Atomic quantum bit state measurement method and device based on convolutional neural network

Abstract

The invention provides an atomic quantum bit state measurement method and device based on a convolutional neural network, and relates to the technical field of quantum computing. The atomic quantum bit state measurement method based on the convolutional neural network comprises the steps that cut original image data is acquired, and the original image data is a neutral atom array fluorescence image of a quantum state or an atom existence state to be measured; processing the original image data based on a preset convolutional neural network to determine a high-dimensional feature vector; obtaining measured experimental environment data, and fusing the experimental environment data with the high-dimensional feature vector according to a preset feature splicing mode to generate an extended feature vector; and according to the convolutional neural network, determining a prediction result of the atomic quantum bit state corresponding to the extended feature vector. By splicing the experimental environment data and the original image data, the actual requirements of high fidelity and rapid real-time measurement can be met at the same time when the experimental parameters are changed due to the influence of various factors.

Description

Technical Field

[0001] The present application relates to the field of quantum computing technology, and for example, to a method and device for measuring the state of an atomic quantum bit based on a convolutional neural network. Background Art

[0002] In quantum mechanics, the state of a physical system is described by a wave function, and the modulus of the wave function gives the probability that the system is in a certain state. The quantum state is a physical quantity that describes the state of a quantum system and is a core concept in quantum mechanics. Quantum state measurement refers to the process of observing and measuring the state of a quantum system under the framework of quantum mechanics. The existence state of an atom is uncertain when it is not observed. When an atom is measured, its state collapses according to the rules of quantum mechanics, thereby showing a definite existence state.

[0003] The detection of quantum states and atomic existence states can be used as a whole to measure the atomic quantum bit state. Convolutional neural network is a feedforward neural network with deep structure that includes convolution calculations. It is one of the representative algorithms in the field of deep learning. High-fidelity quantum state measurement is the core foundation for achieving high-quality quantum computing, real-time error correction, and high-precision quantum logic operations, and directly determines the accuracy of quantum bit initialization, operation, and final calculation results.

[0004] In the related technology, traditional quantum state measurements are generally performed based on fixed thresholds in the scenarios of atomic physics quantum computing, atomic detection, or atomic quantum state detection. Although it is simple and easy to distinguish quantum states by setting a fixed threshold, it is difficult to obtain high fidelity under short exposure conditions due to insufficient signal-to-noise ratio. At present, the existing convolutional neural network (CNN) method is used in the process of quantum state measurement. Although it improves the measurement accuracy under short exposure conditions, the model needs to be retrained every time the experimental conditions (such as exposure time, laser intensity, background noise, etc.) change to ensure the measurement accuracy, and the generalization ability is insufficient. Furthermore, the experimental parameters (such as exposure time, laser intensity, etc.) are affected by various factors such as environmental changes and equipment aging, and there is non-deterministic dynamic drift during the measurement process. Therefore, the generalization performance of the current atomic quantum bit state measurement method is insufficient, and it cannot adapt to the dynamic changes and drifts of experimental parameters in real time or quickly, resulting in reduced measurement accuracy. Summary of the invention

[0005] The present application aims to provide a method and device for measuring the state of atomic quantum bits based on a convolutional neural network.

[0006] According to one aspect of the present application, a method for measuring the state of an atomic quantum bit based on a convolutional neural network is proposed, comprising: obtaining cut original image data, wherein the original image data is a fluorescence image of a neutral atom array of a quantum state or an atomic existence state to be measured; based on a preset convolutional neural network, processing the original image data to determine a high-dimensional feature vector; obtaining measured experimental environment data, and according to a preset feature splicing method, fusing the experimental environment data with the high-dimensional feature vector to generate an extended feature vector; and determining, according to the convolutional neural network, a prediction result of the atomic quantum bit state corresponding to the extended feature vector.

[0007] According to one aspect of the present application, a device for measuring the state of an atomic quantum bit based on a convolutional neural network is proposed, comprising: A data acquisition module, used to acquire the cut original image data, wherein the original image data is a fluorescence image of a neutral atom array of a quantum state or an atomic existence state to be measured; An image data processing module, used for processing the original image data based on a preset convolutional neural network to determine a high-dimensional feature vector; A feature splicing module is used to obtain the measured experimental environment data and fuse the experimental environment data with the high-dimensional feature vector according to a preset feature splicing method to generate an extended feature vector; The prediction module is used to determine the prediction result of the atomic quantum bit state corresponding to the extended feature vector according to the convolutional neural network.

[0008] According to one aspect of the present application, an electronic device is provided, the electronic device comprising: a processor; and a memory storing a computer program, wherein when the computer program is executed by the processor, the processor executes the method as described above.

[0009] According to one aspect of the present application, a non-transitory computer-readable medium is provided, on which readable instructions are stored. When the instructions are executed by a processor, the processor executes the method as described above.

[0010] It should be understood that the foregoing general description and the following detailed description are exemplary only and are not restrictive of the present application.

[0011] Beneficial effects: Through the above-mentioned embodiments provided by the present application, by using convolutional neural networks to perform deep processing on the original image data, high-dimensional feature vectors in the image can be extracted, and these feature vectors contain rich quantum system information. At the same time, combined with experimental environment data, such as laser power, exposure time, background environment parameters, etc., the feature set is further enriched, making the prediction of quantum state more comprehensive and accurate. This multi-dimensional information fusion significantly improves the accuracy of quantum state measurement, and provides strong support for error correction and quantum state manipulation in quantum computing. Since the experimental environment data is taken into consideration, the present application can cope with the quantum state measurement requirements under different experimental conditions. Whether it is laser power jitter, exposure time adjustment, or temperature drift, vector splicing can be used to achieve stable prediction of quantum state, without the need to train convolutional neural networks multiple times, reducing measurement time. On the whole, it can meet the actual needs of high fidelity and fast real-time measurement at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without exceeding the scope of protection required by the present application.

[0013] Figure 1 A flow chart of a method for measuring the state of an atomic quantum bit based on a convolutional neural network provided in an embodiment of the present application; Figure 2 A block diagram of a device for measuring the state of an atomic quantum bit based on a convolutional neural network provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0014] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this application will be comprehensive and complete and will fully convey the concepts of the example embodiments to those skilled in the art. The same reference numerals in the figures represent the same or similar parts, and thus their repeated description will be omitted.

[0015] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0016] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0017] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0018] It should be understood that although the terms first, second, third, etc. may be used herein to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another component. Therefore, the first component discussed below can be referred to as the second component without departing from the teachings of the concepts of the present application. As used herein, the term "and / or" includes any one of the associated listed items and all combinations of one or more.

[0019] At present, quantum measurement technology mainly relies on advanced physical concepts and technical means such as quantum mechanics principles, quantum entanglement, and quantum interference. Feature splicing methods are generally used in the field of image processing. This application provides a method for measuring the state of atomic quantum bits combined with feature splicing in the scenario of quantum state measurement, which helps to capture the complexity and diversity of quantum states and improve the accuracy and reliability of measurements. By splicing multiple features, the dimension and amount of information of the measurement data can be increased, thereby increasing the sensitivity to subtle changes in the quantum state, which is very important for quantum state measurements that require high-fidelity measurements.

[0020] For specific implementation methods, please refer to the following embodiments.

[0021] Figure 1 This is a flow chart of a method for measuring the state of an atomic quantum bit based on a convolutional neural network provided in an embodiment of the present application. The method of this embodiment can be applied to a measurement server. Figure 1As shown, the method includes: step S10, step S11, step S12 and step S13.

[0022] In step S10, the cut original image data is acquired, wherein the original image data is a fluorescence image of a neutral atom array of a quantum state or an atomic existence state to be measured.

[0023] In this application, the quantum state to be measured can be used to characterize the state of the quantum bit that needs to be measured, and the atomic existence state can be used to characterize the state of the existence of the atom that needs to be measured. A high quantum efficiency camera can be used to capture a neutral atom array fluorescence image in real time, which contains multiple atoms. Each atom in the image is used as a sub-image content to perform image cutting to obtain the cut original image data. The terminal device can collect the original image data and send it to the measurement server, which receives it accordingly.

[0024] In some implementations, after acquiring the original image data, the terminal device may first perform preprocessing operations such as noise removal, image enhancement, and standardization on the original image data to improve data quality and reduce the impact of interference signals.

[0025] In step S11, the original image data is processed based on a preset convolutional neural network to determine a high-dimensional feature vector.

[0026] In the present application, a convolutional neural network can be pre-established, which includes a convolutional layer, an activation function, and a pooling layer, etc. The convolutional neural network can extract features from the original image data, and use global average pooling after the last convolutional layer to obtain a high-dimensional feature vector.

[0027] In step S12, the measured experimental environment data is obtained, and according to a preset feature splicing method, the experimental environment data is fused with the high-dimensional feature vector to generate an extended feature vector.

[0028] The experimental environment data in this application may include but is not limited to exposure time, laser intensity, and background environment parameters. The terminal device can collect the experimental environment data in real time and then upload it to the measurement server. The feature splicing method can be pre-set, and feature vectors from different sources or different dimensions are merged along a preset specific dimension to form a new, richer, and complete extended feature vector method.

[0029] In some implementations, the experimental environment data and the high-dimensional feature vector may be combined in a feature concatenation manner to obtain an extended feature vector.

[0030] In step S13, the prediction result of the atomic quantum bit state corresponding to the extended feature vector is determined according to the convolutional neural network.

[0031] In the present application, a neural network allocator or regressor can be pre-set in the convolutional neural network, and the extended feature vector can be mapped to the quantum state as a prediction result of the quantum state to be measured, or the presence of an atom can be detected as a prediction result of the atomic existence state. The prediction result of the quantum state to be measured and the prediction result of the atomic existence state can be used as a prediction result of the atomic quantum bit state.

[0032] This application uses convolutional neural networks to perform deep processing on the original image data, and can extract high-dimensional feature vectors in the image. These feature vectors contain rich quantum system information. At the same time, combined with experimental environment data, such as laser power, exposure time, background environment parameters, etc., the feature set is further enriched, making the prediction of quantum states more comprehensive and accurate. This multi-dimensional information fusion significantly improves the accuracy of quantum state measurement, and provides strong support for error correction and quantum state manipulation in quantum computing. Since the experimental environment data is taken into consideration, this application can cope with the needs of quantum state measurement under different experimental conditions. Whether it is laser power jitter, exposure time adjustment, or temperature drift, vector splicing can be used to achieve stable prediction of quantum states, without the need to train convolutional neural networks multiple times, reducing measurement time. On the whole, it can meet the actual needs of high-fidelity and fast real-time measurement at the same time.

[0033] According to some embodiments, a logistic regression classifier preset on a convolutional neural network may be obtained; and the expanded feature vector is input into the logistic regression classifier to determine a prediction result.

[0034] In this application, an integrated model including CNN and logistic regression classifier can be constructed. The logistic regression classifier can be connected through the fully connected layer of CNN. The extended feature vector is input into the logistic regression classifier, and the logistic regression classifier can output a probability distribution to represent the probability that the extended feature vector belongs to each quantum state category, and the probability is used as the prediction result of the atomic quantum bit state.

[0035] This application can make full use of the high-dimensional feature information extracted by the convolutional neural network and the experimental environment data by inputting the extended feature vector into the logistic regression classifier to achieve accurate prediction of the quantum state. As a linear classifier, the logistic regression classifier can give a more accurate classification result when the feature space is linearly separable, thereby improving the prediction accuracy of quantum state measurement.

[0036] According to some embodiments, the logistic regression classifier may be a softmax classifier. The extended feature vector may be input into the softmax classifier to determine the category of the quantum state and its corresponding category probability; the category and the category probability are determined as the prediction result.

[0037] In this application, softmax classifier (Softmax Classifier, softmax multi-classifier) ​​is a classification model in machine learning, which is a generalization of logistic regression on multi-classification problems. The softmax function can map a vector containing any real number into a probability distribution, that is, the value of each element is between 0 and 1, and the sum of all element values ​​is 1. In the softmax classifier, after the input feature vector is linearly transformed, the probability distribution of each category is obtained by the softmax function, which indicates the probability that the input extended feature vector belongs to each quantum state category. In some implementations, the category with the highest probability can be selected as the prediction result.

[0038] This application uses a softmax classifier, which can output the probability distribution of each category compared to traditional binary logistic regression, thereby more accurately describing the relationship between the input feature vector and each quantum state category. This feature enables this method to cope with more complex classification tasks in quantum state measurement, improving the accuracy and robustness of classification. As a form of logistic regression, the softmax classifier has good scalability and flexibility. In quantum state measurement, as experimental conditions and measurement requirements change, the input feature vector of the softmax classifier can be easily adjusted to adapt to different application scenarios. According to some embodiments, the prediction results may be uploaded to an associated terminal device so that the terminal device displays the results.

[0039] After obtaining the predicted results of the atomic quantum bit state, this application can upload it to the associated terminal device, that is, the device that collects the original image data and experimental environment data in the above steps. Users can obtain the latest measurement and analysis results in real time. This real-time performance is crucial for quantum experiments and quantum computing applications that require rapid response, helping staff to adjust experimental parameters or optimize algorithms in a timely manner, thereby improving research efficiency and accuracy.

[0040] According to some embodiments, a neutral atom array fluorescence image sent by a terminal device may be received; and the neutral atom array fluorescence image may be preprocessed to generate raw image data.

[0041] In this application, the neutral atom array fluorescence image is an image formed by the fluorescence emitted by the neutral atom array captured by a high quantum efficiency camera in the neutral atom quantum computing platform. The terminal device can collect the neutral atom array fluorescence image captured by the high quantum efficiency camera and then send it to the measurement server.

[0042] In some implementations, the measurement server may pre-process the received neutral atom array fluorescence image to obtain raw image data.

[0043] In other implementations, the terminal device may pre-process the neutral atom array fluorescence image, obtain the raw image data, and then send it to the measurement server.

[0044] This application can directly obtain key visual information in quantum computing or quantum simulation experiments by receiving neutral atomic array fluorescence images sent by terminal devices. Preprocessing the image can remove unnecessary background information in the image and focus on the key atomic array area. This greatly improves the accuracy and pertinence of the data, and provides a clearer and more definite image basis for subsequent data analysis. Operations such as denoising and contrast enhancement in the preprocessing step can improve image quality and reduce the misjudgment rate in subsequent analysis. This helps to improve the accuracy of the entire image processing process.

[0045] According to some embodiments, the size of the original image data can be adjusted according to a preset input size of the convolutional neural network to determine the target image data; the target image data can be forward propagated based on the convolutional neural network; a feature map of the target image data can be extracted in a preset intermediate layer of the convolutional neural network; and the feature map can be flattened into a high-dimensional feature vector.

[0046] In this application, forward propagation refers to the process in which data starts from the input layer of the network, passes through the convolution layer, pooling layer, fully connected layer, etc. layer by layer, and finally reaches the output layer. The feature map can contain important information in the image, such as edges, textures, shapes, etc., which can be used for subsequent classification, recognition and other tasks.

[0047] In some implementations, the received raw image data can be resized according to the preset input size requirements of the convolutional neural network. Specifically, the image can be scaled, cropped, or padded to ensure that the image data meets the size standard of the network input. In some implementations, an image processing library can be used to implement the resizing. After resizing, the image data obtained is the target image data. These data will serve as the input of the convolutional neural network. Before the data is sent to the network, it can be normalized or standardized to eliminate the brightness or contrast differences between different images and improve the generalization ability of the network.

[0048] Input the target image data into the preset convolutional neural network for forward propagation. During the forward propagation process, each layer will calculate the input data according to the weights and biases and output feature maps or activation values. In the preset intermediate layer of the convolutional neural network (such as after a convolutional layer or pooling layer and before the fully connected layer), extract the feature map of the target image data. You can obtain the feature map by using the API (Application Programming Interface) provided by the deep learning framework to access the intermediate layer output of the network.

[0049] Through the reshape operation (Reshape Operation, reshaping operation) or flatten operation (Flatten Operation, flattening operation), each element of the feature map can be arranged in order into a long vector, and the extracted feature map can be flattened into a one-dimensional high-dimensional feature vector. Since convolutional neural networks usually output two-dimensional or three-dimensional feature maps, subsequent classifiers or regressors may require one-dimensional input.

[0050] This application ensures that the image data can be seamlessly connected to the network by resizing the original image data according to the preset input size of the convolutional neural network, avoiding data loss or information distortion caused by size mismatch. This step enhances the adaptability of the image data and provides a solid foundation for subsequent feature extraction and recognition. Using the forward propagation mechanism of the convolutional neural network, the target image data can be transmitted layer by layer in the network, and gradually abstracted through operations such as convolution and pooling, thereby extracting deep feature information. Compared with the original image data, these feature information has a higher level of abstraction and stronger expression ability, which provides strong support for subsequent image classification, recognition and other tasks. Extracting feature maps in the preset middle layer of the convolutional neural network makes full use of the output information of the middle layer of the network and avoids information waste. Flattening the feature map into a high-dimensional feature vector converts the original multi-dimensional image data into a one-dimensional vector form, which is convenient for subsequent classifiers or regressors to process. Not only does it simplify the data processing process, but it also improves the efficiency and accuracy of data processing. At the same time, high-dimensional feature vectors have stronger representation and generalization capabilities, and can better adapt to image recognition tasks in different scenarios.

[0051] According to some embodiments, experimental environment data can be obtained and normalized to determine target environment data; the environment data splicing order is extracted from the feature splicing method; and the target environment data is spliced ​​with a high-dimensional feature vector according to the environment data splicing order to generate an extended feature vector.

[0052] In this application, the purpose of normalization is to convert data of different dimensions into the same dimension, eliminate the adverse effects of numerical differences, and ensure that the data is compared and analyzed on a unified scale. Normalization can be performed using a variety of methods, such as linear function conversion, logarithmic function conversion, and inverse cotangent function conversion. The specific method to be selected depends on the characteristics and requirements of the data.

[0053] In some implementations, the experimental environment data sent by the terminal device can be received first, and the data can be normalized to obtain the target environment data. The experimental environment data may be one or more groups. If there is only one group, it can be directly spliced ​​after the high-dimensional feature vector according to the feature splicing method. The spliced ​​data is one dimension higher than the high-dimensional feature vector. If there are multiple groups of experimental environment data, the environment data splicing order can be extracted from the feature splicing method. According to the order, multiple groups of experimental environment data are spliced ​​after the high-dimensional feature vector to obtain an extended feature vector.

[0054] This application realizes the effective fusion of experimental environment data and high-dimensional feature vectors, expands the dimension of feature vectors, and increases the richness and comprehensiveness of feature information. By fusing environmental data, the actual conditions in image recognition or classification tasks can be more comprehensively reflected, and the recognition accuracy and robustness of the system are improved. The target environmental data and high-dimensional feature vectors are spliced ​​in the order of environmental data splicing, so that this application can flexibly adapt to data input of different dimensions, enhancing versatility and scalability.

[0055] The following describes an apparatus embodiment of the present application, which can be used to execute the method embodiment of the present application. For details not disclosed in the apparatus embodiment of the present application, reference can be made to the method embodiment of the present application.

[0056] Figure 2 A block diagram of a device for measuring the state of an atomic quantum bit based on a convolutional neural network provided in an embodiment of the present application. Figure 2 As shown, the device 200 for measuring the state of atomic quantum bits based on a convolutional neural network includes a data acquisition module 201, an image data processing module 202, a feature splicing module 203 and a prediction module 204.

[0057] The data acquisition module 201 is used to acquire the cut original image data, wherein the original image data is a fluorescence image of a neutral atom array of a quantum state or an atomic existence state to be measured; An image data processing module 202 is used to process the original image data based on a preset convolutional neural network to determine a high-dimensional feature vector; The feature splicing module 203 is used to obtain the measured experimental environment data, and fuse the experimental environment data with the high-dimensional feature vector according to a preset feature splicing method to generate an extended feature vector; The prediction module 204 is used to determine the prediction result of the atomic quantum bit state corresponding to the extended feature vector according to the convolutional neural network.

[0058] Optionally, the prediction module 204 is specifically used for: Get the preset logistic regression classifier on the convolutional neural network; The expanded feature vector is input into a logistic regression classifier to determine the prediction result.

[0059] Optionally, the logistic regression classifier is a softmax classifier; when the prediction module 204 inputs the extended feature vector into the logistic regression classifier to determine the prediction result, it is specifically used to: Input the expanded feature vector into the softmax classifier to determine the category of the quantum state and its corresponding category probability; The class and class probability are determined as the prediction results.

[0060] Optionally, the atomic quantum bit state measurement device 200 based on a convolutional neural network includes a prediction result uploading module 205, which is used to: The prediction results are uploaded to the associated terminal device so that the terminal device can display the results.

[0061] Optionally, the data acquisition module 201 is specifically used for: receiving a neutral atom array fluorescence image sent by a terminal device; The neutral atom array fluorescence image is preprocessed to generate raw image data.

[0062] Optionally, the image data processing module 202 is specifically used for: According to a preset input size of the convolutional neural network, the size of the original image data is adjusted to determine the target image data; Perform forward propagation of target image data based on convolutional neural network; Extracting a feature map of the target image data in a preset middle layer of the convolutional neural network; Flatten the feature map into a high-dimensional feature vector.

[0063] Optionally, the feature splicing module 203 is specifically used for: Acquire experimental environment data and normalize the experimental environment data to determine target environment data; Extracting the environment data splicing order from the feature splicing method; According to the concatenation order of the environment data, the target environment data is concatenated with the high-dimensional feature vector to generate an extended feature vector.

[0064] The device performs functions similar to the method provided above. For other functions, please refer to the previous description and will not be repeated here.

[0065] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application, such as Figure 3 As shown, the electronic device 300 of this embodiment may include: a memory 301 and a processor 302 .

[0066] The memory 301 stores a computer program. When the computer program is executed by the processor 302, the processor 302 executes the method in the above embodiment.

[0067] The processor 302 and the memory 301 are connected, for example, via a bus.

[0068] Optionally, the electronic device 300 may further include a transceiver. It should be noted that in actual applications, the number of transceivers is not limited to one, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present application.

[0069] The processor 302 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor 302 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0070] The bus may include a path to transmit information between the above components. The bus may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0071] The memory 301 can be a ROM (Read Only Memory) or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory) or other types of dynamic storage devices that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.

[0072] The memory 301 is used to store the application code for executing the solution of the present application, and the execution is controlled by the processor 302. The processor 302 is used to execute the application code stored in the memory 301 to implement the contents shown in the above method embodiment.

[0073] The electronic devices include, but are not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server, etc. Figure 3 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0074] The electronic device of this embodiment can be used to execute the method of any of the above embodiments, and its implementation principle and technical effect are similar, which will not be described in detail here.

[0075] The present application also provides a non-transitory computer-readable storage medium having computer-readable instructions stored thereon. When the aforementioned instructions are executed by a processor, the processor executes the method in the above embodiment.

[0076] Those skilled in the art can understand that all or part of the steps of implementing the above-mentioned method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a non-transient computer-readable storage medium. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the aforementioned storage medium includes: ROM, RAM, disk or optical disk and other media that can store program codes.

[0077] The embodiments of the present application are described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and its core idea of ​​the present application. At the same time, changes or deformations made by those skilled in the art based on the ideas of the present application, the specific implementation methods and the scope of application of the present application, all belong to the scope of protection of the present application. In summary, the content of this specification should not be construed as a limitation on the present application.

Claims

1. A method for measuring the state of an atomic quantum bit based on a convolutional neural network, characterized in that: include: Acquire the cut original image data, wherein the original image data is a fluorescence image of a neutral atom array of a quantum state or an atomic existence state to be measured; Based on a preset convolutional neural network, the original image data is processed to determine a high-dimensional feature vector; Acquire measured experimental environment data, and fuse the experimental environment data with the high-dimensional feature vector according to a preset feature splicing method to generate an extended feature vector; According to the convolutional neural network, a prediction result of the atomic quantum bit state corresponding to the extended feature vector is determined.

2. The method according to claim 1, characterized in that Determining the prediction result of the atomic quantum bit state corresponding to the extended feature vector according to the convolutional neural network includes: Obtaining a preset logistic regression classifier on the convolutional neural network; The expanded feature vector is input into the logistic regression classifier to determine the prediction result.

3. The method according to claim 2, characterized in that The logistic regression classifier is a softmax classifier; The step of inputting the extended feature vector into the logistic regression classifier to determine the prediction result includes: Inputting the extended feature vector into the softmax classifier to determine the category of the atomic quantum bit state and its corresponding category probability; The category and the category probability are determined as the prediction result.

4. The method according to any one of claims 1 to 3, characterized in that: Also includes: The prediction results are uploaded to the associated terminal device so that the terminal device can display the results.

5. The method according to claim 4, characterized in that The obtaining of the cut original image data comprises: receiving the neutral atom array fluorescence image sent by the terminal device; The neutral atom array fluorescence image is preprocessed to generate the raw image data.

6. The method according to claim 1, characterized in that The processing of the original image data based on a preset convolutional neural network to determine a high-dimensional feature vector includes: According to a preset input size of the convolutional neural network, the size of the original image data is adjusted to determine the target image data; Performing forward propagation on the target image data based on the convolutional neural network; Extracting a feature map of the target image data in a preset middle layer of the convolutional neural network; The feature map is flattened into the high-dimensional feature vector.

7. The method according to claim 1, characterized in that The step of acquiring the measured experimental environment data and fusing the experimental environment data with the high-dimensional feature vector according to a preset feature splicing method to generate an extended feature vector includes: Acquire the experimental environment data, and perform normalization processing on the experimental environment data to determine target environment data; Extracting the environment data splicing order from the feature splicing method; The target environment data is concatenated with the high-dimensional feature vector in the order of concatenating the environment data to generate the extended feature vector.

8. A device for measuring the state of atomic quantum bits based on a convolutional neural network, characterized in that: include: A data acquisition module, used to acquire the cut original image data, wherein the original image data is a fluorescence image of a neutral atom array in a quantum state to be measured; An image data processing module, used for processing the original image data based on a preset convolutional neural network to determine a high-dimensional feature vector; A feature splicing module, used for acquiring measured experimental environment data, and fusing the experimental environment data with the high-dimensional feature vector according to a preset feature splicing method to generate an extended feature vector; A prediction module is used to determine a prediction result of the atomic quantum bit state corresponding to the extended feature vector based on the convolutional neural network.

9. An electronic device, characterized in that: include: processor; A memory storing a computer program, which, when executed by the processor, enables the processor to perform the method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the instructions are executed by a processor, the processor is caused to execute the method according to any one of claims 1 to 7.

Images