCOL FOR REMOTE STATE PREPARATION WITH A WEAK COHERENT PULSE SOURCE

Resumen de: WO2026216996A1

The present disclosure presents a single-click protocol for use in a quantum network, nodes of such quantum network, and related methods. The quantum network includes client nodes, server nodes, and measurement nodes, wherein at least one client node is operatively coupled via a first quantum communication link with at least one measurement node, and wherein at least one server node is operatively coupled via a second quantum communication link with the at least one measurement node, wherein the at least one server node and the at least one client node are configured to transmit quantum states toward the measurement node for remote state preparation, and wherein the server node and the client node are configured to receive heralding signals from the measurement node, wherein the measurement node is configured to transmit the heralding signals upon using a single-click protocol.

QUANTUM COMPUTER AND CONTROL METHOD THEREFOR

Nº publicación: WO2026214675A1 15/10/2026

Solicitante:

HUAWEI TECH CO LTD [CN]
UNIV OF SCIENCE AND TECHNOLOGY OF CHINA [CN]
\u534E\u4E3A\u6280\u672F\u6709\u9650\u516C\u53F8
\u4E2D\u56FD\u79D1\u5B66\u6280\u672F\u5927\u5B66

Resumen de: WO2026214675A1

Provided in the embodiments of the present application are a quantum computer and method for quantum computing. The quantum computer comprises: a control module, an acousto-optic deflector (AOD) and an ion trap or a neutral atom array. The method comprises: the control module determining, for the ion trap or the neutral atom array using the AOD, the interference of an addressing laser on neighboring qubits of an addressed qubit; moreover, on the basis of the interference, the control module determining a target electrical signal for target interference cancellation laser for canceling the interference on the neighboring qubits; and the AOD acquiring the target interference cancellation laser on the basis of the target electrical signal, wherein the target interference cancellation laser covers the neighboring qubits. In this way, the optical intensity crosstalk of addressing light on the neighboring qubits can be effectively canceled, thereby reducing crosstalk errors in laser addressing.