Chaotic signal prediction method and device based on quantum memristive system
Patent Information
- Application Number
- CN202511567427.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-30
- Publication Date
- 2026-10-30
AI Technical Summary
[0005]本申请实施例提供一种基于量子忆阻系统的混沌信号预测方法、装置、计算机设备、计算机可读存储介质、计算机程序产品，以解决或缓解上面提出的一项或更多项技术问题
[0024]本申请实施例采用上述技术方案可以包括如下优势：通过融合量子机器学习与光量子忆阻器技术，量子忆阻系统利用量子态叠加特性，能够高效捕捉并学习混沌系统中复杂的非线性特征，相较于经典忆阻系统，能够有效降低预测误差，使预测结果更贴近实际值，提高混沌信号预测任务的计算精度与效率，突破高维数据处理中的性能限制，为人工智能硬件实现提供新路径。
Smart Images
Figure CN123014044A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of neuromorphic technology, and in particular to a method and apparatus for predicting chaotic signals based on a quantum memristor system. Background Technology
[0002] In the field of electrical neuromorphic computing, memristors have been widely studied due to their ability to efficiently simulate synaptic plasticity; however, their development is limited by challenges such as crosstalk and system-level energy consumption. In contrast, photonic memristors, using light as the information carrier, possess great potential for reducing energy consumption and improving computational performance due to the multidimensional properties and low crosstalk of light. Building on this, the concept of memristors is extended to quantum systems to construct quantum memristor architectures. By utilizing the nonlinear dynamic and quantum properties of memristors to achieve information processing, novel solutions are provided for complex computational tasks.
[0003] Optical quantum memristors, composed of fiber optic devices and equipped with tunable parameters (such as the phase of a phase modulator), can be combined with quantum machine learning techniques to construct quantum memristor systems centered on optical quantum memristors, enabling efficient processing of complex multidimensional data tasks. However, performance bottlenecks exist when processing high-dimensional nonlinear data in complex machine learning tasks, especially when dealing with turbid signal prediction tasks, where efficiency is low.
[0004] It should be noted that the above content is not necessarily prior art, nor is it intended to limit the scope of patent protection of this application. Summary of the Invention
[0005] This application provides a method, apparatus, computer device, computer-readable storage medium, and computer program product for predicting chaotic signals based on a quantum memristor system, in order to solve or alleviate one or more of the technical problems mentioned above.
[0006] One aspect of this application provides a method for predicting chaotic signals based on a quantum memristor system, wherein the quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detector terminal, and the method includes: A single-photon source signal is generated through the input terminal, and the single-photon source signal is encoded into a first quantum state; The first quantum state is transformed into the second quantum state using the optical quantum memristor; Density matrix information is obtained by detecting the second quantum state through the probe end; The density matrix information is analyzed using a preset signal prediction model to obtain the turbidity signal prediction result.
[0007] Optionally, the input terminal includes an information encoding module for encoding the single-photon source signal, the encoding process of which satisfies the following expression:
[0008]
[0009] in, This indicates the phase of the phase modulator in the information encoding module. Represents the first quantum state of the output; coefficients These are the complex elements of the transformation matrix T, representing the three photons appearing in the output channel respectively. The probability amplitude of this event, and matrix T satisfies ( (The identity matrix); This represents the creation operator for the i-th output channel. This represents the creation operator for the j-th output channel. This represents the creation operator for the k-th output channel. It is in a vacuum state.
[0010] Optionally, the photonic quantum memristor is constructed using a Mach-Zehnder interferometer (MZI); the mathematical expression for the transformation of the first quantum state by the photonic quantum memristor is:
[0011] in, Represents the first quantum state. Represents the second quantum state. This is the unitary transformation matrix of the photonic quantum memristor, which depends on the state variables of the photonic quantum memristor. First quantum state and time .
[0012] Optionally, the state variable R of the optical quantum memristor satisfies the following expression:
[0013] One of the output ports in the MZI is defined as the feedback port. The expected number of photons measured at this feedback port.
[0014] Optionally, the state variable R is phase-controlled by the phase modulator in the MZI and is defined as:
[0015] Where θ represents the phase of the phase modulator in MZI.
[0016] Optionally, the method further includes: Based on the measurement results from the feedback port, the phase of the phase modulator is updated to update the state variable R.
[0017] Optionally, the training process of the preset signal prediction model satisfies the following expression:
[0018] in, The actual output value vector of the training set samples. The input feature matrix of the training set samples, This is the weight matrix of the model.
[0019] Optionally, the method further includes: The preset signal prediction model is tested, and the testing process satisfies the following expression:
[0020] in, This is the weight matrix of the model. The input feature matrix for the test set samples. This is the corresponding predicted output value.
[0021] Optionally, the method further includes: The performance of the preset signal prediction model is evaluated according to the standardized mean square error loss function, which satisfies the following expression:
[0022] in, Indicates the first The actual output values of each test set sample. This represents the corresponding predicted output value. This is the sum of squares of the actual output values for all test set samples.
[0023] Another aspect of this application provides a chaotic signal prediction device based on a quantum memristor system, characterized in that the device comprises: An encoding module is used to generate a single-photon source signal through the input terminal and encode the single-photon source signal into a first quantum state; A memristor module is used to convert the first quantum state into a second quantum state through the optical quantum memristor; The detection module is used to obtain density matrix information by detecting the second quantum state through the detection terminal; The turbidity signal prediction module is used to analyze the density matrix information through a preset signal prediction model to obtain the turbidity signal prediction result.
[0024] The embodiments of this application employing the above-mentioned technical solutions may include the following advantages: by integrating quantum machine learning and optical quantum memristor technology, the quantum memristor system utilizes the superposition characteristics of quantum states to efficiently capture and learn complex nonlinear features in chaotic systems. Compared with classical memristor systems, it can effectively reduce prediction errors, make prediction results closer to actual values, improve the computational accuracy and efficiency of chaotic signal prediction tasks, break through the performance limitations in high-dimensional data processing, and provide a new path for the realization of artificial intelligence hardware. Attached Figure Description
[0025] The accompanying drawings exemplify embodiments and form part of the specification, serving together with the textual description to explain exemplary implementations of the embodiments. The illustrated embodiments are for illustrative purposes only and do not limit the scope of the claims. Throughout the drawings, the same reference numerals refer to similar but not necessarily identical elements.
[0026] Figure 1 A flowchart illustrating a chaotic signal prediction method based on a quantum memristor system according to Embodiment 1 of this application is shown in the schematic diagram. Figure 2 A schematic diagram of the structure of a quantum memristor system is shown. Figure 3 A schematic diagram of the input structure of a quantum memristor system is shown, wherein, Figure 3 (a) shows a schematic diagram of the structure of MZI. Figure 3 (b) is a structural diagram of the information encoding module; Figure 4 A flowchart illustrating the analysis process for a chaotic signal prediction task is shown schematically. Figure 5 A block diagram of a chaotic signal prediction device based on a quantum memristor system according to Embodiment 2 of this application is shown schematically. Detailed Implementation
[0027] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.
[0028] It should be noted that the descriptions involving "first," "second," etc., in the embodiments of this application are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.
[0029] In the description of this application, it should be understood that the numerical labels before the steps do not indicate the order of the steps, but are only used to facilitate the description of this application and to distinguish each step, and therefore should not be construed as a limitation of this application.
[0030] First, a definition of the terminology used in this application is provided: The first Mach-Zehnder interferometer (MZI) is an interferometer used to observe the relative phase shifts that occur when a beam of light emitted from a single source splits into two collimated beams and travels through different paths and media.
[0031] Phase modulator: A phase modulator is an optoelectronic device that modulates signals by changing the phase of light waves, used to make the phase of light change according to a certain pattern.
[0032] Supervised learning is the process of adjusting the parameters of a classifier using a set of samples of known classes to achieve the required performance. It is also known as supervised training or teacher-led learning.
[0033] Secondly, to facilitate understanding of the technical solutions provided in the embodiments of this application by those skilled in the art, the relevant technologies are described below: The rapid development of artificial intelligence has led to challenges such as computing power bottlenecks and excessive energy consumption in existing electronic computing systems. To address these challenges, neuromorphic computing, by simulating the neural functions of the human brain and using electronic devices to simulate synapses, neurons, and axons, constructs brain-like neural networks, providing an important approach for novel computing architectures.
[0034] In the field of electrical neuromorphic computing, memristors have been widely studied due to their ability to efficiently simulate synaptic plasticity; however, their development is limited by challenges such as crosstalk and system-level energy consumption. In contrast, photonic memristors, using light as the information carrier, possess great potential for reducing energy consumption and improving computational performance due to the multidimensional properties and low crosstalk of light. Building on this, the concept of memristors is extended to quantum systems to construct quantum memristor architectures. By utilizing the nonlinear dynamic and quantum properties of memristors to achieve information processing, novel solutions are provided for complex computational tasks.
[0035] Quantum memristors are designed based on classical memristors, the concept of which originated from a theoretical prediction made by Professor Shao-Tang Tsai of the University of California, Berkeley in 1971. Hewlett-Packard Laboratories first linked the resistive switching behavior of solid-state devices with the conceptual memristor in 2008. Quantizing memristors was a complex challenge, only recently achieved. In 2016, P. Pfeiffer et al. solved this problem by constructing a memristor system using a tunable resistor, a weak measurement protocol, and classical feedback acting on the system's resistive coupling, proving that this memristor system satisfies both memristor dynamics and quantum dynamics, thus enabling its use in quantum information tasks. In 2017, J. Salmilehto et al. realized a quantum memristor on a superconducting platform by utilizing the phenomenon of naturally occurring memristor behavior in the presence of Josephson junctions.
[0036] Optical quantum memristors are composed of fiber optic devices equipped with adjustable parameters (such as the phase of a phase modulator). By combining quantum machine learning techniques, a quantum memristor system with optical quantum memristors at its core can be constructed to efficiently process complex multidimensional data tasks, providing a new approach for the integration of quantum computing and artificial intelligence.
[0037] Therefore, this application provides a technical solution for chaotic signal prediction based on a quantum memristor system. In this solution, the quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detector terminal. The method includes: generating a single-photon source signal through the input terminal and encoding the single-photon source signal into a first quantum state; converting the first quantum state into a second quantum state through the optical quantum memristors; obtaining density matrix information by detecting the second quantum state through the detector terminal; and analyzing the density matrix information using a preset signal prediction model to obtain a chaotic signal prediction result. By integrating quantum machine learning and optical quantum memristor technology, the quantum memristor system utilizes the superposition property of quantum states to efficiently capture and learn complex nonlinear characteristics in chaotic systems. Compared to classical memristor systems, it can effectively reduce prediction errors, making the prediction results closer to the actual values, improving the computational accuracy and efficiency of chaotic signal prediction tasks, breaking through performance limitations in high-dimensional data processing, and providing a new path for the implementation of artificial intelligence hardware. See below for details.
[0038] Example 1 Figure 1 A flowchart illustrating a chaotic signal prediction method based on a quantum memristor system according to Embodiment 1 of this application is shown.
[0039] like Figure 1 As shown, the quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detector terminal. The chaotic signal prediction method based on the quantum memristor system may include steps S101-S104, wherein: Step S101: Generate a single-photon source signal through the input terminal and encode the single-photon source signal into a first quantum state; In this embodiment, the input terminal includes a light source module and an information encoding module. The light source module generates a single-photon source signal, and the information encoding module maps the classical information (such as binary format information) of the single-photon source signal to a first quantum state. The encoded first quantum state is then input to an optical quantum memristor for processing.
[0040] Step S102: The first quantum state is converted to a second quantum state using the optical quantum memristor; In this embodiment, the optical quantum memristor module is composed of a Mach-Zehnder interferometer (MZI), which consists of a phase modulator and an optical fiber beam splitter (50:50). The MZI performs the function of the optical quantum memristor, processing the input first quantum state and outputting a second quantum state. The processed second quantum state can be transmitted through a single-mode optical fiber to the detection end for single-photon detection.
[0041] Step S103: Density matrix information is obtained by detecting the second quantum state through the detection end; In this embodiment, the detection end includes a polarization controller and a single-photon detector for receiving the second quantum state from the optical quantum memristor module. The single-photon detector records the detection event of the second quantum state to obtain density matrix information. Finally, the detection results can be transmitted to a classical computer for subsequent neural network training and analysis and modeling of chaotic signals.
[0042] Step S104: Analyze the density matrix information using a preset signal prediction model to obtain the turbidity signal prediction result.
[0043] In this embodiment, the preset signal prediction model is a pre-trained neural network model used to handle the prediction task of turbid signals.
[0044] In one optional embodiment of this application, the input terminal includes an information encoding module for encoding the single-photon source signal, the encoding process satisfying the following expression:
[0045]
[0046] in, This indicates the phase of the phase modulator in the information encoding module. Represents the first quantum state of the output; coefficients These are the complex elements of the transformation matrix T, representing the three photons appearing in the output channel respectively. The probability amplitude of this event, and matrix T satisfies ( (The identity matrix). Transform matrix elements. Phase parameters of all MZI It was jointly decided that, during the encoding process, classic information The phase parameter is set for all MZIs . This represents the creation operator for the i-th output channel. This represents the creation operator for the j-th output channel. This represents the creation operator for the k-th output channel. It is in a vacuum state.
[0047] In an optional embodiment of this application, the photonic quantum memristor is constructed using a Mach-Zehnder interferometer (MZI); the mathematical expression for the transformation of the first quantum state by the photonic quantum memristor is:
[0048] in, Represents the first quantum state. Represents the second quantum state. This is the unitary transformation matrix of the photonic quantum memristor, which depends on the state variables of the photonic quantum memristor. First quantum state and time .
[0049] In an optional embodiment of this application, the state variable R of the photonic quantum memristor satisfies the following expression:
[0050] One of the output ports in the MZI is defined as the feedback port. The expected number of photons measured at this feedback port. It represents the first derivative of the state variable R with respect to time t, that is, the rate of change of the state variable with time.
[0051] In an optional embodiment of this application, the state variable R is phase-controlled by the phase modulator in the MZI and is defined as:
[0052] in, This indicates the phase of the phase modulator in MZI.
[0053] In an optional embodiment of this application, the method further includes: Based on the measurement results from the feedback port, the phase of the phase modulator is updated to update the state variable R.
[0054] In this embodiment, the state update mechanism of the photonic quantum memristor is as follows: based on the measurement results of the MZI feedback port, the phase θ of the phase modulator is updated, thereby realizing the update of the state variable R.
[0055] In an optional embodiment of this application, the training process of the preset signal prediction model satisfies the following expression:
[0056] in, The actual output value vector of the training set samples. The input feature matrix of the training set samples, This is the weight matrix of the model.
[0057] In an optional embodiment of this application, the method further includes: The preset signal prediction model is tested, and the testing process satisfies the following expression:
[0058] in, This is the weight matrix of the model. The input feature matrix for the test set samples. This is the corresponding predicted output value.
[0059] In an optional embodiment of this application, the method further includes: The performance of the preset signal prediction model is evaluated according to the standardized mean square error loss function, which satisfies the following expression:
[0060] in, Indicates the first The actual output values of each test set sample. This represents the corresponding predicted output value. This is the sum of squares of the actual output values for all test set samples.
[0061] To make the chaotic signal prediction scheme based on the quantum memristor system in this embodiment easier to understand, an exemplary application is provided below. This exemplary application includes two parts: an experimental setup and a data post-processing method. The structural framework of the quantum memristor system in the experimental setup is as follows: Figure 2 As shown, it mainly consists of an input terminal, memristor modules A, B, and C, and a probe terminal D.
[0062] The input end includes a light source module and an information encoding module. Light sources 1, 2 and 3 generate single-photon sources, and information encoding module 4 maps classical information to the first quantum state, that is, generates a quantum state for processing chaotic signal prediction tasks.
[0063] The memristor module mainly comprises a phase modulator and an optical fiber beam splitter (BS 50:50). Memristor module A performs the function of an optical quantum memristor, processes the input first quantum state, and uses a single-photon detector 9 to measure the number of output photons at the feedback port to update the phase of the phase modulator 5, preparing for processing subsequent input first quantum states. The optical signal of the output second quantum state is transmitted to the detector. Memristor modules B and C have the same device architecture as memristor module A. Figure 3 (a) shows a schematic diagram of the structure of MZI, which consists of two beam splitters (50:50) and a phase modulator. Figure 3 (b) is a schematic diagram of the information encoding module. This structure realizes the quantum information encoding of three single-photon sources a, b, and c, and six output modes. The encoded first quantum state is transmitted to the quantum memristor system for processing.
[0064] Specifically, by adjusting Phase parameters of mid-phase modulator This allows for the encoding of classical information from three input single-photon sources. The encoded output quantum state can be represented as:
[0065] Among them, coefficient These are the complex elements of the transformation matrix T, representing the three photons appearing in the output channel respectively. The probability amplitude of this event, and matrix T satisfies ( (The identity matrix). This represents the creation operator for the i-th output channel. This represents the creation operator for the j-th output channel. This represents the creation operator for the k-th output channel. It is in a vacuum state.
[0066] Transformation matrix elements Phase parameters of all MZI Jointly decided. During the encoding process, classic information... The phase parameter is set for all MZIs The phase parameter With the classic information to be encoded The mapping relationship between them is defined as follows:
[0067] This information encoding module enables the transformation from classic information. To output quantum state The mapping completes the task of encoding quantum states.
[0068] Memristor modules A, B, and C output photons which are transmitted to detector D for detection by single-photon detectors 8-13. The measured single-photon detection information is sent to a classical computer for training and processing to complete the analysis and modeling of chaotic signals.
[0069] The analysis process for chaotic signal prediction tasks includes three main stages: dataset partitioning, neural network model training, and performance evaluation. The specific process is as follows: Figure 4 As shown. The goal of this task is to reach a known system time step. Based on previous values, the prediction system predicts at time step The output. System modeling is based on variables. The equations satisfied, where the input is defined as The target output is To this end, a dataset containing 2000 data points was constructed, each data point consisting of an input... With corresponding output Composition: The first 1000 data points are used as the training set, and the last 1000 data points are used as the test set.
[0070] First, the training set data is sequentially input into the information encoding module to complete quantum state encoding, and then input into the quantum memristor system for processing. Subsequently, the output quantum state is expanded based on the distribution of photons at different output ports of the quantum memristor system (represented by Fock states), generating all possible Fock states and constructing them as measurement basis vectors. The quantum information output by the quantum memristor system is subjected to quantum state tomography by single-photon detectors 8-13 to construct the output density matrix information, and the diagonal elements of the density matrix are extracted as input features of the training set. These features are then transmitted to a classical computer for further processing. Simultaneously, measurements are performed at the feedback port using single-photon detectors 9, 11, and 13. The number of photons within is used to update the phase of phase modulators 5, 6, and 7 in the memristor module, preparing for subsequent data processing. Specifically, when a single data point... After being input into the quantum memristor system, the data undergoes information encoding and processing via a quantum memristor network. Then, quantum state tomography extracts the diagonal elements of the output density matrix to generate eigenvectors. ).
[0071] In classical computers, supervised learning methods are used to train neural network models: the training set... Corresponding feature vector As input features, and the actual output value Pairing samples creates training samples used to optimize model parameters. This training process can be represented in matrix form as follows: =
[0072] in This is the actual output value vector of the training set. Input the feature matrix into the training set. This represents the weight matrix (i.e., the learnable parameters) of the neural network model. The core objective of the training process is to iteratively optimize the weight matrix. The goal is to minimize the error between the model's predicted values and the actual output values, thereby obtaining the optimized weight matrix.
[0073] The test set data is processed using the same procedure as the training set: First, the quantum state is encoded via an information encoding module, then processed by a quantum memristor system. Subsequently, quantum state tomography extracts the diagonal elements of the output density matrix as input features, which are then transmitted to a classical computer for further analysis. The feature vectors from the test set are input into the trained neural network model for prediction. This process can be represented as: =
[0074] in The optimized weight matrix, For the feature matrix of the test set, This is the corresponding predicted output value.
[0075] Model performance evaluation uses standardized mean squared error. As a primary indicator, it is defined as mean square error. Divide by the sum of squares of the actual output values as the normalization factor, and the specific calculation formula is as follows:
[0076] in Indicates the first The actual output values of each test set sample. The denominator represents the corresponding predicted value. This is the sum of squares of the actual output values for all test samples. This metric effectively quantifies the accuracy and robustness of quantum memristor systems in chaotic signal prediction tasks, providing a quantitative basis for system optimization.
[0077] In this embodiment, the quantum memristor system utilizes the superposition property of quantum states to efficiently capture and learn complex nonlinear features in chaotic systems. Compared to classical memristor systems, it effectively reduces prediction errors, making prediction results closer to actual values. Furthermore, this embodiment requires only a small number of memristors to achieve high prediction performance, reducing system size and hardware requirements. By increasing the number of single-photon sources, the system can encode and process higher-dimensional quantum states, thereby enhancing its ability to process high-dimensional chaotic information. This embodiment possesses the ability to efficiently mine deep feature correlations, extracting key features of chaotic signals with only a limited number of training samples to achieve high-precision predictions. This characteristic reduces dependence on large-scale datasets, improves data utilization efficiency, and makes the system more applicable in data-constrained scenarios.
[0078] Example 2 Figure 5 A block diagram schematically illustrates a chaotic signal prediction device based on a quantum memristor system according to Embodiment 2 of this application. This device can be divided into one or more program modules, which are stored in a storage medium and executed by one or more processors to complete the embodiments of this application. The program modules referred to in the embodiments of this application are a series of computer program instruction segments capable of performing specific functions. The following description will specifically introduce the functions of each program module in this embodiment. The quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detection terminal, such as... Figure 5 As shown, the chaotic signal prediction device 500 based on a quantum memristor system may include: an encoding module 501, a memristor module 502, a detection module 503, and a chaotic signal prediction module 504, wherein: Encoding module 501 is used to generate a single-photon source signal through the input terminal and encode the single-photon source signal into a first quantum state; Memristor module 502 is used to convert the first quantum state into a second quantum state through the photonic quantum memristor; The detection module 503 is used to obtain density matrix information by detecting the second quantum state through the detection end; The turbidity signal prediction module 504 is used to analyze the density matrix information through a preset signal prediction model to obtain the turbidity signal prediction result.
[0079] In one optional embodiment of this application, the input terminal includes an information encoding module for encoding the single-photon source signal, the encoding process satisfying the following expression:
[0080]
[0081] in, This indicates the phase of the phase modulator in the information encoding module. Represents the first quantum state of the output; coefficient These are the complex elements of the transformation matrix T, representing the three photons appearing in the output channel respectively. The probability amplitude of this event, and matrix T satisfies ( (the identity matrix); This represents the creation operator for the i-th output channel. This represents the creation operator for the j-th output channel. This represents the creation operator for the k-th output channel. It is in a vacuum state.
[0082] In an optional embodiment of this application, the photonic quantum memristor is constructed using a Mach-Zehnder interferometer (MZI); the mathematical expression for the transformation of the first quantum state by the photonic quantum memristor is:
[0083] in, Represents the first quantum state. Represents the second quantum state. This is the unitary transformation matrix of the photonic quantum memristor, which depends on the state variables of the photonic quantum memristor. First quantum state and time .
[0084] In an optional embodiment of this application, the state variable R of the photonic quantum memristor satisfies the following expression:
[0085] One of the output ports in the MZI is defined as the feedback port. The expected number of photons measured at this feedback port.
[0086] In an optional embodiment of this application, the state variable R is phase-controlled by the phase modulator in the MZI and is defined as:
[0087] in, This indicates the phase of the phase modulator in MZI.
[0088] In an optional embodiment of this application, the apparatus further includes: The phase update module is used to update the phase of the phase modulator based on the measurement results of the feedback port, so as to update the state variable R.
[0089] In an optional embodiment of this application, the training process of the preset signal prediction model satisfies the following expression:
[0090] in, This is the actual output value vector of the training set samples. The input feature matrix of the training set samples, This is the weight matrix of the model.
[0091] In an optional embodiment of this application, the apparatus further includes: The model testing module is used to test the preset signal prediction model. The testing process satisfies the following expression:
[0092] in, This is the weight matrix of the model. The input feature matrix for the test set samples. This is the corresponding predicted output value.
[0093] In an optional embodiment of this application, the apparatus further includes: The model evaluation module is used to evaluate the performance of the preset signal prediction model according to the standardized mean square error loss function, which satisfies the following expression:
[0094] in, Indicates the first The actual output values of each test set sample. This represents the corresponding predicted output value. This is the sum of squares of the actual output values for all test set samples.
[0095] It should be noted that the above are merely preferred embodiments of this application and do not limit the scope of patent protection of this application. Any equivalent structural or procedural changes made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of this application.
Claims
1. A method for predicting chaotic signals based on a quantum memristor system, characterized in that, The quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detector terminal; the method includes: A single-photon source signal is generated through the input terminal, and the single-photon source signal is encoded into a first quantum state; The first quantum state is transformed into the second quantum state using the optical quantum memristor; Density matrix information is obtained by detecting the second quantum state through the probe end; The density matrix information is analyzed using a preset signal prediction model to obtain the turbidity signal prediction result.
2. The chaotic signal prediction method based on a quantum memristor system according to claim 1, characterized in that, The input terminal includes an information encoding module for encoding the single-photon source signal. The encoding process satisfies the following expression: in, This indicates the phase of the phase modulator in the information encoding module. Represents the first quantum state of the output; coefficient These are the complex elements of the transformation matrix T, representing the three photons appearing in the output channel respectively. The probability amplitude of this event, and matrix T satisfies ( (the identity matrix); This represents the creation operator for the i-th output channel. This represents the creation operator for the j-th output channel. This represents the creation operator for the k-th output channel. It is in a vacuum state.
3. The chaotic signal prediction method based on a quantum memristor system according to claim 1, characterized in that, The optical quantum memristor is constructed using a Mach-Zehnder interferometer (MZI); the mathematical expression for the transformation of the first quantum state by the optical quantum memristor is as follows: in, Represents the first quantum state. Represents the second quantum state. This is the unitary transformation matrix of the photonic quantum memristor, which depends on the state variables of the photonic quantum memristor. First quantum state and time .
4. The chaotic signal prediction method based on a quantum memristor system according to claim 3, characterized in that, The state variable R of the optical quantum memristor satisfies the following expression [I1][A2]: One of the output ports in the MZI is defined as the feedback port. The expected number of photons measured at this feedback port.
5. The chaotic signal prediction method based on a quantum memristor system according to claim 3 or 4, characterized in that, The state variable R is phase-controlled by the phase modulator in the MZI and is defined as follows: in, This indicates the phase of the phase modulator in MZI.
6. The chaotic signal prediction method based on a quantum memristor system according to claim 5, characterized in that, The method further includes: Based on the measurement results from the feedback port, the phase of the phase modulator is updated to update the state variable R.
7. The chaotic signal prediction method based on a quantum memristor system according to claim 1, characterized in that, The training process of the preset signal prediction model satisfies the following expression: in, The actual output value vector of the training set samples. The input feature matrix of the training set samples, This is the weight matrix of the model.
8. The chaotic signal prediction method based on a quantum memristor system according to claim 7, characterized in that, The method further includes: The preset signal prediction model is tested, and the testing process satisfies the following expression: in, This is the weight matrix of the model. The input feature matrix for the test set samples. This is the corresponding predicted output value.
9. The chaotic signal prediction method based on a quantum memristor system according to claim 8, characterized in that, The method further includes: The performance of the preset signal prediction model is evaluated according to the standardized mean square error loss function, which satisfies the following expression: in, Indicates the first The actual output values of each test set sample. This represents the corresponding predicted output value. This is the sum of squares of the actual output values for all test set samples.
10. A chaotic signal prediction device based on a quantum memristor system, characterized in that, The quantum memristor system includes an input terminal, multiple optical quantum memristors, and a detector terminal. The device includes: An encoding module is used to generate a single-photon source signal through the input terminal and encode the single-photon source signal into a first quantum state; A memristor module is used to convert the first quantum state into a second quantum state through the optical quantum memristor; The detection module is used to obtain density matrix information by detecting the second quantum state through the detection terminal; The turbidity signal prediction module is used to analyze the density matrix information through a preset signal prediction model to obtain the turbidity signal prediction result.