CN115828999A - Quantum convolution neural network construction method and system based on quantum state amplitude transformation - Google Patents
Quantum convolution neural network construction method and system based on quantum state amplitude transformation Download PDFInfo
- Publication number
- CN115828999A CN115828999A CN202211292790.6A CN202211292790A CN115828999A CN 115828999 A CN115828999 A CN 115828999A CN 202211292790 A CN202211292790 A CN 202211292790A CN 115828999 A CN115828999 A CN 115828999A
- Authority
- CN
- China
- Prior art keywords
- quantum
- neural network
- layer
- convolutional
- convolutional neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of quantum model calculation, and particularly relates to a quantum convolutional neural network construction method and system based on quantum state amplitude transformation, wherein the properties of a convolutional layer and a pooling layer are expanded to a quantum domain according to the local connectivity and parameter sharing properties of the convolutional layer and the pooling layer in a convolutional neural network, and a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer is constructed; and generating a quantum data set according to the quantum bit cluster state, training the quantum convolution neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain an expected value in the training process. The invention can realize the construction of the quantum convolution neural network model by using the low-depth quantum circuit, so that the training efficiency and the convergence rate of the constructed network model can be greatly improved, and the application in the classification processing of quantum data and classical data is facilitated.
Description
Technical Field
The invention belongs to the technical field of quantum model calculation, and particularly relates to a quantum convolution neural network construction method and system based on quantum state amplitude transformation.
Background
Quantum machine learning library TensorFlow Quantum (TFQ) can be used for constructing classical or Quantum data's mixed Quantum classical model fast, supports high performance Quantum circuit simulator, its Quantum convolution neural network model, but has the following problem: the pure quantum model has long training time in the experiment and is difficult to converge; the quantum lines of the constructed model have more parameters and deeper line depth. Since it is still in the noisy mesoscale quantum (NISQ) era, the construction of quantum neural networks requires less quantum cost, for example: the line depth and the parameter number of the quantum parameterized line are consistent with the expected line depth and parameter number.
Disclosure of Invention
The invention provides a quantum convolution neural network construction method and system based on quantum state amplitude transformation, which can realize construction of a quantum convolution neural network model by using a low-depth quantum circuit, greatly improve the training efficiency and convergence rate of the constructed network model, and facilitate application in classification processing of quantum data and classical data.
According to the design scheme provided by the invention, a quantum convolution neural network construction method based on quantum state amplitude transformation is provided, and comprises the following contents:
expanding the attributes of the convolutional layer and the pooling layer to a quantum domain according to the local connectivity and the parameter sharing attributes of the convolutional layer and the pooling layer in the convolutional neural network, and constructing a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural features of an input quantum state through unitary transformation of quantum bits, and the quantum full-connection layer utilizes a quantum circuit to map the structural features to a corresponding label space;
and generating a quantum data set according to the quantum bit cluster state, training the quantum convolutional neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain an expected value in the training process.
As a quantum convolution neural network construction method based on quantum state amplitude transformation, further, the quantum convolution neural network is multiplied by unitary transformation
To indicate the manner in which, among others,
unitary matrix respectively representing quantum convolution layer, quantum pooling layer and quantum full-link layer, wherein L is not more than log 2 N,N＝2 n ,n∈N * Where l denotes the l-th quantum convolutional layer and pooling layer, and N is the number of quantum bits in the model. Theta represents the angle of rotation of the qubit about the Y-axis of the bloch sphere,
representing the network model output.
As the quantum convolution neural network construction method based on quantum state amplitude transformation, further, in unitary transformation in the quantum convolution neural network, a unitary transformation operation process in the quantum convolution neural network is constructed through the addition and/or repetition operation of a single-quantum-bit revolving gate and a double-quantum-bit controlled NOT gate.
As a quantum convolution neural network construction method based on quantum state amplitude transformation, further, a quantum convolution layer performs unitary computation W through two quantum bits l To extract the input data features, wherein the unitary operation process is expressed as:
U ent representing a dual qubit controlled not gate operation, U representing a single qubit rotation gate, n representing the number of qubits, and θ representing the angle of rotation of the qubits about the X-axis of the bloch sphere.
As the quantum convolution neural network construction method based on quantum state amplitude transformation, further, the quantum pooling layer completes the projection of characteristic multi-quantum bit to single-quantum bit by unitary transformation of two quantum bits, wherein the unitary transformation process is expressed as:
as the quantum convolution neural network construction method based on quantum state amplitude transformation, further, a quantum full-connection layer adopts a function
And mapping the quantum circuit structure with the effect, wherein x is the final output of the convolutional-stacking pooling layer, b is a preset deviation value, and w is a parameter trained by adjusting the rotation angle.
As the quantum convolution neural network construction method based on the quantum state amplitude transformation, furthermore, in a quantum full-connection layer, a quantum circuit utilizes unitary transformation
To perform an input-output mapping in which,
n represents the number of qubits and CNOT represents a double-qubit controlled not gate.
As the quantum convolution neural network construction method based on quantum state amplitude transformation, further, a full connection layer in the quantum convolution neural network obtains quantum bit expected values by measuring quantum bits in corresponding quantum lines, and maps the expected values to classification labels.
As the quantum convolution neural network construction method based on quantum state amplitude transformation, further, a quantum data set is generated according to a quantum bit cluster state, firstly, a controlled Z gate operation is executed on adjacent quantum bits to generate a correct cluster state on the quantum bits; then, rotating the quantum bit around the X axis of the Bloch sphere to simulate the error in the cluster state, and generating the error cluster state of the quantum bit; then, a number of quantum states conforming to the magnitude of the training quantities are randomly generated.
Further, the present invention also provides a quantum convolution neural network construction system based on quantum state amplitude transformation, including: a model building module and a model training module, wherein,
the model building module is used for expanding the attributes of the convolutional layer and the pooling layer to a quantum domain according to the local connectivity and the parameter sharing characteristics of the convolutional layer and the pooling layer in the convolutional neural network CNN and building a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural features of an input quantum state through unitary transformation of quantum bits, and the quantum full-connection layer maps the structural features to a corresponding label space by using quantum circuits;
and the model training module is used for generating a quantum data set according to the quantum bit cluster state, training the quantum convolutional neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain the expected value in the training process.
The invention has the beneficial effects that:
the invention constructs the quantum convolutional neural network ATQCNN based on the low-depth quantum circuit, realizes extraction and classification application of quantum state input data through a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer in the network, and improves the training efficiency and the convergence speed of a pure quantum convolutional neural network model. And further through experimental result verification, compared with the QCNN model of Google, the method can reduce the parameter quantity by about 30%, has faster and more stable convergence, and improves the training efficiency by 35%, so that the model has the advantages of reducing the quantum cost overhead, such as circuit depth and parameter quantity, being suitable for processing the classification problems of quantum data and classical data, such as image recognition, quantum state classification, malicious code detection and the like, and providing possibility for the application of a quantum computer in the NISQ era.
Description of the drawings:
FIG. 1 is a schematic diagram of a construction process of a quantum convolution neural network in an embodiment;
FIG. 2 is a schematic diagram of an 8-qubit ATQCNN model architecture in an embodiment;
FIG. 3 is a circuit structure diagram of a quantum fully-connected layer in an embodiment;
FIG. 4 is a graph showing the loss function, test accuracy and training time of the sample during the training and testing process in the example.
The specific implementation mode is as follows:
in order to make the objects, technical solutions and advantages of the present invention clearer and more obvious, the present invention is further described in detail below with reference to the accompanying drawings and technical solutions.
The embodiment of the invention, as shown in fig. 1, provides a quantum convolution neural network construction method based on quantum state amplitude transformation, including:
s101, expanding the attributes of a convolutional layer and a pooling layer to a quantum domain according to the local connectivity and parameter sharing attributes of the convolutional layer and the pooling layer in the convolutional neural network, and constructing a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural features of an input quantum state through unitary transformation of quantum bits, and the quantum full-connection layer maps the structural features to a corresponding label space by using quantum circuits;
s102, generating a quantum data set according to the quantum bit cluster state, training the quantum convolution neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain an expected value in the training process.
Unlike other neural networks, CNNs are composed of multiple convolutional and pooling layers. These two unique layers have the property of local connectivity and parameter sharing, allowing the CNN to extract structured features with relatively few parameters. In convolutional layers, each neuron is connected to only a portion of the input neurons, and this local connection ensures that the learned filter can respond positively to local input features. The filters used in the neuron computation share the same depth, which can significantly reduce the parameters that need to be solved. In the embodiment of the scheme, the key attributes are extended to a quantum domain to construct an ATQCNN model. The input of the ATQCNN is quantum state data, after characteristics are extracted layer by layer, an expected value is obtained through quantum measurement of specific quantum bits, and a loss function is calculated; optimization updates are then performed until the appropriate parameters are learned so that the encoded quantum states can be correctly mapped to the corresponding labels.
The framework of the 8-qubit ATQCNN model shown in FIG. 2 is (a) a framework structure schematic, which is composed of three quantum convolution, pooling layers and a quantum full-link layer; (b) Quantum wires, quantum convolutional layers and quantum well layers, where the subscripts of W and V indicate which two qubits the unitary transform acts on. (c) A two-quantum bit unitary transform for the quantum convolution layer and the quantum pool layer, respectively. The model can be written as the product of a unitary transform:
wherein
Unitary matrices of quantum convolutional layers, quantum pooling layers, and quantum fully-connected layers, respectively. Specifically, if the input quantum state is | q 0 q 1 ...q k >After one convolution pooling operation, the number of quantum bits is halved:
after l convolution pooling operations:
wherein the content of the first and second substances,
k is an odd number.
As a preferred embodiment, further, in the unitary transformation in the quantum convolutional neural network, the unitary transformation operation process in the quantum convolutional neural network is constructed by the additional and/or repeated operations of the single-quantum-bit rotating gate and the double-quantum-bit controlled not gate.
Unitary transforms for quantum convolutional layers and quantum pooling layers can be constructed by appending and repeating in { RY, CNOT }. RY is a single-quantum bit-rotation gate (rotation around the Y axis of the Bloch sphere), while CNOT is a double-quantum bit-controlled NOT gate. After the RY gate operation is performed, only the amplitude of the quantum state changes. Thus, the entire bloch sphere needs to be explored initially to obtain the solution vector, but now only the surface around the y-axis needs to be explored. Furthermore, CNOT gates are used to build the entanglement relationship between qubits. Low depth circuits with highly entangled states have potential advantages in capturing the relationship between quantum data and data classification.
The main purpose of convolutional layers is to extract features from the input data, while quantum convolution has the advantage of enhancing the mapping. FIG. 2 (b) shows that the quantum convolution layer is a unitary operation W using two qubits l Implemented, where l denotes the l-th quantum convolutional layer, can be written as the product of a unitary transform:
wherein, U ent ＝CNOT,
This unitary operation acts on adjacent qubits and all applied unitary operations in one layer have the same parameters, which reflects the same two-feature local connectivity and parameter sharing as in classical CNNs. The dashed box is repeated a number of times to increase the depth of the layer and thus the number of parameters.
Quantum pooling layer unitary transform V by using two qubits l Where l denotes the l-th quantum pooling layer, the product of the unitary transform can be written:
it allows information to be extracted from two quantum ratiosThe bits are projected onto a single qubit to achieve the effect of reducing the feature map dimension. As shown in FIG. 2 (c), the unitary operation of two qubits is applied to m and m + k/2 l Where k is the number of quantum bits contained in the model and m is the mth quantum bit. Like the quantum convolutional layer, it has the same parameters within the layer.
After convolution and pooling layers, the dimensionality of the data is reduced. And the quantum full-connection layer maps the remaining bit reserved characteristic information to a corresponding sample label space. The quantum fully connected layer circuit may be as shown in fig. 3. Based on the definition of the classical full connection layer, the function can be realized by utilizing
An effective quantum wire structure. The input x is the output of the final convolution pooling layer and b is the set offset value. The parameter training of w and b is done by adjusting the angle of the RY gate, where the dimension of the output measurement is the same as the dimension of the input x. It can be written as the product of a unitary transform:
wherein, the first and the second end of the pipe are connected with each other,
finally, a measurement is made of a particular qubit and the obtained expected value is mapped to a class label.
As a preferred embodiment, further, in generating the quantum data set and the quantum state according to the qubit cluster state, first, a controlled Z-gate operation is performed on adjacent qubits to generate the correct cluster state on the qubit; then, rotating the quantum bit around the X axis of the Bloch sphere to simulate the error in the cluster state, and generating the error cluster state of the quantum bit; then, a number of quantum states conforming to the magnitude of the training quantities are randomly generated.
Further, based on the above method, an embodiment of the present invention further provides a quantum convolution neural network construction system based on quantum state amplitude transformation, including: a model building module and a model training module, wherein,
the model building module is used for expanding the attributes of the convolutional layer and the pooling layer to a quantum domain according to the local connectivity and the parameter sharing characteristics of the convolutional layer and the pooling layer in the convolutional neural network CNN and building a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural features of an input quantum state through unitary transformation of quantum bits, and the quantum full-connection layer maps the structural features to a corresponding label space by using quantum circuits;
and the model training module is used for generating a quantum data set according to the quantum bit cluster state, training the quantum convolutional neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain the expected value in the training process.
In order to verify the validity of the scheme of the present application, the validity of the ATQCNN model constructed in the embodiment of the present application is evaluated by combining test data and training efficiency and accuracy as follows:
the TensorFlow Quantum framework developed by Google was mainly used for the experiments. In having an Inter R Core TM Our experiments were performed on PCs with i7-8700 CPU (3.2 GHz) and 32GB RAM. The software environment is Python 3.6 in the windows 10 system. The experimental contents are as follows: it is discriminated whether or not the quantum cluster state is excited.
A quantum data set consists of a collection of cluster states that are correctly and incorrectly prepared on 8 qubits. First, the correct cluster state is generated by performing a CZ gate operation on adjacent qubits. Second, the error in the constellation state is modeled by rotating the qubit by an amount 0 ≦ θ ≦ 2 π around the X-axis of the Bloch sphere.
Considered as excitations and marked as 1, otherwise-1. 400 quantum states were randomly generated for the experiment.
Pure quantum CNN of ATQCNN and google were compared. Both models are 8-qubit hierarchies, with 3 repetitions of the application of quantum convolution and quantum pooling. In contrast, ATQCNN has one more quantum fully connected layer compared to google's pure quantum CNN. Google's pure quantum CNN contains 63 parameters that need to be trained, whereas ATQCNN contains only 44 parameters. Finally, the Pauli-Z expectation of the last qubit is measured using the mean square error between the model output and the tag as a cost function.
Under the same experimental conditions, the training epochs was set to 25, the batch size was set to 16, and the learning rate was 0.025. The loss function curve, the test accuracy and the training time of each sample in the training and testing process are shown in fig. 4, (a) is the loss function curve of two models in the training and testing process; (b) is the test precision; (c) average training time per sample per epoch round.
According to the experimental result, the loss function has an obvious downward trend and finally tends to be stable and converges at about 0.2. The convergence speed of the ATQCNN model is higher in the training process, the convergence is achieved on the average in the fifth epoch, and the accuracy is up to 100%. The average training time of a single sample of the Google model on the simulation processor is about 55ms, and the atqcnn model can be improved by about 40% to about 35ms. The conclusion is drawn that the ATQCNN model has higher convergence speed, shorter training time and lower accuracy. This result has two reasons. On the one hand, the parameters of the ATQCNN model are approximately one-third less than the Google model. On the other hand, the quantum circuit of Google simultaneously changes the phase and amplitude of the quantum state. However, the ATQCNN only changes the amplitude, reduces the space for exploring a solution set, improves the training efficiency, is similar to space time change, reduces the quantum cost overhead, provides possibility for the application of a quantum computer in the NISQ era, and can be further researched so as to fully exert the quantum advantages.
Unless specifically stated otherwise, the relative steps, numerical expressions, and values of the components and steps set forth in these embodiments do not limit the scope of the present invention.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.
The elements of the various examples and method steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and the components and steps of the examples have been described in a functional generic sense in the foregoing description for clarity of hardware and software interchangeability. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
Those skilled in the art will appreciate that all or part of the steps of the above methods may be implemented by instructing the relevant hardware through a program, which may be stored in a computer-readable storage medium, such as: read-only memory, magnetic or optical disk, and the like. Alternatively, all or part of the steps of the foregoing embodiments may also be implemented by using one or more integrated circuits, and accordingly, each module/unit in the foregoing embodiments may be implemented in the form of hardware, and may also be implemented in the form of a software functional module. The present invention is not limited to any specific form of combination of hardware and software.
Finally, it should be noted that: although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the following descriptions are only illustrative and not restrictive, and that the scope of the present invention is not limited to the above embodiments: those skilled in the art can still make modifications or changes to the embodiments described in the foregoing embodiments, or make equivalent substitutions for some features, within the scope of the disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the embodiments of the present invention, and they should be construed as being included therein. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. A quantum convolution neural network construction method based on quantum state amplitude transformation is characterized by comprising the following contents:
expanding the attributes of the convolutional layer and the pooling layer to a quantum domain according to the local connectivity and parameter sharing attributes of the convolutional layer and the pooling layer in the convolutional neural network, and constructing a quantum convolutional neural network consisting of a quantum convolutional layer, a quantum pooling layer and a quantum full-connection layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural features of input quantum states through unitary transformation of quantum bits, and the quantum full-connection layer maps the structural features to corresponding label spaces by using quantum lines;
and generating a quantum data set according to the quantum bit cluster state, training the quantum convolutional neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain an expected value in the training process.
2. The method for constructing quantum convolutional neural network based on quantum state amplitude transformation as claimed in claim 1, wherein the quantum convolutional neural network is multiplied by unitary transformation
To indicate the manner in which, among others,
unitary matrix respectively representing quantum convolution layer, quantum pooling layer and quantum full-link layer, wherein L is not more than log 2 N,N＝2 n ,n∈N * Where l denotes the first quantum convolution layer and pooling layer, N is the number of quantum bits in the model, θ denotes the angle of rotation of the quantum bits about the Y-axis of the Bloch sphere,
representing the network model output.
3. The quantum convolutional neural network construction method based on quantum state amplitude transformation as claimed in claim 1 or 2, wherein in the unitary transformation in the quantum convolutional neural network, the unitary transformation operation process in the quantum convolutional neural network is constructed by the additional and/or repeated operation of a single-quantum-bit rotating gate and a double-quantum-bit controlled non-gate.
4. The method of claim 1, wherein the quantum convolutional neural network is constructed by performing a unitary operation on the quantum convolutional layer through two qubits l To extract the input data features, wherein the unitary operation process is expressed as:
U ent representing a dual qubit controlled not gate operation, U representing a single qubit rotation gate, n representing the number of qubits, and θ representing the angle of rotation of the qubits about the Y-axis of the bloch sphere.
5. The method for constructing the quantum convolutional neural network based on quantum state amplitude transform of claim 4, wherein the quantum pooling layer completes the projection of multi-qubit to single-qubit of the feature by unitary transform of two qubits, wherein the unitary transform process is expressed as:
6. the quantum convolutional neural network construction method based on quantum state amplitude transformation as claimed in claim 1 or 4 or 5, wherein the quantum fully-connected layer adopts a function
A quantum wire structure with mapping effect, wherein x is the final output of the convolutional laminated pooling layer, b is a preset deviation value, and w is the deviation value obtained by adjusting the rotation angleParameters for training.
7. The method of claim 6, wherein the quantum convolution neural network is constructed by using unitary transformation for quantum circuit in the quantum fully-connected layer
To perform an input-output mapping in which,
n represents the number of qubits and CNOT represents a double-qubit controlled not gate.
8. The method as claimed in claim 1, wherein the fully connected layer in the quantum convolutional neural network obtains expected values of the qubits by measuring the qubits in the corresponding quantum wires and maps the expected values to the class labels.
9. The method of claim 1, wherein the quantum data set is generated according to the qubit cluster state, and first, a controlled Z-gate operation is performed on adjacent qubits to generate the correct cluster state on the qubit; then, the quantum bit is rotated around the X axis of the Bloch sphere to simulate the error in the cluster state, and the error cluster state of the quantum bit is generated; then, a number of quantum states conforming to the magnitude of the training quantities are randomly generated.
10. A quantum convolution neural network construction system based on quantum state amplitude transformation is characterized by comprising: a model building module and a model training module, wherein,
the model building module is used for expanding the attributes of the convolutional layer and the pooling layer to a quantum domain according to the local connectivity and the parameter sharing characteristics of the convolutional layer and the pooling layer in the convolutional neural network CNN and building a quantum convolutional neural network consisting of the quantum convolutional layer, the quantum pooling layer and a quantum fully-connected layer, wherein the quantum convolutional layer and the quantum pooling layer extract the structural characteristics of an input quantum state through unitary transformation of quantum bits, and the quantum fully-connected layer maps the structural characteristics to a corresponding label space by using quantum circuits;
and the model training module is used for generating a quantum data set according to the quantum bit cluster state, training the quantum convolutional neural network by using the generated quantum data set, and taking the mean square error between the model output and the label as a cost function for measuring the quantum bit to obtain the expected value in the training process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211292790.6A CN115828999B (en) | 2022-10-21 | 2022-10-21 | Quantum convolution neural network construction method and system based on quantum state amplitude transformation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211292790.6A CN115828999B (en) | 2022-10-21 | 2022-10-21 | Quantum convolution neural network construction method and system based on quantum state amplitude transformation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115828999A true CN115828999A (en) | 2023-03-21 |
| CN115828999B CN115828999B (en) | 2023-09-19 |
Family
ID=85525177
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211292790.6A Active CN115828999B (en) | 2022-10-21 | 2022-10-21 | Quantum convolution neural network construction method and system based on quantum state amplitude transformation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115828999B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117669753A (en) * | 2024-01-31 | 2024-03-08 | 北京航空航天大学杭州创新研究院 | Quantum model training method, multi-mode data processing method and device |
| CN117710513A (en) * | 2024-02-06 | 2024-03-15 | 中国科学院深圳先进技术研究院 | Quantum convolution neural network-based magnetic resonance imaging method and device |
| CN117710513B (en) * | 2024-02-06 | 2024-06-11 | 中国科学院深圳先进技术研究院 | Quantum convolution neural network-based magnetic resonance imaging method and device |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110692067A (en) * | 2017-06-02 | 2020-01-14 | 谷歌有限责任公司 | Quantum neural network |
| US20210182728A1 (en) * | 2018-08-30 | 2021-06-17 | Google Llc | Two-qubit gates implemented with a tunable coupler |
| CN113361664A (en) * | 2021-08-10 | 2021-09-07 | 北京航空航天大学 | Image recognition system and method based on quantum convolution neural network |
| CN113792870A (en) * | 2021-10-29 | 2021-12-14 | 广西大学 | All-optical control optical neural network system based on acousto-optic technology |
| US20220149823A1 (en) * | 2020-11-12 | 2022-05-12 | equal1.labs Inc. | System and method of generating quantum unitary noise using silicon based quantum dot arrays |
| CN114821217A (en) * | 2021-06-28 | 2022-07-29 | 合肥本源量子计算科技有限责任公司 | Image identification method and device based on quantum classical hybrid neural network |
-
2022
- 2022-10-21 CN CN202211292790.6A patent/CN115828999B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110692067A (en) * | 2017-06-02 | 2020-01-14 | 谷歌有限责任公司 | Quantum neural network |
| US20210182728A1 (en) * | 2018-08-30 | 2021-06-17 | Google Llc | Two-qubit gates implemented with a tunable coupler |
| US20220149823A1 (en) * | 2020-11-12 | 2022-05-12 | equal1.labs Inc. | System and method of generating quantum unitary noise using silicon based quantum dot arrays |
| CN114821217A (en) * | 2021-06-28 | 2022-07-29 | 合肥本源量子计算科技有限责任公司 | Image identification method and device based on quantum classical hybrid neural network |
| CN113361664A (en) * | 2021-08-10 | 2021-09-07 | 北京航空航天大学 | Image recognition system and method based on quantum convolution neural network |
| CN113792870A (en) * | 2021-10-29 | 2021-12-14 | 广西大学 | All-optical control optical neural network system based on acousto-optic technology |
Non-Patent Citations (4)
| Title |
|---|
| IRIS CONG等: "Quantum convolutional neural networks", 《NATURE PHYSICS》, pages 1273 - 1278 * |
| SHIJIE WEI等: "A quantum convolutional neural network on NISQ devices", 《AAPPS BULLETIN》, pages 1 - 11 * |
| 刘兴奥等: "量子线性卷积及其在图像处理中的应用", 《自动化学报》, vol. 48, no. 6, pages 1504 - 1519 * |
| 范兴奎等: "基于量子卷积神经网络的图像识别新模型", 《电子科技大学学报》, vol. 51, no. 5, pages 641 - 650 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117669753A (en) * | 2024-01-31 | 2024-03-08 | 北京航空航天大学杭州创新研究院 | Quantum model training method, multi-mode data processing method and device |
| CN117669753B (en) * | 2024-01-31 | 2024-04-16 | 北京航空航天大学杭州创新研究院 | Quantum model training method, multi-mode data processing method and device |
| CN117710513A (en) * | 2024-02-06 | 2024-03-15 | 中国科学院深圳先进技术研究院 | Quantum convolution neural network-based magnetic resonance imaging method and device |
| CN117710513B (en) * | 2024-02-06 | 2024-06-11 | 中国科学院深圳先进技术研究院 | Quantum convolution neural network-based magnetic resonance imaging method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115828999B (en) | 2023-09-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113159239B (en) | Method for processing graph data by quantum graph convolutional neural network | |
| US9836698B2 (en) | Method and system for decomposing single-qubit quantum circuits into a discrete basis | |
| US20190354316A1 (en) | Effective Quantum RAM Architecture for Quantum Database | |
| CN115828999A (en) | Quantum convolution neural network construction method and system based on quantum state amplitude transformation | |
| CN114358319A (en) | Machine learning framework-based classification method and related device | |
| CN114358216B (en) | Quantum clustering method based on machine learning framework and related device | |
| JP7404597B2 (en) | Molecular structure acquisition method, apparatus, electronic device and storage medium | |
| DE102019130311A1 (en) | TRANSPOSED WEAK MATRIX MULTIPLICATED WITH DENSE MATRIX FOR TRAINING NEURONAL NETWORKS | |
| CN114676837A (en) | Evolutionary quantum neural network architecture searching method based on quantum simulator | |
| CN113836319A (en) | Knowledge completion method and system for fusing entity neighbors | |
| EP4242936A1 (en) | Reducing resources in quantum circuits | |
| CN114511092B (en) | Graph attention machine mechanism implementation method based on quantum circuit | |
| CN113222160A (en) | Quantum state conversion method and device | |
| CN114446412A (en) | Molecular generation method and device based on quantum gate control recurrent neural network | |
| CN114974405A (en) | Quantum GNN-based binding energy prediction method | |
| CN114550849A (en) | Method for solving chemical molecular property prediction based on quantum graph neural network | |
| CN113780518B (en) | Network architecture optimization method, terminal equipment and computer readable storage medium | |
| CN115130655A (en) | Method for solving product reaction center prediction in inverse synthesis | |
| CN114548412A (en) | Quantum convolution model, quantum circuit and construction method thereof | |
| Stein | Quantum Computing Aided Machine Learning Through Quantum State Fidelity | |
| CN114282000A (en) | Text clustering method, text clustering device, text clustering medium and electronic device based on quantum computation | |
| CN117875370A (en) | Task processing method and device utilizing molecular data | |
| CN118038000A (en) | Voxel connected domain dividing method, device, equipment and storage medium | |
| CN117610671A (en) | Model training method, device, equipment and medium for generating quantum state | |
| CN117610670A (en) | Model training method, device, equipment and storage medium for quantum state generation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |