81 results about "Quantum neural network" patented technology

The invention provides a signal decoding method based on a PSI5 interface, and relates to the technical field of signal decoding, and the decoding method comprises the following steps: processing and adjusting a received PSI5 signal to obtain an accurate identification signal; and extracting a clock signal from the PSI5 signal, designing a clock compensation algorithm according to the quantum neural network, and dynamically adjusting the clock signal to obtain clock synchronization data. And determining a data frame boundary, and decoding the accurate identification signal to obtain original data. And performing comparative analysis on the original data and the PSI5 signal to obtain comparative difference data, judging whether the comparative difference data accords with a preset error type or not, and performing corresponding processing. According to the invention, fractional calculus processing is carried out on the received signal, then signal distortion is corrected by using a channel equalization method, and the quantum neural network is constructed to dynamically adjust the clock signal, so that clock synchronization is realized, a reliable time reference is provided for accurate processing of the signal, and the accuracy of signal decoding is improved.

The invention provides a multi-mode environment cooperative regulation and control system for poultry egg preservation. The system comprises a sensor network and a server, wherein the sensor network is used for collecting poultry egg data; the server is used for constructing a digital twinborn model according to the poultry egg data; perceptual data output by the digital twin model are input into a quantum neural network model, and a predicted fresh-keeping result and an optimal environment regulation and control parameter combination are obtained; according to the optimal environment regulation and control parameter combination, driving a regulation and control actuator group and feeding back actually executed regulation and control parameter values to the digital twinborn model; comparing the difference between the actual fresh-keeping effect reflected by the digital twinborn model and the predicted fresh-keeping result, and optimizing a mapping strategy from the current poultry egg state to the optimal regulation and control parameter by taking the fresh-keeping period extension rate as a reward function; the change condition of biochemical indexes is monitored through a digital twinborn model, and a monitoring result serves as a feedback signal to be input into the reinforcement learning module for strategy adjustment. The fresh-keeping period can be prolonged, intelligent prediction is achieved, quality traceability is guaranteed, and the energy consumption cost is reduced.

The invention discloses a machine room operation and maintenance management automatic inspection system based on intellectualization, and belongs to the technical field of intellectualization. Comprising a multi-source data fusion processing module, a hidden fault probability distribution module, a fault root cause analysis module, a dynamic adaptive baseline generation module, a multivariate strategy self-healing decision module, a cooperative control execution module and a feedback optimization module. According to the method, a full-link intelligent system of multi-source data fusion, quantum neural network prediction, graph neural network traceability, dynamic baseline generation, reinforcement learning decision, transactional scheduling execution and closed-loop optimization is constructed, the hidden fault recognition capability is improved through quantum classical hybrid calculation, and accurate root cause positioning is realized in combination with a causal atlas; and an optimal repair scheme is generated based on a dynamic baseline and a reinforcement learning strategy library, and finally, the reliability of cross-system operation is ensured through transactional scheduling, so that the fault prediction accuracy is remarkably improved.

The invention provides a Q / R dual-adaptive hybrid quantum filtering method for an OTFS (Over the The Over the File System) communication and inductance integrated system, and the method comprises the following steps: constructing an OTFS communication and inductance integrated signal model, and obtaining a target measurement vector; establishing a state vector for describing a target motion state, and constructing a state transition equation and a measurement equation; the method comprises the following steps: constructing a Q / R mixed quantum adaptive adjustment module based on dual-channel features to construct a dual-channel feature input vector, generating a dual-path adjustment factor through a mixed quantum neural network, and constructing an adaptive Q matrix and an adaptive R matrix; and performing adaptive state prediction and updating at each time step by using the adaptive Q and R matrixes constructed in the step 3 to realize tracking of the motion state of the target. According to the method, the hybrid quantum neural network is introduced for driving, collaborative optimization of filter parameters is achieved, and therefore the precision, robustness and response speed of target tracking are remarkably improved in a complex dynamic noise environment.

The invention discloses a high-precision quantum sensing environment physical quantity real-time monitoring system and method, and belongs to the technical field of environment monitoring. The system comprises a quantum sensing front end, a signal excitation and regulation module, a quantum classical hybrid computing engine, an edge cloud collaborative transmission module and a terminal application module. The quantum sensing front end adopts a quantum entanglement sensing array integrated by NV color center diamond and a rubidium atom gas chamber to synchronously sense a magnetic field, temperature, pressure and vibration; the signal excitation and regulation module optimizes regulation parameters and inhibits noise through a quantum optimal control algorithm; a quantum-classical hybrid calculation engine is combined with a quantum neural network and classical multi-algorithm collaboration to realize efficient resolving and data fusion of quantum state features. According to the invention, the monitoring precision, the real-time response speed and the anti-interference capability are obviously improved, synchronous monitoring of multiple physical quantities is supported, and the system is suitable for various scenes such as industry, scientific research and deep space exploration, and has wide application value.

The invention discloses an anti-quantum computing attack data encryption method based on a deep quantum neural network, which is applied to the technical field of data processing, and comprises the following steps: obtaining quantum associated data and security constraint parameters in the fields of government affairs and finance, and completing classical data quantum bit coding, high-dimensional feature extraction and noise filtering through quantization preprocessing; generating high-dimensional quantum state data adaptive to the quantum neural network; a deep quantum neural network encryption architecture is constructed, a dynamic key pool is generated based on quantum superposition-entanglement characteristics, and a layered encryption scheme is formed in combination with a weight sub-update mechanism; encryption parameters are optimized through a quantum neural network reinforcement learning framework, robustness is improved, and a self-learning optimization strategy is generated; an evolution trajectory is monitored by means of a quantum neural network to construct an encryption execution closed loop, and finally comprehensive evaluation information containing indexes such as a training convergence rate and key efficiency is generated through a quantum gate inverse operation learning model and a Hash verification neural network, so that the quantum attack resistance is enhanced.

Embodiments of the present application provide a quantum neural network classifier training method and device, electronic equipment and medium. The scheme is as follows: obtaining a training data set and a to-be-trained classifier; for each training sample data, classifying the training sample data by using the to-be-trained classifier to obtain a first predicted label; calculating a first loss value of the to-be-trained classifier according to a sample label corresponding to each training sample data and the first predicted label; when the to-be-trained classifier has not converged, adjusting the classifier parameters based on the first loss value, and returning to execute the step of classifying each training sample data by using the to-be-trained classifier to obtain the first predicted label corresponding to the training sample data until the to-be-trained classifier converges at the current time. Through the technical scheme provided by the embodiments of the present application, the optimization of the quantum neural network classifier is realized, and the classification accuracy and attack resistance of the quantum neural network classifier are improved.

The invention discloses a health state monitoring method and system based on big data multi-dimensional evaluation, and relates to the cross technical field of health monitoring, big data processing and quantum machine learning. Through multi-source health data collection and in combination with quantization coding and high-dimensional data dimension reduction technologies, core low-dimensional feature vectors are effectively extracted, further, nonlinear correlation features in the feature vectors are deeply mined through a quantum neural network model, comprehensive and accurate evaluation of the health state of the user is achieved, and the user health state evaluation accuracy is improved. The multi-dimensional and deep evaluation mode can reflect the health state of the user more accurately, and provides a more reliable basis for subsequent early warning and intervention.

The invention discloses an Ansatz self-adaptive intelligent design optimization method based on hardware awareness, and relates to the field of quantum machine learning, and the method comprises the following specific steps: obtaining a preset number of target task data sets which are labeled with labels, dividing the target task data sets into K groups, and further based on a load balancing strategy, obtaining a target task data set; dividing a preset number of quantum bits to form quantum circuit partitions; constructing a full binary merging tree, obtaining quantum bit information corresponding to each node, constructing an Ansatz training line in combination with a preset Ansatz structure, and further optimizing the Ansatz training line based on the target task data set; and deploying the obtained optimal Ansatz line into the target quantum neural network to realize classification of the target task. According to the method, the Ansatz line architecture is designed through dynamic structure design and task awareness optimization, the problem that hardware efficient Ansatz is limited by a fixed architecture is avoided, and the convergence speed and the training stability are remarkably improved.

The invention provides an original real-time electricity price prediction method based on a quantum complex neural network and Hilbert-Huang Transform (HHT), and the real-time electricity price prediction method based on the quantum complex neural network and the Hilbert-Huang Transform (HHT). Aiming at the non-stationarity of electricity price data on the historical level, firstly, the time internal correlation of each feature channel is extracted through an HHT time sequence analysis method, and complex multi-dimensional time sequence prediction is simplified into a simple regression task, so that a tedious time sequence modeling process is avoided; for the non-linear problem of electricity price data, a quantum neural network is used for capturing the coupling relation between different factors, and the calculation speed and the model efficiency are further improved by means of the parallelism of quantum calculation. Based on the two advantages, the method can realize accurate real-time electricity price prediction.

The application relates to a method, device and equipment for anti-deepfake speech based on quantum resonance peak disturbance, wherein the method comprises the following steps: reading an original speech signal, extracting a fundamental frequency and a formant frequency of the original speech signal; creating a parameterized quantum circuit to generate quantum noise; adding the quantum noise to the extracted fundamental frequency and formant frequency to obtain a disturbed signal; and preprocessing the disturbed signal to obtain a reconstructed speech signal. The optimized quantum noise generated by the quantum neural network can effectively interfere with the learning and generation of speech features by a deepfake model. Through the design of a loss function and the adjustment of an optimization algorithm, the speech after adding the noise is ensured to have the smallest difference in hearing from the original speech.

A method for modifying a quantum neural network, comprising: identifying via computer means information that needs to be erased from a trained quantum neural network; localizing via the computer means the identified information in the trained quantum neural network; erasing via the computer means the identified information from the trained quantum neural network without erasing from the trained quantum neural network other information that needs not to be erased; compressing via the computer means the quantum neural network which results when the identified information has been erased. Also, a system.

The application discloses a quantum neural network hybrid system based on a memristor Hopfield and a training method thereof, and aims to solve the problems of low storage precision of complex quantum weights, poor quantum-classical collaboration and high energy consumption of the quantum neural network. The system comprises a quantum computing layer, a classical interface layer and a memristor Hopfield layer, and constitutes a closed loop through error feedback. The quantum computing layer extracts quantum entanglement features by using a quantum bit and a quantum convolution layer, and outputs complex quantum weights; the classical interface layer converts the complex quantum weights into memristor conductance values through polar coordinate mapping; and the memristor Hopfield core layer adopts a double-end complex memristor array to complete storage and associative memory operation. The training method optimizes quantum gate parameters and memristor conductance values through closed loop feedback collaboration. The application can reduce precision loss, realize stable and low-energy consumption storage, and is suitable for quantum key storage, medical image diagnosis and other scenes.

The invention discloses an unmanned aerial vehicle network slice resource optimization method based on a hybrid quantum neural network, which solves the optimization bottleneck problem under complex constraints and high-dimensional space in the prior art, reduces the service time delay and improves the resource utilization rate. The method comprises the following steps: constructing an unmanned aerial vehicle network system model, establishing an optimization problem based on the unmanned aerial vehicle network system model by taking minimization of service delay as a target, and decoupling the optimization problem into a communication delay quantile minimization sub-problem and a processing delay minimization sub-problem; sequentially solving the communication delay quantile minimization sub-problem and the processing delay minimization sub-problem by adopting a multi-layer intelligent agent architecture to obtain an association indication variable set, a power distribution variable set and a function segmentation decision variable set; and realizing joint optimization of unmanned aerial vehicle network slice resources according to the association indication variable set, the power distribution variable set and the function segmentation decision variable set.

The invention discloses a memory management method for a quantum neural network system and a related device, and belongs to the technical field of quantum computing, and the method comprises the following steps: determining a memory pool with different memory spaces as a specified memory pool according to the characteristics of a quantum neural network; in response to an idle buffer block which meets the memory space applied by the user and is used for storing the minimum memory value of data and is found in the specified memory pool according to the memory application request of the user, marking the idle buffer block as used; wherein the specified memory pool comprises a plurality of buffer blocks with different memory values; dividing the free buffer block into a first buffer block and a second buffer block according to the remaining memory space after the data is stored; and storing the data in the first buffer block, and obtaining a first address of the first buffer block as a memory application address. By introducing the memory segmentation mechanism of the free buffer blocks of the memory space, the generation of memory fragments is reduced, so that the use efficiency of the whole memory is improved.

The application discloses a bank fixed deposit target customer classification method based on a quantum neural network, relates to the technical field of classification based on quantum computing, and comprises the following steps: preprocessing a bank fixed deposit data set, obtaining feature information of bank fixed deposit customers, and loading the feature information into a quantum state; processing the quantum state by using a quantum neural network to obtain a prediction result of the feature information of the bank fixed deposit customers, constructing a loss function according to the prediction result of the feature information of the bank fixed deposit customers and a real label of the bank fixed deposit customers, updating parameters of the quantum neural network according to the loss function until a termination condition is met; and classifying feature information of a bank fixed deposit customer of a preset user according to the latest obtained quantum neural network. The quantum neural network is superior to a classical neural network in extraction of global features of bank deposit customer information and convergence speed, and requires less data for training.

The invention discloses a backdoor detection method for a quantum neural network. The backdoor detection method comprises the following steps: acquiring a quantum neural network to be detected; constructing a test sample; inputting the test sample into a quantum neural network to be detected, performing measurement to obtain an expected value of each quantum bit, and generating a corresponding measurement activation matrix; projecting the measurement activation matrix to obtain a dimension reduction feature matrix; clustering the dimension reduction feature matrix by adopting a clustering scheme; calculating a separation index between clustering results; and completing backdoor detection of the to-be-detected quantum neural network. The invention also discloses a system for realizing the backdoor detection method for the quantum neural network. According to the method, the backdoor detection for the quantum neural network is realized, the reliability is higher, and the applicability is better.

The present invention relates to a quantum convolutional neural network system and a method of operation for processing quantum states to estimate intrinsic properties. More specifically, the invention relates to a technique for transforming an input quantum state, measuring the first qubit, processing the state based on the measurement result, and calculating the intrinsic properties of the transformed state. The present invention provides a quantum convolutional neural network system composed of a learning unit, a quantum layer structure unit, and a computation unit. The learning unit receives an input quantum state and is designed to maximize the measurement probability of the first qubit by learning a unitary operator based on a local cost function. The learned unitary operator transforms the quantum state through the quantum layer structure unit; the transformed state is then measured by the first qubit, and the residual state is transmitted to the next layer or re-measured based on the measurement result. The present invention provides a technique that efficiently processes input data during the quantum neural network learning process and enables accurate analysis of quantum states and efficient dimensionality reduction. Through this, high performance and reliability can be secured in various application fields such as quantum information theory, quantum data analysis, and quantum machine learning.

The invention discloses a self-supervised training method and device based on a multi-scale residual quantum neural network. The self-supervised training method based on the multi-scale residual quantum neural network comprises the following steps: performing rotation processing on an original input image according to a plurality of preset rotation angles to generate a rotation image set, and taking the rotation angles as self-supervised labels; constructing a multi-scale residual quantum neural network comprising a quantum convolution kernel used for feature extraction, a feature fusion module used for fusing features of different scales and a residual connection module used for connecting input and output; training the multi-scale residual quantum neural network by using the rotating image set, and optimizing network parameters by minimizing a self-supervision loss function; and generating an upstream model for a downstream image task according to the trained network parameters. According to the embodiment of the invention, more abundant feature information can be extracted while the calculation is accelerated through the quantum convolution kernel.

An object detection device may include a variational autoencoder (VAE) configured to encode image data to generate a latent vector, and decode the latent vector to generate new image data, at least one quantum computing circuit configured to perform quantum subset summing, and a processor. The processor may be configured to cooperate with the at least one quantum computing circuit to train a plurality of deep learning models based upon a Quantum Neural Network (QNN), generate a game theory reward matrix for the plurality of deep learning models, perform quantum subset summing of the game theory reward matrix, select a deep learning model from the plurality thereof based upon the quantum subset summing of the game theory reward matrix, and process the new image data using the selected deep learning model for object detection.

The invention discloses a quantum neural network training method and a related device. The method comprises the following steps: acquiring a quantum logic gate sequence; setting at least one check point from the quantum logic gate sequence according to a certain number of interval nodes, and storing quantum state intermediate data obtained by forward propagation calculation of the at least one check point into a memory; if the current computing node is not the check point and the check point exists before the current computing node, backtracking to the nearest check point from the current computing node, and reading the quantum state intermediate data of the nearest check point from the memory; and performing reverse calculation based on the quantum state intermediate data of the closest check point to update parameters in the quantum neural network so as to realize training of the quantum neural network. By introducing a check point setting mechanism, the memory consumption in the quantum neural network training process is reduced, so that the resource space configuration is optimized, and the model training efficiency is improved.

The application provides an architecture search method, an electronic device, a chip and a storage medium for a quantum neural network. The method comprises the following steps: in response to receiving a quantum artificial intelligence task type, constructing a corresponding task objective function based on the task type; configuring a set of selectable quantum gates; based on a reinforcement learning strategy or an evolutionary algorithm, generating a candidate quantum neural network architecture in a search space constituted by the number of quantum bits, the number of network layers and the type and sequence of quantum gates used in each layer; performing parameter training on the candidate quantum neural network architecture; introducing a preset hardware noise model during the training process; applying a quantum measurement error compensation mechanism to the measurement results in the inference stage to correct the output deviation caused by the readout error; calculating the performance score of the candidate quantum neural network architecture; performing pruning optimization on the candidate architecture set, and outputting an optimal quantum neural network architecture.

The present application relates to the technical field of quantum state preparation, and particularly relates to a multi-quantum state preparation method, comprising: S1, randomly initializing an Ansatz circuit parameter; S2, inputting an initial quantum state into a variational quantum circuit under a current circuit parameter configuration to obtain a corresponding output state density matrix; S3, calculating a loss value between the output state density matrix and a target state according to the output state density matrix, feeding back the loss value to a classical optimizer, updating the Ansatz circuit parameter, returning to S2 until a preset condition is reached, and obtaining a trained variational quantum circuit; and S4, using the trained variational quantum circuit to realize preparation of multiple groups of target quantum states. The present application learns and approximates to prepare any target quantum state by constructing a quantum neural network structure with expression capability.

The application discloses a machine room operation and maintenance management automatic inspection system based on intelligence, and belongs to the technical field of intelligence, comprising a multi-source data fusion processing module, a hidden fault probability distribution module, a fault root cause analysis module, a dynamic self-adaptive baseline generation module, a multi-element strategy autonomous healing decision module, a collaborative control execution module and a feedback optimization module. The application constructs a full-link intelligent system of multi-source data fusion, quantum neural network prediction, graph neural network tracing, dynamic baseline generation, reinforcement learning decision, transactional scheduling execution and closed-loop optimization. The hidden fault recognition capability is improved through quantum-classical hybrid calculation. The root cause is accurately positioned in combination with a causal graph. An optimal repair scheme is generated based on a dynamic baseline and a reinforcement learning strategy library. Finally, the reliability of cross-system operation is ensured through transactional scheduling, and the fault prediction accuracy is significantly improved.

A method for determining a quantum neural network (QNN) configuration for classification in a problem domain by generating a population of candidate QNN configurations based on a number of feature val

The invention discloses a synchronous phasor measurement data anomaly detection method based on a quantum generative adversarial network, and the method comprises the steps: constructing a quantum-classical hybrid generative adversarial network model, introducing a quantum neural network based on a parameterized quantum circuit as a generator, and carrying out the adversarial training through combining with a Wasserstein distance, thereby achieving the anomaly detection of synchronous phasor measurement data. Therefore, while the number of trainable parameters of the model is reduced and the training stability is improved, the abnormality caused by the equipment fault or the power system operation event in the synchronous phasor measurement data is efficiently and reliably detected. According to the method, the training cost is reduced while the anomaly detection performance is ensured, the model training stability is enhanced in combination with the Wasserstein distance, and an efficient and reliable technical means is provided for operation monitoring and data quality control of a power system.

The invention provides a method for recognizing a target after reconstruction of a hyperspectral image by using a quantum generative adversarial network and a quantum neural network in quantum machine learning, and the method comprises the following steps: 10) carrying out the preprocessing of a to-be-measured hyperspectral image, and obtaining a standardized low-resolution hyperspectral image; step 20) inputting the preprocessed image into a quantum generative adversarial network for image reconstruction to obtain a hyperspectral reconstruction image; and step 30) inputting the reconstructed hyperspectral image into a target recognition model based on the quantum neural network, and outputting a repaired image and a recognition result. According to the hyperspectral image reconstruction identification method based on the quantum generative adversarial network and the quantum neural network, the hyperspectral image is not influenced by low resolution, and the identification efficiency and accuracy are improved.

The invention discloses a power distribution network high-resistance grounding fault identification method and device and electronic equipment. The method comprises the following steps: acquiring current electric signal data of a power distribution network line; feature extraction is carried out on the current electric signal data to obtain current electric power features, and the current electric power features comprise time domain features, frequency domain features and time-frequency domain features of the current electric signal data; based on the current electric power characteristics, a high-resistance grounding fault identification model is adopted to obtain a high-resistance grounding fault identification result of the power distribution network line at the current moment, and the high-resistance grounding fault identification model is based on historical electric power characteristics corresponding to multiple historical sampling periods and the high-resistance grounding fault identification result; and training the initial quantum neural network model. According to the invention, the technical problem of low identification accuracy of the high-resistance grounding fault of the power distribution network in the prior art is solved.

The invention discloses an energy-saving green traffic sign classification method based on a quantum pulse neural network. The method comprises the following steps: constructing a QTCAC-SNN model fusing quantum time channel attention coding QTCAC and a membrane residual neural network MS-ResNet; the method comprises the following steps: extracting attention features of an input image in a time dimension and a channel dimension by using a quantum neural network QNNs through a QTCAC module, carrying out fusion processing on the features through an attention fusion unit, and encoding the image into a pulse sequence with higher expression capability; according to the method, a QNN and SNN mixed training method is adopted to train the model, and the trained model is deployed in a traffic sign image classification task to realize green and efficient calculation. According to the method, a quantum time channel attention coding mechanism is innovatively provided, classification tasks are executed in combination with quantum calculation and a mixed training method, and energy consumption is remarkably reduced while high accuracy of the model is guaranteed.

This application provides a data processing method, apparatus, device, and computer program product. The method includes: measuring the quantum circuits in a quantum neural network model to be trained based on a first measurement basis and a second measurement basis, and obtaining measurement results; the first measurement basis is a random measurement basis; the second measurement basis is a quantum state feature enhancement measurement basis; extracting multiple feature fluctuation data and multiple feature information entropy data corresponding to multiple features from the measurement results; determining whether to perform dropout on multiple features based on the multiple feature fluctuation data and multiple feature information entropy data, and obtaining multiple judgment results corresponding to multiple features; training the quantum neural network model to be trained based on the multiple judgment results, and obtaining a trained quantum neural network model; the trained quantum neural network model is used at least for processing quantum data. This improves dropout efficiency.