Federal learning-based noise identification method and system

Abstract

The invention relates to a federated learning-based noise recognition method and system, and the method comprises the steps: firstly distributing a target global noise recognition model, and obtaining a text feature corresponding to a local text, and a trained model parameter; secondly, clustering the customer service nodes through the text features, and classifying the customer service nodes with similar text feature similarity into the same cluster, so that the distribution difference of the text features of the customer service nodes in the same cluster is relatively small; and then intra-cluster and inter-cluster aggregation is carried out, and finally continuous iteration can be carried out on the target global noise identification model, so that the noise identification reliability of the model can be continuously improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of text noise recognition, and in particular, to a noise recognition method and system based on federated learning. Background Art

[0002] With the rapid development of the Internet and the improvement of people's service awareness, online customer service has been popularized in all walks of life and penetrated into every aspect of daily business services.

[0003] Online customer service needs to give accurate responses based on customer information. If irrelevant content (text noise) appears in the response text, it will affect the matching response items.

[0004] The existing method is to build a noise recognition model, and then identify the noise of the text of customer information based on the recognition model, so as to remove the noise impact and improve the satisfaction with the information reply. However, with the diversification of customer information text and the processing of a large amount of user personal information, the reliability of the recognition model has been challenged, and the iterative upgrade of the recognition model has become a pain point that the industry needs to solve urgently. Summary of the invention

[0005] The following is a summary of the subject matter described in detail herein. This summary is not intended to limit the scope of the claims.

[0006] The main purpose of the embodiments of the present disclosure is to propose a noise recognition method and system based on federated learning, so as to continuously improve the reliability of the model in recognizing noise.

[0007] A first aspect of an embodiment of the present application proposes a noise recognition method based on federated learning, the method comprising:

[0008] Sending a target global noise recognition model to M customer service nodes so that each of the customer service nodes performs a model training process; M is a positive integer greater than 1, and the model training process includes:

[0009] Performing noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and trained model parameters;

[0010] Perform clustering, inter-cluster aggregation and intra-cluster aggregation processes, where:

[0011] The clustering process includes:

[0012] receiving text features of the M customer service nodes, and dividing the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of the customer service nodes; N is a positive integer greater than 1, and M is greater than N;

[0013] The inter-cluster aggregation process includes:

[0014] Receiving the model parameters of all the customer service nodes corresponding to each cluster, performing intra-cluster model aggregation according to the model parameters, and obtaining an intra-cluster noise recognition model corresponding to each cluster;

[0015] The intra-cluster aggregation process includes:

[0016] Perform inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model;

[0017] In the case that the iterated target global noise recognition model does not meet the preset requirements, the iterated target global noise recognition model is used as the target global noise recognition model and sent to all customer service nodes, and the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process are performed.

[0018] This application proposes a noise identification method based on federated learning, which has at least the following beneficial effects:

[0019] This method first distributes the target global noise recognition model to obtain the text features corresponding to the local text and the model parameters after training; then the customer service nodes are clustered by text features, and a class of customer service nodes with relatively similar text features can be classified into the same cluster, so that the text feature distribution differences of customer service nodes in the same cluster are relatively small; then intra-cluster and inter-cluster aggregation are performed, and finally the target global noise recognition model can be continuously iterated, thereby continuously improving the model's recognition reliability of noise.

[0020] In some implementations, dividing the M customer service nodes into N clusters according to the text features includes:

[0021] Based on the K nearest neighbor classification algorithm, the M customer service nodes are divided into N clusters according to the text features.

[0022] In some implementations, the aggregating the intra-cluster models according to the model parameters to obtain the intra-cluster noise recognition model corresponding to each cluster includes:

[0023]

[0024]

[0025] in, Cluster A j The intra-cluster noise recognition model, F i is the target global noise recognition model for customer service node i, θ i is the model parameter of customer service node i, ωi is the weight value of customer service node i, S i is the data volume of the local text of customer service node i, Cluster A j The total amount of local text data.

[0026] In some implementations, performing inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to all clusters to obtain an iterative global noise recognition model includes:

[0027]

[0028]

[0029] Among them, F is the iterative global noise recognition model, Cluster C j The weight value, S total The amount of local text data for all customer service nodes.

[0030] In some implementations, after performing inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model, the method further includes:

[0031] When the iterated target global noise recognition model meets the preset requirement, the iterated target global noise recognition model is sent to each of the customer service nodes.

[0032] In some embodiments, the target global noise recognition model is a model constructed by a graph neural network model.

[0033] A second aspect of the embodiment of the present application proposes a noise recognition system based on federated learning, the system comprising:

[0034] A model training module is used to send a target global noise recognition model to M customer service nodes so that each of the customer service nodes performs a model training process; M is a positive integer greater than 1, and the model training process includes:

[0035] Performing noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and trained model parameters;

[0036] The model aggregation module is used to perform clustering, inter-cluster aggregation and intra-cluster aggregation processes, where:

[0037] The clustering process includes:

[0038] receiving text features of the M customer service nodes, and dividing the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of the customer service nodes; N is a positive integer greater than 1, and M is greater than N;

[0039] The inter-cluster aggregation process includes:

[0040] Receiving the model parameters of all the customer service nodes corresponding to each cluster, performing intra-cluster model aggregation according to the model parameters, and obtaining an intra-cluster noise recognition model corresponding to each cluster;

[0041] The intra-cluster aggregation process includes:

[0042] Perform inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model;

[0043] The model iteration judgment module is used to use the iterated target global noise recognition model as the target global noise recognition model and send it to all customer service nodes when the iterated target global noise recognition model does not meet the preset requirements, and execute the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process.

[0044] A third aspect of an embodiment of the present application proposes an electronic device, comprising at least one controller and a memory for communicating with the controller; the memory stores instructions that can be executed by the at least one controller, and the instructions are executed by the at least one controller so that the at least one controller performs the noise recognition method based on federated learning as described above.

[0045] A fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed, the noise recognition method based on federated learning as described above is implemented.

[0046] A fifth aspect of an embodiment of the present application proposes a computer program product. When instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the noise recognition method based on federated learning as described above.

[0047] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application.

[0048] It can be understood that the beneficial effects of the second to fifth aspects compared with the related art are the same as the beneficial effects of the first aspect compared with the related art. Please refer to the relevant description in the first aspect, and no further details will be given here. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0050] Figure 1 It is a flowchart of the noise identification method based on federated learning provided in this application;

[0051] Figure 2 It is a structural diagram of a noise identification system based on federated learning provided in this application;

[0052] Figure 3 It is a structural schematic diagram of the electronic device provided by this application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0056] With the rapid development of the Internet and the improvement of people's service awareness, online customer service has been popularized in all walks of life and penetrated into every aspect of daily business services.

[0057] Online customer service needs to give accurate responses based on customer information. If irrelevant content (text noise) appears in the response text, it will affect the matching response items.

[0058] The existing method is to build a noise recognition model, and then identify the noise of the text of customer information based on the recognition model, so as to remove the noise impact and improve the satisfaction with the information reply. However, with the diversification of customer information text and the processing of a large amount of user personal information, the reliability of the recognition model has been challenged, and the iterative upgrade of the recognition model has become a pain point that the industry needs to solve urgently.

[0059] An embodiment of the present application provides a noise recognition method based on federated learning, the method comprising:

[0060] Step S110: the server sends the target global noise recognition model to M customer service nodes, so that each customer service node performs a model training process; M is a positive integer greater than 1, and the model training process includes:

[0061] Perform noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and model parameters after training;

[0062] Step S120, performing clustering, inter-cluster aggregation and intra-cluster aggregation processes, wherein:

[0063] The clustering process includes:

[0064] The server receives text features of M customer service nodes, and divides the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of customer service nodes; N is a positive integer greater than 1, and M is greater than N;

[0065] The inter-cluster aggregation process includes:

[0066] The server receives the model parameters of all the customer service nodes corresponding to each cluster, aggregates the models within the cluster according to the model parameters, and obtains the cluster noise recognition model corresponding to each cluster;

[0067] The intra-cluster aggregation process includes:

[0068] The server aggregates the inter-cluster models according to the intra-cluster noise recognition models corresponding to the N clusters to obtain the iterative target global noise recognition model;

[0069] Step S130, when the iterated target global noise recognition model does not meet the preset requirements, the iterated target global noise recognition model is used as the target global noise recognition model and sent to all customer service nodes, and the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process are performed.

[0070] In this embodiment, the customer service node can be a connection node corresponding to a customer service center in each region, such as a connection node of a customer service center in Nanning, Guangxi. The customer service node has a communication device for communicating with a server. The target global noise recognition model is issued by the server.

[0071] Local text is the local text converted by the voice-to-text conversion technology after the connection node receives the customer's voice.

[0072] This application has the following beneficial effects:

[0073] This method first distributes the target global noise recognition model to obtain the text features corresponding to the local text and the model parameters after training; then the customer service nodes are clustered by text features, and a class of customer service nodes with relatively similar text features can be classified into the same cluster, so that the text feature distribution differences of customer service nodes in the same cluster are relatively small; then intra-cluster and inter-cluster aggregation are performed, and finally the target global noise recognition model can be continuously iterated, thereby continuously improving the model's recognition reliability of noise.

[0074] In some embodiments of the present application, M customer service nodes are divided into N clusters according to text features, including:

[0075] Based on the K nearest neighbor classification algorithm, M customer service nodes are divided into N clusters according to text features.

[0076] The K nearest neighbor classification algorithm is in the feature space. If most of the k nearest samples (i.e. the nearest samples in the feature space) near a sample belong to a certain category, then the sample also belongs to this category. The K nearest neighbor classification algorithm can be used to divide a class of customer service nodes with similar text features into the same cluster, so that the distribution difference of text features in the same cluster is small.

[0077] In some embodiments of the present application, cluster model aggregation is performed according to model parameters to obtain a cluster noise recognition model corresponding to each cluster, including:

[0078]

[0079]

[0080] in, Cluster A j The intra-cluster noise recognition model, F i is the target global noise recognition model for customer service node i, θ i is the model parameter of customer service node i, ω i is the weight value of customer service node i, S i is the data volume of the local text of customer service node i, Cluster A j The total amount of local text data.

[0081] In this embodiment, the weight value of customer service node i is introduced, where ω i The data volume of the local text of the customer service node i and cluster A j The ratio of the local text data to the total data volume is determined.

[0082] In some embodiments of the present application, cluster model aggregation is performed according to the intra-cluster noise recognition models corresponding to all clusters to obtain an iterative global noise recognition model, including:

[0083]

[0084]

[0085] Among them, F is the iterative global noise recognition model, Cluster A j The weight value, S total The amount of local text data for all customer service nodes.

[0086] In this embodiment, cluster A is introduced. j The weight value of From cluster A j The data volume is determined by the ratio of the total data volume of the local text of the client node to the data volume of the local text of all customer service nodes.

[0087] In some embodiments of the present application, after performing inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain the iterative target global noise recognition model, the method further includes:

[0088] When the iterative target global noise recognition model meets the preset requirements, the iterative target global noise recognition model is sent to each customer service node.

[0089] Each customer service node can use the iterated target global noise recognition model to identify the noise in the customer's text.

[0090] In some embodiments of the present application, the target global noise recognition model is a model constructed by a graph neural network model.

[0091] An embodiment of the present application provides a noise recognition system based on federated learning, the system comprising:

[0092] The model training module is used to send the target global noise recognition model to M customer service nodes so that each customer service node performs a model training process; M is a positive integer greater than 1, and the model training process includes:

[0093] Perform noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and model parameters after training;

[0094] The model aggregation module is used to perform clustering, inter-cluster aggregation and intra-cluster aggregation processes, where:

[0095] The clustering process includes:

[0096] receiving text features of M customer service nodes, and dividing the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of customer service nodes; N is a positive integer greater than 1, and M is greater than N;

[0097] The inter-cluster aggregation process includes:

[0098] Receive the model parameters of all customer service nodes corresponding to each cluster, aggregate the intra-cluster models according to the model parameters, and obtain the intra-cluster noise recognition model corresponding to each cluster;

[0099] The intra-cluster aggregation process includes:

[0100] According to the intra-cluster noise recognition models corresponding to N clusters, cluster models are aggregated to obtain the iterative target global noise recognition model;

[0101] The model iteration judgment module is used to use the iterated target global noise recognition model as the target global noise recognition model and send it to all customer service nodes when the target global noise recognition model after iteration does not meet the preset requirements, and execute the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process.

[0102] The device first distributes the target global noise recognition model to obtain the text features corresponding to the local text and the model parameters after training; then the customer service nodes are clustered by the text features, and a class of customer service nodes with relatively similar text features can be classified into the same cluster, so that the text feature distribution differences of the customer service nodes in the same cluster are relatively small; then intra-cluster and inter-cluster aggregation are performed, and finally the target global noise recognition model can be continuously iterated, thereby continuously improving the model's recognition reliability of noise.

[0103] Specifically, the processor 401 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiment of the present invention.

[0104] The memory 402 may include a large capacity memory for data or instructions. For example, but not limitation, the memory 402 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 402 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 402 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 402 is a non-volatile solid-state memory. The memory may include a read-only memory (ROM), a random access memory (RAM), a disk storage medium device, an optical storage medium device, a flash memory device, an electrical, optical or other physical / tangible memory storage device. Therefore, typically, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described in the noise identification method based on federated learning provided in the above-mentioned embodiments.

[0105] The processor 401 implements any noise recognition method based on federated learning in the above embodiments by reading and executing computer program instructions stored in the memory 402 .

[0106] In one example, the electronic device may further include a communication interface 403 and a bus 410. The processor 401, the memory 402, and the communication interface 403 are connected via the bus 410 and communicate with each other.

[0107] The communication interface 403 is mainly used to implement the communication between the modules, devices, units and / or devices in the embodiment of the present invention.

[0108] Bus 410 includes hardware, software or both, and the parts of electronic equipment are coupled to each other. For example, but not limitation, bus may include accelerated graphics port (AGP) or other graphics bus, enhanced industrial standard architecture (EISA) bus, front side bus (FSB), hypertransport (HT) interconnection, industrial standard architecture (ISA) bus, infinite bandwidth interconnection, low pin count (LPC) bus, memory bus, micro channel architecture (MCA) bus, peripheral component interconnection (PCI) bus, PCI-Express (PCI-X) bus, serial advanced technology attachment (SATA) bus, video electronics standard association local (VLB) bus or other suitable bus or two or more of these combinations. In appropriate cases, bus 410 may include one or more buses. Although the embodiment of the present invention describes and shows a specific bus, the present invention considers any suitable bus or interconnection.

[0109] The electronic device can execute the data processing method in the embodiment of the present invention, thereby realizing Figure 1 Any of the described federated learning based noise identification methods.

[0110] This device first distributes the target global noise recognition model to obtain the text features corresponding to the local text and the model parameters after training; then the customer service nodes are clustered by the text features, and a class of customer service nodes with relatively similar text features can be classified into the same cluster, so that the text feature distribution differences of the customer service nodes in the same cluster are relatively small; then intra-cluster and inter-cluster aggregation are performed, and finally the target global noise recognition model can be continuously iterated, thereby continuously improving the model's recognition reliability of noise.

[0111] In addition, in combination with the noise identification method based on federated learning in the above embodiment, the embodiment of the present invention can provide a readable storage medium for implementation. The readable storage medium stores program instructions; when the program instructions are executed by the processor, any noise identification method based on federated learning in the above embodiment is implemented.

[0112] It should be clear that the present invention is not limited to the specific configuration and processing described above and shown in the figures. For the sake of brevity, a detailed description of the known methods is omitted here.

[0113] In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present invention.

[0114] The functional blocks shown in the structural block diagram described above can be implemented as hardware, software, firmware or a combination thereof.

[0115] When implemented in hardware, it may be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments used to perform the required tasks.

[0116] The program or code segments may be stored in a machine-readable medium or transmitted over a transmission medium or a communication link via a data signal carried in a carrier wave.

[0117] "Machine-readable media" may include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio frequency (RF) links, and the like. Code segments may be downloaded via computer networks such as the Internet, intranets, and the like.

[0118] It should also be noted that the exemplary embodiments mentioned in the present invention describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiments, or in a different order from the embodiments, or several steps can be performed simultaneously.

[0119] Aspects of the present application are described above with reference to flowcharts and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present application.

[0120] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine, so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0121] Such a processor may be, but is not limited to, a general purpose processor, a dedicated processor, a special application processor, or a field programmable logic circuit. It is also understood that each box in the block diagram and / or flow chart and the combination of boxes in the block diagram and / or flow chart may also be implemented by dedicated hardware that performs a specified function or action, or may be implemented by a combination of dedicated hardware and computer instructions.

Claims

1. A noise recognition method based on federated learning, characterized in that: The method comprises: Sending a target global noise recognition model to M customer service nodes so that each of the customer service nodes performs a model training process; M is a positive integer greater than 1, and the model training process includes: Performing noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and trained model parameters; Perform clustering, inter-cluster aggregation and intra-cluster aggregation processes, where: The clustering process includes: receiving text features of the M customer service nodes, and dividing the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of the customer service nodes; N is a positive integer greater than 1, and M is greater than N; The inter-cluster aggregation process includes: Receiving the model parameters of all the customer service nodes corresponding to each cluster, performing intra-cluster model aggregation according to the model parameters, and obtaining an intra-cluster noise recognition model corresponding to each cluster; The intra-cluster aggregation process includes: Perform inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model; In the case that the iterated target global noise recognition model does not meet the preset requirements, the iterated target global noise recognition model is used as the target global noise recognition model and sent to all customer service nodes, and the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process are performed.

2. The noise identification method based on federated learning according to claim 1, characterized in that: The dividing the M customer service nodes into N clusters according to the text features includes: Based on the K nearest neighbor classification algorithm, the M customer service nodes are divided into N clusters according to the text features.

3. The noise identification method based on federated learning according to claim 1, characterized in that: The step of aggregating the intra-cluster models according to the model parameters to obtain an intra-cluster noise recognition model corresponding to each cluster includes: in, Cluster A j The intra-cluster noise recognition model, F i is the target global noise recognition model for customer service node i, θ i is the model parameter of customer service node i, ω i is the weight value of customer service node i, S i is the data volume of the local text of customer service node i, Cluster A j The total amount of local text data.

4. The noise identification method based on federated learning according to claim 3 is characterized in that: The inter-cluster model aggregation is performed according to the intra-cluster noise recognition models corresponding to all clusters to obtain an iterative global noise recognition model, including: Among them, F is the iterative global noise recognition model, Cluster C j The weight value, S total The amount of local text data for all customer service nodes.

5. The noise identification method based on federated learning according to claim 1, characterized in that: After performing inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model, the method further includes: When the iterated target global noise recognition model meets the preset requirement, the iterated target global noise recognition model is sent to each of the customer service nodes.

6. The noise identification method based on federated learning according to claim 1, characterized in that: The target global noise recognition model is a model constructed by a graph neural network model.

7. A noise recognition system based on federated learning, characterized in that: The system comprises: A model training module is used to send a target global noise recognition model to M customer service nodes so that each of the customer service nodes performs a model training process; M is a positive integer greater than 1, and the model training process includes: Performing noise recognition on the local text based on the target global noise recognition model to obtain text features corresponding to the local text and trained model parameters; The model aggregation module is used to perform clustering, inter-cluster aggregation and intra-cluster aggregation processes, where: The clustering process includes: receiving text features of the M customer service nodes, and dividing the M customer service nodes into N clusters according to the text features, wherein each cluster includes a number of the customer service nodes; N is a positive integer greater than 1, and M is greater than N; The inter-cluster aggregation process includes: Receiving the model parameters of all the customer service nodes corresponding to each cluster, performing intra-cluster model aggregation according to the model parameters, and obtaining an intra-cluster noise recognition model corresponding to each cluster; The intra-cluster aggregation process includes: Perform inter-cluster model aggregation according to the intra-cluster noise recognition models corresponding to the N clusters to obtain an iterative target global noise recognition model; The model iteration judgment module is used to use the iterated target global noise recognition model as the target global noise recognition model and send it to all customer service nodes when the iterated target global noise recognition model does not meet the preset requirements, and execute the model training process, clustering, inter-cluster aggregation and intra-cluster aggregation process.

8. An electronic device, characterized in that: It includes at least one controller and a memory for communicating with the controller; the memory stores instructions that can be executed by the at least one controller, and the instructions are executed by at least one of the controllers to enable at least one of the controllers to perform the noise recognition method based on federated learning as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the noise identification method based on federated learning as described in any one of claims 1 to 6.

10. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the noise recognition method based on federated learning as described in any one of claims 1 to 6.

Images