WO2021114618A1 - Procédé et appareil d'apprentissage fédéré, dispositif informatique et support de stockage lisible - Google Patents

Procédé et appareil d'apprentissage fédéré, dispositif informatique et support de stockage lisible Download PDF

Info

Publication number
WO2021114618A1
WO2021114618A1 PCT/CN2020/098890 CN2020098890W WO2021114618A1 WO 2021114618 A1 WO2021114618 A1 WO 2021114618A1 CN 2020098890 W CN2020098890 W CN 2020098890W WO 2021114618 A1 WO2021114618 A1 WO 2021114618A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
sample data
vector
federated learning
learning model
Prior art date
Application number
PCT/CN2020/098890
Other languages
English (en)
Chinese (zh)
Inventor
周学立
朱恩东
张茜
刘丽扬
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2021114618A1 publication Critical patent/WO2021114618A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60—Protecting data
    • G06F21/602—Providing cryptographic facilities or services
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance

Definitions

  • This application relates to the field of artificial intelligence technology, in particular to a federated learning method, device, computer equipment, and computer-readable storage medium.
  • the first aspect of this application provides a federated learning method, and the federated learning method includes:
  • the feature vector and the identification code are transmitted to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and performs federated learning model training according to the feature vector and the label.
  • a second aspect of the present application provides a federated learning device, the federated learning device includes:
  • An obtaining module used to obtain sample data and an identification code of the sample data
  • a conversion module configured to convert the sample data into a vector to obtain a vector representation of the sample data
  • An encoding module configured to encode the vector representation to obtain the feature vector of the sample data
  • the transmission module is configured to transmit the feature vector and the identification code to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and performs processing according to the feature vector and the label Federation learning model training.
  • a third aspect of the present application provides a computer device that includes a processor, and the processor is configured to execute computer-readable instructions stored in a memory to implement the following steps:
  • the feature vector and the identification code are transmitted to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and performs federated learning model training according to the feature vector and the label.
  • a fourth aspect of the present application provides a computer-readable storage medium having computer-readable instructions stored on the computer-readable storage medium, and when the computer-readable instructions are executed by a processor, the following steps are implemented:
  • the feature vector and the identification code are transmitted to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and performs federated learning model training according to the feature vector and the label.
  • This application uses the coding model to perform feature learning and feature integration on the vector representation. Without a decoder, the data requesting end cannot interpret and obtain the sample data corresponding to the vector representation, which ensures the security of the data. The data requesting end does not directly obtain the data of the data providing end, which improves the security of the data in the federated learning process.
  • Fig. 1 is a flowchart of a federated learning method provided by an embodiment of the application.
  • Figure 2 is a structural diagram of a federated learning device provided by an embodiment of the present application.
  • Fig. 3 is a schematic diagram of a computer device provided by an embodiment of the present application.
  • the federated learning method of this application is applied to one or more computer devices.
  • the computer device is a device that can automatically perform numerical calculation and/or information processing in accordance with pre-set or stored instructions.
  • Its hardware includes, but is not limited to, a microprocessor and an application specific integrated circuit (ASIC) , Programmable Gate Array (Field-Programmable Gate Array, FPGA), Digital Processor (Digital Signal Processor, DSP), embedded equipment, etc.
  • ASIC application specific integrated circuit
  • FPGA Field-Programmable Gate Array
  • DSP Digital Processor
  • embedded equipment etc.
  • This application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, small computers, large computers, including Distributed computing environment for any of the above systems or equipment, etc.
  • This application may be described in the general context of computer-executable instructions executed by a computer, such as a program module.
  • program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types.
  • This application can also be practiced in distributed computing environments. In these distributed computing environments, tasks are performed by remote processing devices connected through a communication network.
  • program modules can be located in local and remote computer storage media including storage devices.
  • the computer device may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.
  • the computer device can interact with the user through a keyboard, a mouse, a remote control, a touch panel, or a voice control device.
  • Fig. 1 is a flow chart of the federated learning method provided in Embodiment 1 of the present application.
  • the federated learning method is applied to a data provider, and the data provider is a computer device for generating a federated learning model through federated learning.
  • the federated learning method specifically includes the following steps. According to different needs, the order of the steps in the flowchart can be changed, and some can be omitted.
  • the identification code is the unique identification information of the sample data, and is used to identify the sample data between the data provider and the third party requesting the data.
  • the sample data may include different data types such as text data, one-hot data, numerical data, and embedding data.
  • the data provider may be a financial company
  • the third party requesting the data may be an insurance company
  • the sample data may be user behavior data of the financial company
  • the data type of the sample data may be text data.
  • the identification code may be the user's mobile phone number or ID number corresponding to the user behavior data.
  • the insurance company needs to obtain the coded user behavior data of the financial company, and use the insurance reliability scoring model to score the reliability of the user based on the obtained coded user behavior data. That is, insurance companies do not need to directly obtain user behavior data of financial companies, which protects the data security of financial companies.
  • the insurance reliability scoring model can be a specific federated learning model, and the local federated learning model of an insurance company can be a deep learning model.
  • the data provider may be an e-commerce company
  • the third party requesting the data may be an advertising company
  • the sample data may be product click behavior data of the e-commerce company
  • the data type of the sample data may be numeric data .
  • the identification code may be the user's mobile phone number or ID number corresponding to the commodity click behavior data.
  • the advertising company needs to obtain the coded product click behavior data of the e-commerce company, and recommend products to users according to the obtained coded product click behavior data through the product recommendation model. That is, the advertising company does not need to directly obtain the product click behavior data of the e-commerce company, which protects the data security of the e-commerce company.
  • the product recommendation model can be a specific federated learning model, and the local federated learning model of the advertising company can be a deep learning model.
  • the converting the sample data into a vector includes:
  • the sample data is converted into a vector according to a preset conversion method corresponding to the data type of the sample data.
  • the sample data is converted into a vector according to the word2vec method.
  • the data type of the sample data is numeric data (the preset conversion method corresponding to the numeric data is a standardization method)
  • the sample data is converted into a vector according to the standardization method.
  • the judging whether the sample data needs to be converted into a vector according to the data type of the sample data includes:
  • the sample data needs to be converted into a vector
  • the sample data does not need to be converted into a vector.
  • the data to be converted is a vector, and no conversion is required.
  • the encoding the vector representation includes:
  • the encoding model consisting of an encoder and a decoder
  • the vector representation is encoded with the trained encoder.
  • the data provider In order to ensure the security of the sample data, the data provider cannot directly request the sample data from a third party.
  • the trained encoding model is optimized through the Deep auto-encoder (deep encoding) or sparse auto-encoder (sparse encoding) method.
  • the trained coding model may be optimized by a sparse coding method according to the difference between the output of the trained coding model and the input.
  • the sparse coding method mainly optimizes the trained coding model by adding sparsity restriction conditions to the neural units in the trained coding model.
  • the sparsity restriction condition may include, when the output value of the neuron is close to 1 (for example, greater than 0.9), the neuron is activated; when the output value of the neuron is close to 0 (for example, less than or equal to 0.9), the neuron is not activated.
  • the coding model is used to perform feature learning and feature integration on the vector representation. Without a decoder, the data requesting end cannot interpret and obtain the sample data corresponding to the vector representation, which ensures the security of the data.
  • the federated learning model includes: LR, XGB, DNN, etc. LR, XGB, DNN and other models are used for machine learning training and algorithm models for business use purposes.
  • the federated learning model may be a specific artificial intelligence model, such as an artificial intelligence classification model, an artificial intelligence recognition model, and the like.
  • the transmitting the feature vector and the identification code to the data requesting terminal includes:
  • the feature vector and the identification code are transmitted to the data requesting end through an encryption algorithm.
  • the feature vector and the identification code are encrypted by the private key of the data provider; the encrypted feature vector and the identification code are transmitted to the data requesting terminal, so that the data requesting terminal passes through the data provider's
  • the public key decrypts the encrypted feature vector and identification code.
  • the labels of the sample data are "risk user” and "normal user”.
  • the labels of the sample data are "recommended product one”, “recommended product two”, and so on.
  • the training of the federated learning model according to the feature vector and the label includes:
  • the data requesting terminal obtains the initial parameters of the federated learning model from a preset server;
  • the data requesting terminal initializes the federated learning model with the initial parameters
  • the data requesting terminal locally trains the initialized federated learning model according to the feature vector and the label, updates the parameters of the initialized federated learning model, and obtains the updated parameters;
  • the data requesting end uploads the updated parameters to the preset server, so that the preset server performs aggregation processing on the parameters uploaded by each requesting end to obtain aggregation parameters, and when it is detected that the aggregation is used
  • the preset server performs aggregation processing on the parameters uploaded by each requesting end to obtain aggregation parameters, and when it is detected that the aggregation is used
  • deliver the updated federated learning model to the data requester
  • the data requesting terminal receives the federated learning model issued by the preset server.
  • the preset server before sending the updated federated learning model to the data requester, when the preset server detects that the federated learning model updated with the aggregation parameters is in a non-convergent state When the time, the preset server returns the aggregation parameter to the data requesting terminal, so that the data requesting terminal continues iterative training.
  • the federated learning method of the first embodiment generates a federated learning model through federated learning.
  • the coding model is used to perform feature learning and feature integration on the vector representation, and the data requester cannot interpret and obtain sample data corresponding to the vector representation without a decoder, which ensures the security of the data To prevent data leakage.
  • the coding model does not need to add noise to the vector representation, and avoids the generation of additional interference information due to the addition of noise.
  • the establishment of the federation model has direct feedback on the coding results, which is beneficial to optimization and adjustment.
  • the coding model can adjust the information loss degree and the information security degree of the feature learning and feature integration of the vector representation, find the compromise point of information security and information loss, and obtain more optimized parameters of the entire federated learning model.
  • the data requesting end does not directly obtain the data of the data providing end, which improves the security of the data in the federated learning process.
  • the federated learning method further includes:
  • Hyperparameters include network structure, number of neural layers, number of neurons in each layer, activation function, learning rate, regularization and penalty coefficients, loss function, and so on. Specifically, when the loss function floats and does not converge, the loss function, learning rate, and/or network structure can be adjusted. When the gradient disappears or the gradient explodes, adjust the activation function.
  • the federated learning method further includes:
  • the updated local federated learning model is used to process the to-be-processed data.
  • the federated learning method further includes:
  • the parameters and/or hyperparameters of the coding model and/or the federated learning model are adjusted according to the processing result of the to-be-processed data and the preset result of the to-be-processed data.
  • the coding model can be determined whether the coding model over-encodes data according to the processing result of the data to be processed and the preset result of the data to be processed.
  • Over-encoding data may cause the encoding model to lose its ability to extract effective features;
  • the judgment result adjusts the coding model to improve the feature extraction capability of the coding model and balance the feature extraction capability with the data security achieved through coding.
  • Fig. 2 is a structural diagram of a federated learning device provided in the second embodiment of the present application.
  • the federated learning device 20 is applied to a data provider, and the data provider is a computer device.
  • the federated learning device 20 is used to generate federated learning models through federated learning.
  • the federated learning device 20 may include an acquisition module 201, a conversion module 202, an encoding module 203, and a transmission module 204.
  • the obtaining module 201 is used to obtain sample data and an identification code of the sample data.
  • the identification code is the unique identification information of the sample data, and is used to identify the sample data between the data provider and the third party requesting the data.
  • the sample data may include different data types such as text data, one-hot data, numerical data, and embedding data.
  • the data provider may be a financial company
  • the third party requesting the data may be an insurance company
  • the sample data may be user behavior data of the financial company
  • the data type of the sample data may be text data.
  • the identification code may be the user's mobile phone number or ID number corresponding to the user behavior data.
  • the insurance company needs to obtain the coded user behavior data of the financial company, and use the insurance reliability scoring model to score the reliability of the user based on the obtained coded user behavior data. That is, insurance companies do not need to directly obtain user behavior data of financial companies, which protects the data security of financial companies.
  • the insurance reliability scoring model can be a specific federated learning model, and the local federated learning model of an insurance company can be a deep learning model.
  • the data provider may be an e-commerce company
  • the third party requesting the data may be an advertising company
  • the sample data may be product click behavior data of the e-commerce company
  • the data type of the sample data may be numeric data .
  • the identification code may be the user's mobile phone number or ID number corresponding to the commodity click behavior data.
  • the advertising company needs to obtain the coded product click behavior data of the e-commerce company, and recommend products to users according to the obtained coded product click behavior data through the product recommendation model. That is, the advertising company does not need to directly obtain the product click behavior data of the e-commerce company, which protects the data security of the e-commerce company.
  • the product recommendation model can be a specific federated learning model, and the local federated learning model of the advertising company can be a deep learning model.
  • the conversion module 202 is configured to convert the sample data into a vector to obtain a vector representation of the sample data.
  • the converting the sample data into a vector includes:
  • the sample data is converted into a vector according to a preset conversion method corresponding to the data type of the sample data.
  • the sample data is converted into a vector according to the word2vec method.
  • the data type of the sample data is numeric data (the preset conversion method corresponding to the numeric data is a standardization method)
  • the sample data is converted into a vector according to the standardization method.
  • the judging whether the sample data needs to be converted into a vector according to the data type of the sample data includes:
  • the sample data needs to be converted into a vector
  • the sample data does not need to be converted into a vector.
  • the data to be converted is a vector, and no conversion is required.
  • the encoding module 203 is configured to encode the vector representation to obtain the feature vector of the sample data.
  • the encoding the vector representation includes:
  • the encoding model consisting of an encoder and a decoder
  • the vector representation is encoded with the trained encoder.
  • the data provider In order to ensure the security of the sample data, the data provider cannot directly request the sample data from a third party.
  • the federated learning device 20 further includes an optimization module for optimizing the trained encoding model through a Deep auto-encoder (deep encoding) or sparse auto-encoder (sparse encoding) method.
  • the trained coding model may be optimized by a sparse coding method according to the difference between the output of the trained coding model and the input.
  • the sparse coding method mainly optimizes the trained coding model by adding sparsity restriction conditions to the neural units in the trained coding model.
  • the sparsity restriction condition may include, when the output value of the neuron is close to 1 (for example, greater than 0.9), the neuron is activated; when the output value of the neuron is close to 0 (for example, less than or equal to 0.9), the neuron is not activated.
  • the coding model is used to perform feature learning and feature integration on the vector representation. Without a decoder, the data requesting end cannot interpret and obtain the sample data corresponding to the vector representation, which ensures the security of the data.
  • the transmission module 204 is configured to transmit the feature vector and the identification code to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and according to the feature vector and the label Perform federated learning model training.
  • the federated learning model includes: LR, XGB, DNN, etc. LR, XGB, DNN and other models are used for machine learning training and algorithm models for business use purposes.
  • the federated learning model may be a specific artificial intelligence model, such as an artificial intelligence classification model, an artificial intelligence recognition model, and the like.
  • the transmitting the feature vector and the identification code to the data requesting terminal includes:
  • the feature vector and the identification code are transmitted to the data requesting end through an encryption algorithm.
  • the feature vector and the identification code are encrypted by the private key of the data provider; the encrypted feature vector and the identification code are transmitted to the data requesting terminal, so that the data requesting terminal passes through the data provider's
  • the public key decrypts the encrypted feature vector and identification code.
  • the labels of the sample data are "risk user” and "normal user”.
  • the labels of the sample data are "recommended product one”, “recommended product two”, and so on.
  • the training of the federated learning model according to the feature vector and the label includes:
  • the data requesting terminal obtains the initial parameters of the federated learning model from a preset server;
  • the data requesting terminal initializes the federated learning model with the initial parameters
  • the data requesting terminal locally trains the initialized federated learning model according to the feature vector and the label, updates the parameters of the initialized federated learning model, and obtains the updated parameters;
  • the data requesting end uploads the updated parameters to the preset server, so that the preset server performs aggregation processing on the parameters uploaded by each requesting end to obtain aggregation parameters, and when it is detected that the aggregation is used
  • the preset server performs aggregation processing on the parameters uploaded by each requesting end to obtain aggregation parameters, and when it is detected that the aggregation is used
  • deliver the updated federated learning model to the data requester
  • the data requesting terminal receives the federated learning model issued by the preset server.
  • the federated learning device 20 of the second embodiment generates a federated learning model through federated learning.
  • the coding model is used to perform feature learning and feature integration on the vector representation, and the data requester cannot interpret and obtain sample data corresponding to the vector representation without a decoder, which ensures the security of the data To prevent data leakage.
  • the coding model does not need to add noise to the vector representation, and avoids the generation of additional interference information due to the addition of noise.
  • the establishment of the federation model has direct feedback on the coding results, which is beneficial to optimization and adjustment.
  • the coding model can adjust the information loss degree and the information security degree of the feature learning and feature integration of the vector representation, find the compromise point of information security and information loss, and obtain more optimized parameters of the entire federated learning model.
  • the data requesting end does not directly obtain the data of the data providing end, which improves the security of the data in the federated learning process.
  • the federated learning device 20 further includes an adjustment module for adjusting the hyperparameters of the encoder and/or the federated learning model.
  • Hyperparameters include network structure, number of neural layers, number of neurons in each layer, activation function, learning rate, regularization and penalty coefficients, loss function, and so on. Specifically, when the loss function floats and does not converge, the loss function, learning rate, and/or network structure can be adjusted. When the gradient disappears or the gradient explodes, adjust the activation function.
  • the federated learning device 20 further includes a processing module for obtaining the parameters of the trained federated learning model from the data requesting terminal; acquiring the data to be processed; using the trained federated learning model Update the local federated learning model with the parameters of, and process the data to be processed with the updated local federated learning model.
  • the adjustment module is further configured to obtain the processing result of the to-be-processed data; obtain the preset result of the to-be-processed data; according to the processing result of the to-be-processed data and the to-be-processed data The preset result of adjusting the parameters and/or hyperparameters of the coding model and/or the federated learning model.
  • the coding model can be determined whether the coding model over-encodes data according to the processing result of the data to be processed and the preset result of the data to be processed.
  • Over-encoding data may cause the encoding model to lose its ability to extract effective features;
  • the judgment result adjusts the coding model to improve the feature extraction capability of the coding model and balance the feature extraction capability with the data security achieved through coding.
  • This embodiment provides a computer-readable storage medium having computer-readable instructions stored thereon.
  • the computer-readable storage medium may be non-volatile or volatile.
  • the steps in the above-mentioned federated learning method embodiment are implemented, for example, steps 101-104 shown in Fig. 1:
  • each module in the above-mentioned device embodiment is realized, for example, the modules 201-204 in Fig. 2:
  • the obtaining module 201 is used to obtain sample data and an identification code of the sample data
  • the conversion module 202 is configured to convert the sample data into a vector to obtain a vector representation of the sample data
  • the encoding module 203 is configured to encode the vector representation to obtain the feature vector of the sample data
  • the transmission module 204 is configured to transmit the feature vector and the identification code to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and according to the feature vector and the label Perform federated learning model training.
  • FIG. 3 is a schematic diagram of the computer equipment provided in the fourth embodiment of the application.
  • the computer device 30 includes a memory 301, a processor 302, and computer-readable instructions 303 stored in the memory 301 and running on the processor 302, such as a federated learning program.
  • the processor 302 implements the steps in the embodiment of the federated learning method when the computer readable instruction 303 is executed, for example, steps 101-104 shown in FIG. 1:
  • each module in the above-mentioned device embodiment is realized, for example, the modules 201-204 in FIG. 2:
  • the obtaining module 201 is used to obtain sample data and an identification code of the sample data
  • the conversion module 202 is configured to convert the sample data into a vector to obtain a vector representation of the sample data
  • the encoding module 203 is configured to encode the vector representation to obtain the feature vector of the sample data
  • the transmission module 204 is configured to transmit the feature vector and the identification code to the data requesting terminal, so that the data requesting terminal searches for the label of the sample data according to the identification code, and according to the feature vector and the label Perform federated learning model training.
  • the computer-readable instruction 303 may be divided into one or more modules, and the one or more modules are stored in the memory 301 and executed by the processor 302 to complete the method.
  • the one or more modules may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer readable instruction 303 in the computer device 30.
  • the computer-readable instruction 303 can be divided into the acquisition module 201, the conversion module 202, the encoding module 203, and the transmission module 204 in FIG. 2.
  • the specific functions of each module refer to the second embodiment.
  • the computer device 30 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.
  • a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.
  • the schematic diagram 3 is only an example of the computer device 30, and does not constitute a limitation on the computer device 30. It may include more or less components than those shown in the figure, or combine certain components, or different components.
  • the computer device 30 may also include input and output devices, network access devices, buses, and so on.
  • the so-called processor 302 may be a central processing unit (Central Processing Unit, CPU), other general processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor can be a microprocessor or the processor 302 can also be any conventional processor, etc.
  • the processor 302 is the control center of the computer device 30, which uses various interfaces and lines to connect the entire computer device 30. Various parts.
  • the memory 301 may be used to store the computer-readable instructions 303, and the processor 302 executes or executes the computer-readable instructions or modules stored in the memory 301 and calls the data stored in the memory 301 to implement Various functions of the computer device 30.
  • the memory 301 may mainly include a storage program area and a storage data area, where the storage program area may store an operating system, an application program required by at least one function (such as a sound playback function, an image playback function, etc.); the storage data area may Data and the like created in accordance with the use of the computer device 30 are stored.
  • the memory 301 may include a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a Secure Digital (SD) card, a flash memory card (Flash Card), at least one disk storage device, flash memory Devices, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), or other non-volatile/volatile storage devices.
  • the integrated module of the computer device 30 may be stored in a computer-readable storage medium.
  • the computer-readable storage medium may be non-volatile or volatile. Based on this understanding, this application implements all or part of the processes in the above-mentioned embodiments and methods, and can also be completed by instructing relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium.
  • the computer-readable instruction when executed by the processor, it can implement the steps of the foregoing method embodiments.
  • the computer-readable instructions may be in the form of source code, object code, executable file, or some intermediate forms, etc.
  • the computer-readable storage medium may include: any entity or device capable of carrying the computer-readable instructions, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, read only memory (ROM), random access memory ( RAM).
  • modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional modules in the various embodiments of the present application may be integrated into one processing module, or each module may exist alone physically, or two or more modules may be integrated into one module.
  • the above-mentioned integrated modules can be implemented in the form of hardware, or in the form of hardware plus software functional modules.
  • the above-mentioned integrated modules implemented in the form of software functional modules may be stored in a computer-readable storage medium.
  • the above-mentioned software function module is stored in a storage medium and includes several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor execute the federated learning described in each embodiment of this application. Part of the method.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Finance (AREA)
  • Biophysics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Accounting & Taxation (AREA)
  • Medical Informatics (AREA)
  • Computer Security & Cryptography (AREA)
  • Computer Hardware Design (AREA)
  • Bioethics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Technology Law (AREA)
  • General Business, Economics & Management (AREA)
  • Image Analysis (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

L'invention concerne un procédé et un appareil d'apprentissage fédéré, un dispositif informatique et un support de stockage lisible. Le procédé d'apprentissage fédéré comprend les étapes consistant à : acquérir des données d'échantillon et un code d'identification des données d'échantillon ; convertir les données d'échantillon en vecteur pour obtenir une représentation vectorielle des données d'échantillon ; coder la représentation vectorielle pour obtenir un vecteur propre des données d'échantillon ; et transmettre le vecteur propre et le code d'identification à une extrémité de demande de données, de telle sorte que l'extrémité de demande de données recherche une étiquette des données d'échantillon sur la base du code d'identification et met en oeuvre un apprentissage de modèle d'apprentissage fédéré sur la base du vecteur propre et de l'étiquette. Le présent procédé augmente la sécurité des données pendant le processus d'apprentissage fédéré.
PCT/CN2020/098890 2020-05-14 2020-06-29 Procédé et appareil d'apprentissage fédéré, dispositif informatique et support de stockage lisible WO2021114618A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010408557.4 2020-05-14
CN202010408557.4A CN111695674B (zh) 2020-05-14 2020-05-14 联邦学习方法、装置、计算机设备及可读存储介质

Publications (1)

Publication Number Publication Date
WO2021114618A1 true WO2021114618A1 (fr) 2021-06-17

Family

ID=72477764

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/098890 WO2021114618A1 (fr) 2020-05-14 2020-06-29 Procédé et appareil d'apprentissage fédéré, dispositif informatique et support de stockage lisible

Country Status (2)

Country Link
CN (1) CN111695674B (fr)
WO (1) WO2021114618A1 (fr)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781397A (zh) * 2021-08-11 2021-12-10 中国科学院信息工程研究所 基于联邦学习的医疗影像病灶检测建模方法、装置及系统
CN114154392A (zh) * 2021-10-15 2022-03-08 海南火链科技有限公司 基于区块链和联邦学习的模型共建方法、装置及设备
CN114648130A (zh) * 2022-02-07 2022-06-21 北京航空航天大学 纵向联邦学习方法、装置、电子设备及存储介质
CN114996317A (zh) * 2022-07-05 2022-09-02 中国电信股份有限公司 基于纵向联邦学习的异步优化方法、装置及存储介质
CN117424765A (zh) * 2023-12-19 2024-01-19 天津医康互联科技有限公司 分布式独热编码方法、装置、电子设备及计算机存储介质

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112364919A (zh) * 2020-11-11 2021-02-12 深圳前海微众银行股份有限公司 数据处理方法、装置、设备及存储介质
CN112396189B (zh) * 2020-11-27 2023-09-01 中国银联股份有限公司 一种多方构建联邦学习模型的方法及装置
CN112463161B (zh) * 2020-11-27 2023-07-28 平安科技（深圳）有限公司 基于联邦学习的代码注释生成方法、系统及装置
CN113239351B (zh) * 2020-12-08 2022-05-13 武汉大学 一种面向物联网系统的新型数据污染攻击防御方法
CN113283185B (zh) * 2021-07-23 2021-11-12 平安科技（深圳）有限公司 联邦模型训练、客户画像方法、装置、设备及介质
CN113722987B (zh) * 2021-08-16 2023-11-03 京东科技控股股份有限公司 联邦学习模型的训练方法、装置、电子设备和存储介质
CN114357521A (zh) * 2022-01-12 2022-04-15 平安科技（深圳）有限公司 数据保护方法和系统
CN114710415B (zh) * 2022-05-23 2022-08-12 北京理工大学 一种冗余编码的无源消息可靠传输与处理系统
CN115034816A (zh) * 2022-06-07 2022-09-09 青岛文达通科技股份有限公司 一种基于无监督和联邦学习的需求预测方法及系统
CN115495541B (zh) * 2022-11-18 2023-04-07 深译信息科技(珠海)有限公司 语料数据库、语料数据库的维护方法、装置、设备和介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110297848A (zh) * 2019-07-09 2019-10-01 深圳前海微众银行股份有限公司 基于联邦学习的推荐模型训练方法、终端及存储介质
CN110738323A (zh) * 2018-07-03 2020-01-31 百度在线网络技术（北京）有限公司 基于数据共享建立机器学习模型的方法和装置
CN110807207A (zh) * 2019-10-30 2020-02-18 腾讯科技（深圳）有限公司 数据处理方法、装置、电子设备及存储介质
CN111027715A (zh) * 2019-12-11 2020-04-17 支付宝(杭州)信息技术有限公司 基于蒙特卡洛的联邦学习模型训练方法及装置

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014008403A1 (fr) * 2012-07-03 2014-01-09 Visa International Service Association Concentrateur de protection de données

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110738323A (zh) * 2018-07-03 2020-01-31 百度在线网络技术（北京）有限公司 基于数据共享建立机器学习模型的方法和装置
CN110297848A (zh) * 2019-07-09 2019-10-01 深圳前海微众银行股份有限公司 基于联邦学习的推荐模型训练方法、终端及存储介质
CN110807207A (zh) * 2019-10-30 2020-02-18 腾讯科技（深圳）有限公司 数据处理方法、装置、电子设备及存储介质
CN111027715A (zh) * 2019-12-11 2020-04-17 支付宝(杭州)信息技术有限公司 基于蒙特卡洛的联邦学习模型训练方法及装置

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113781397A (zh) * 2021-08-11 2021-12-10 中国科学院信息工程研究所 基于联邦学习的医疗影像病灶检测建模方法、装置及系统
CN113781397B (zh) * 2021-08-11 2023-11-21 中国科学院信息工程研究所 基于联邦学习的医疗影像病灶检测建模方法、装置及系统
CN114154392A (zh) * 2021-10-15 2022-03-08 海南火链科技有限公司 基于区块链和联邦学习的模型共建方法、装置及设备
CN114648130A (zh) * 2022-02-07 2022-06-21 北京航空航天大学 纵向联邦学习方法、装置、电子设备及存储介质
CN114648130B (zh) * 2022-02-07 2024-04-16 北京航空航天大学 纵向联邦学习方法、装置、电子设备及存储介质
CN114996317A (zh) * 2022-07-05 2022-09-02 中国电信股份有限公司 基于纵向联邦学习的异步优化方法、装置及存储介质
CN114996317B (zh) * 2022-07-05 2024-02-23 中国电信股份有限公司 基于纵向联邦学习的异步优化方法、装置及存储介质
CN117424765A (zh) * 2023-12-19 2024-01-19 天津医康互联科技有限公司 分布式独热编码方法、装置、电子设备及计算机存储介质
CN117424765B (zh) * 2023-12-19 2024-03-22 天津医康互联科技有限公司 分布式独热编码方法、装置、电子设备及计算机存储介质

Also Published As

Publication number Publication date
CN111695674A (zh) 2020-09-22
CN111695674B (zh) 2024-04-09

Similar Documents

Publication Publication Date Title
WO2021114618A1 (fr) Procédé et appareil d'apprentissage fédéré, dispositif informatique et support de stockage lisible
He et al. Neural factorization machines for sparse predictive analytics
WO2022089256A1 (fr) Procédé, appareil et dispositif de formation de modèle de réseau neuronal fédéré, ainsi que produit programme d'ordinateur et support de stockage lisible par ordinateur
Younus et al. Video steganography using knight tour algorithm and LSB method for encrypted data
US20210117417A1 (en) Real-time content analysis and ranking
Bhme Advanced statistical steganalysis
WO2021114616A1 (fr) Procédé d'apprentissage de modèle d'apprentissage fédéré et dispositif associé
CA3059610A1 (fr) Calcul multipartite securise sans initialisateur de confiance
US8918382B1 (en) Learning common spelling errors through content matching
CN111026858B (zh) 基于项目推荐模型的项目信息处理方法及装置
WO2021174922A1 (fr) Procédé de classification de sentiments d'énoncé et dispositif associé
CN113011646B (zh) 一种数据处理方法、设备以及可读存储介质
WO2021120779A1 (fr) Procédé et appareil de construction de profil d'utilisateur basés sur le dialogue homme-machine, terminal et support de stockage
US12026614B2 (en) Interpretable tabular data learning using sequential sparse attention
WO2021196474A1 (fr) Procédé de profilage d'intérêt d'utilisateur et dispositif associé
Aparna et al. A blind medical image watermarking for secure e-healthcare application using crypto-watermarking system
US20230006819A1 (en) Systems and methods for homomorphic encryption-based triggering
US10902215B1 (en) Social hash for language models
Duan et al. A coverless steganography method based on generative adversarial network
KR102254097B1 (ko) 법률 컨텐츠 제공을 위한 지능적인 네트워크 어레인징 시스템 및 방법
US20200320343A1 (en) Method for automatically identifying signals or patterns in time series data by treating series as image
US20240054141A1 (en) Authoritative Factual Service for Blockchain Smart Contracts
US20230186353A1 (en) System for intelligent assessment models for non-fungible electronic resources
US11366825B1 (en) Method and system for managing relational transactions in a workspace inventory and analytics service
Hu et al. Multi-modal steganography based on semantic relevancy

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20898489

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20898489

Country of ref document: EP

Kind code of ref document: A1