Telecommunication fraud detection model training method and device based on block chain and federal learning

Abstract

The invention discloses a telecommunication fraud detection model training method and device based on a block chain and federated learning, and the method comprises the steps: obtaining a model training task issued by a model demander, determining a task participant and a federated committee according to the model training task, and distributing corresponding first public key and private key secret shares for the task participant and the federated committee; downloading a current global model from an alliance chain through a task participant, training the current global model according to a local data set to obtain a local local model, encrypting the local local model according to a first public key to obtain an encrypted local model, and uploading the encrypted local model to the alliance chain; and downloading the encrypted local models from the alliance chain through a federal committee, aggregating the encrypted local models to obtain an encrypted global model, decrypting the encrypted global model according to the private key secret share to obtain an optimized current global model, and uploading the optimized current global model to the alliance chain. The method improves the safety and reliability of telecommunication fraud detection model training, and can be applied to the technical field of communication.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a telecommunications fraud detection model training method and device based on blockchain and federated learning. Background Art

[0002] With the improvement of the level of network intelligence in modern society, traditional criminal activities have been moving towards cyberspace. Telecom fraud, a non-contact crime carried out by means of communication and network technology, has become a hot and difficult topic of social concern. Research and governance of telecommunications fraud methods have become a necessary and urgent need. In the development of business, the communications industry carries a large amount of data generated by a large number of users and devices. These data resources are rich and diverse, including high-value data such as blacklist data, malicious websites, user tag libraries, call data, Internet data, and network experience. If the communications industry data can be used to share and participate in the collaborative construction of anti-telecom fraud detection models, it will be of great help in combating fraud crimes. However, in real application scenarios, on the one hand, due to industry competition, conflicts of interest and other reasons, most of the telecom operator industry data is still in the form of information islands; on the other hand, because the telecom operator industry data involves user privacy and security issues, sharing telecom operator data for collaborative anti-fraud governance needs to consider data privacy protection and security issues.

[0003] Federated learning applies the concept of distribution to machine learning, and multiple participants work together to complete model training with the assistance of a central server. Its main advantage is that the training data of all participants is retained locally, and the participants will not directly exchange local data sets during the training process, but only exchange model gradient information during training, and finally train a model that is close to the effect of centralized machine learning. However, since federated learning uses a central server for model aggregation, the gradient information of the model aggregation process may be maliciously tampered with, thus facing the risk of information being easily tampered with and data privacy being easily leaked. Summary of the invention

[0004] The purpose of the present invention is to solve one of the technical problems existing in the prior art to at least a certain extent.

[0005] To this end, one purpose of an embodiment of the present invention is to provide a telecommunications fraud detection model training method based on blockchain and federated learning, which improves the security and reliability of telecommunications fraud detection model training.

[0006] Another object of an embodiment of the present invention is to provide a telecommunications fraud detection model training device based on blockchain and federated learning.

[0007] In order to achieve the above technical objectives, the technical solutions adopted by the embodiments of the present invention include:

[0008] On the one hand, an embodiment of the present invention provides a telecommunications fraud detection model training method based on blockchain and federated learning, comprising the following steps:

[0009] Obtain a model training task issued by a model demander, determine a task participant and a federal committee according to the model training task, and allocate corresponding first public key and private key secret shares to the task participant and the federal committee;

[0010] The task participant downloads the current global model from the alliance chain, trains the current global model according to the local data set to obtain a local partial model, encrypts the local partial model according to the first public key to obtain an encrypted partial model, and uploads the encrypted partial model to the alliance chain;

[0011] The encrypted local model is downloaded from the alliance chain by the federal committee, the encrypted local model is aggregated to obtain an encrypted global model, and then the encrypted global model is decrypted according to the private key secret share to obtain the optimized current global model, and the optimized current global model is uploaded to the alliance chain.

[0012] Furthermore, in one embodiment of the present invention, the telecommunications fraud detection model training method further includes the step of registering to join the alliance chain, and the specific steps are:

[0013] Receive a consortium chain registration request from a user, where the consortium chain registration request includes an identity identifier and a current timestamp of the user;

[0014] Verifying the identity and the current timestamp through the alliance chain;

[0015] When the verification is successful, the user is determined to be a member of the alliance chain, and a second public key and a second private key are allocated to the user.

[0016] Further, in one embodiment of the present invention, the step of obtaining a model training task issued by a model demander, determining a task participant and a federal committee according to the model training task, and allocating corresponding first public keys and private key secret shares to the task participants and the federal committee specifically includes:

[0017] Acquire the model training task published by the model demander, where the model training task includes an initial global model and task incentives;

[0018] Determine a plurality of task participants from members of the alliance chain according to the model training task, and select a plurality of task participants to form the federal committee according to reputation values, and then generate a homomorphic key, wherein the homomorphic key includes the first public key and a corresponding first private key;

[0019] The first private key is divided into a plurality of the private key secret shares according to the number of the task participants, and the first public key and the private key secret shares are distributed to the corresponding task participants.

[0020] Further, in one embodiment of the present invention, the training of the current global model according to the local data set to obtain the local local model specifically includes:

[0021] Dividing the local data set into a training set and a test set;

[0022] According to the training set, the current global model is trained by a gradient boosting tree algorithm to obtain a current local model;

[0023] The current local model is tested according to the test set to obtain the local local model.

[0024] Further, in one embodiment of the present invention, encrypting the local partial model according to the first public key to obtain an encrypted partial model, and uploading the encrypted partial model to the alliance chain specifically includes:

[0025] Encrypting the local partial model according to the first public key to obtain a first partial model;

[0026] Sign the first local model according to the identity identifier, the local training duration, and the second private key to obtain the encrypted local model;

[0027] The encrypted local model is uploaded to the alliance chain, so that the alliance chain verifies the encrypted local model according to the second public key, and receives the encrypted local model when the verification passes.

[0028] Further, in one embodiment of the present invention, the aggregating the encrypted local models to obtain the encrypted global model, and then decrypting the encrypted global model according to the private key secret share to obtain the optimized current global model, specifically includes:

[0029] Randomly selecting a member from the federal committee as a federal committee leader, and performing aggregation calculation on a plurality of the encrypted local models by the federal committee leader to obtain the encrypted global model;

[0030] Aggregate validation of the cryptographic global model by all members of the federal committee;

[0031] When the aggregate verification succeeds, the corresponding decryption share is calculated according to the private key secret share, and then all members of the federal committee collaboratively decrypt the encrypted global model according to the corresponding decryption share to obtain the optimized current global model.

[0032] Furthermore, in one embodiment of the present invention, the telecommunications fraud detection model training method further includes the following steps:

[0033] Calculate the reward value and reputation value incentive of each task participant according to the preset incentive mechanism;

[0034] Each of the task participants is rewarded according to the reward value, and the reputation value of each of the task participants is updated according to the reputation value incentive.

[0035] On the other hand, an embodiment of the present invention provides a telecommunications fraud detection model training device based on blockchain and federated learning, comprising:

[0036] A task publishing module, used to obtain a model training task published by a model demander, determine a task participant and a federal committee according to the model training task, and allocate corresponding first public keys and private key secret shares to the task participants and the federal committee;

[0037] A local model training module, used for downloading the current global model from the alliance chain through the task participant, training the current global model according to the local data set to obtain a local local model, further encrypting the local local model according to the first public key to obtain an encrypted local model, and uploading the encrypted local model to the alliance chain;

[0038] A global model aggregation module is used to download the encrypted local model from the alliance chain through the federal committee, aggregate the encrypted local model to obtain an encrypted global model, and then decrypt the encrypted global model according to the private key secret share to obtain the optimized current global model, and upload the optimized current global model to the alliance chain.

[0039] On the other hand, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory. When the program is executed by the processor, the telecommunications fraud detection model training method based on blockchain and federated learning as described above is realized.

[0040] On the other hand, an embodiment of the present invention further provides a storage medium, which is a computer-readable storage medium for computer-readable storage, and which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the telecommunications fraud detection model training method based on blockchain and federated learning as described above.

[0041] The advantages and beneficial effects of the present invention will be partly given in the following description, partly become apparent from the following description, or be understood through the practice of the present invention:

[0042] The embodiment of the present invention obtains a model training task issued by a model demander, determines a task participant and a federal committee according to the model training task, and allocates corresponding first public keys and private key secret shares to the task participants and the federal committee, downloads the current global model from the alliance chain through the task participants, trains the current global model according to a local data set to obtain a local local model, and then encrypts the local local model according to the first public key to obtain an encrypted local model, and uploads the encrypted local model to the alliance chain, downloads the encrypted local model from the alliance chain through the federal committee, aggregates the encrypted local model to obtain an encrypted global model, and then decrypts the encrypted global model according to the private key secret share to obtain an optimized current global model, and uploads the optimized current global model to the alliance chain. The embodiment of the present invention combines blockchain technology with federated learning and cryptographic mechanisms. Since blockchain has the characteristics of decentralization, tamper-proofness, and traceability, it can complement federated learning well and provide privacy protection for collaborative training. It can effectively solve security issues such as a single data source for training telecommunications fraud detection models and easy leakage of data privacy. It is conducive to multi-party collaboration between telecommunications operators and public security departments, financial institutions, Internet companies and other organizations in a decentralized environment, and to jointly train telecommunications fraud detection models safely and reliably, thereby realizing the value circulation of telecommunications operator data, improving the security and reliability of telecommunications fraud detection model training, and effectively protecting the data privacy of each participant while achieving efficient detection of telecommunications fraud. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solution in the embodiments of the present invention, the following introduction is made to the drawings required for use in the embodiments of the present invention. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solution of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A flowchart of a method for training a telecommunications fraud detection model based on blockchain and federated learning provided in an embodiment of the present invention;

[0045] Figure 2 A schematic diagram of multi-party interaction of a telecommunications fraud detection model training method based on blockchain and federated learning provided in an embodiment of the present invention;

[0046] Figure 3 A timing diagram of a telecommunications fraud detection model training method based on blockchain and federated learning provided in an embodiment of the present invention;

[0047] Figure 4 A flowchart of the steps for registering to join a consortium chain provided in an embodiment of the present invention;

[0048] Figure 5 A step flow chart of step S101 provided in an embodiment of the present invention;

[0049] Figure 6 A step flow chart of step S102 provided in an embodiment of the present invention;

[0050] Figure 7 Another step flow chart of step S102 provided in an embodiment of the present invention;

[0051] Figure 8 A step flow chart of step S103 provided in an embodiment of the present invention;

[0052] Fig. 9 Another step flow chart of a telecommunications fraud detection model training method based on blockchain and federated learning provided in an embodiment of the present invention;

[0053] Fig.10 A schematic diagram of the structure of a telecommunications fraud detection model training device based on blockchain and federated learning provided in an embodiment of the present invention;

[0054] Fig.11 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention;

[0055] Fig.12 A schematic diagram of the structure of a storage medium provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] Embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as limitations on the present application. It should be noted that, although the functional module division is performed in the system schematic diagram and the logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in a different order from the module division in the system schematic diagram or the flow chart. For the step numbers in the following embodiments, they are only set for the convenience of explanation, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0057] In the description of the present invention, the meaning of "a plurality" is two or more. If there is a description of the first or the second, it is only for the purpose of distinguishing the technical features, and it cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features. In addition, unless otherwise defined, all technical and scientific terms used in this document have the same meaning as those commonly understood by technicians in the technical field of this application. The terms used in this document are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0058] The telecommunications fraud detection model training method based on blockchain and federated learning provided in the embodiments of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a set-top box, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the telecommunications fraud detection model training method based on blockchain and federated learning, etc., but is not limited to the above forms.

[0059] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0060] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with the relevant laws, regulations, and standards of the relevant countries and regions. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0061] The implementation process of the embodiment of the present invention includes four main entities: telecommunications fraud detection model demander MD, task participant TP, consortium chain CBC and federal committee FC.

[0062] Model Demander (MD): Responsible for publishing the training requirements for telecommunications fraud detection models. The model demander first selects a machine learning model and initializes the selected model; then organizes participants to join the telecommunications fraud detection model training team. The model demander will eventually obtain an efficient telecommunications fraud detection model.

[0063] Task Participant (TP): Participants have some of the data required for telecommunications fraud detection model training, and can be organizations such as telecommunications operators, public security departments, and financial institutions. Public security departments have mastered customer data of telecommunications fraud in investigating cases, telecommunications operators have customer communication data, and financial institutions have accumulated customer behavior data in the financial field. However, due to the limited data resources they hold and security issues such as data privacy protection, they cannot independently complete the entire telecommunications fraud training task, or the model trained through a single data resource is poor.

[0064] Consortium Blockchain (CBC): Telecom operators, public security departments, financial institutions and other organizations are selected as consensus nodes to form a consortium chain. The consortium chain has an access mechanism. The system needs to conduct a qualification review of the nodes joining the consortium chain. Only users who pass the review can join the consortium chain, which can avoid attacks and damage from malicious nodes.

[0065] Federal Council (FC): It is composed of a group of participants with high credibility and is responsible for the aggregation of intermediate parameters of the telecommunications fraud detection model, collaborative decryption, and allocation of incentives.

[0066] like Figure 1 The figure is a flowchart of a method for training a telecommunications fraud detection model based on blockchain and federated learning provided by an embodiment of the present invention, referring to Figure 1 The embodiment of the present invention provides a telecommunications fraud detection model training method based on blockchain and federated learning, which specifically includes the following steps:

[0067] S101. Obtain the model training task issued by the model demander, determine the task participants and the federal committee according to the model training task, and allocate corresponding first public key and private key secret shares to the task participants and the federal committee;

[0068] S102, downloading the current global model from the alliance chain through the task participant, training the current global model according to the local data set to obtain a local local model, further encrypting the local local model according to the first public key to obtain an encrypted local model, and uploading the encrypted local model to the alliance chain;

[0069] S103. The encrypted local model is downloaded from the alliance chain through the federal committee, the encrypted local model is aggregated to obtain the encrypted global model, and then the encrypted global model is decrypted according to the private key secret share to obtain the optimized current global model, and the optimized current global model is uploaded to the alliance chain.

[0070] The above steps S102 to S103 are iterated repeatedly until the model converges or reaches a preset accuracy. Finally, the model demander downloads the final telecommunications fraud detection model from the alliance chain.

[0071] Specifically, Figure 2 The figure shows a multi-party interaction diagram of a telecommunications fraud detection model training method based on blockchain and federated learning provided by an embodiment of the present invention, wherein:

[0072] Telecom fraud detection model demander MD: Publish the training requirements for the telecom fraud detection model, select a machine learning model, upload its initial parameters to the alliance chain, and pay a token as a prize pool. Participants who are interested in this and hold training data sets can apply to join the federation training task. As more and more participants join the model training team, the model demander will eventually obtain a more efficient telecom fraud detection model.

[0073] Task participant TP: has the data required for the telecommunications fraud federated learning task, and can be a telecommunications operator, public security department, financial institution and other organizations. The public security department has mastered the customer data of telecommunications fraud in the investigation of the case, the telecommunications operator has the customer's communication data, and the financial institution has accumulated the customer's behavior data in the financial field. However, due to the limited data resources held by themselves and security issues such as data privacy protection, they cannot independently complete the entire training task, or the model trained by a single data resource is poor. Participants can apply to join the federal training team and become a member of the federated learning after being reviewed and approved by the demander. When uploading local model parameters to the alliance chain, task participants must also submit information such as the size of their local training data set and the time required for training. The system will use this as a basis for examining the contribution of participants.

[0074] Alliance chain CBC: Telecom operator A, telecom operator B, telecom operator C, public security departments, bank A, bank B, bank C and other organizations are used as consensus nodes to form an alliance chain. Because the alliance chain has an access mechanism, the system will conduct a qualification review of the nodes joining the alliance chain. Only users who pass the review can join the alliance chain, thereby avoiding attacks and damage from malicious nodes. The blockchain ledger is jointly maintained by each user in the system. The failure of a single node in the system will not affect the operation of the entire system, which avoids the impact of a single point of failure on the system. In addition, due to the characteristics of blockchain information disclosure, traceability, and non-repudiation, the system's historical incentive evaluation information for all participants is also stored on the alliance chain, making model training more transparent, reliable, and fair.

[0075] Federal Committee FC: Initially, the system randomly selects a group of participants from the federated learning team to form a federal committee, and randomly selects a member as the leader of the training task. The federal committee downloads the local model parameters of each participant from the alliance chain, aggregates them, and then collaboratively decrypts the aggregated model and uploads it to the alliance chain. When the leader of the federal committee goes offline or the term set by the system expires, the system will re-rank the users participating in the current training task according to their reputation, and then select the best to generate a new federal committee and leader. The main work of the federal committee is to be responsible for model parameter aggregation, collaborative decryption, and incentive allocation.

[0076] It can be recognized that the embodiments of the present invention combine blockchain technology with federated learning and cryptographic mechanisms. Since blockchain has the characteristics of decentralization, tamper-proofing, and traceability, it can complement federated learning well and provide privacy protection for collaborative training. It can effectively solve security issues such as a single data source for training telecommunications fraud detection models and easy leakage of data privacy. It is conducive to multi-party collaboration between telecommunications operators and public security departments, financial institutions, Internet companies and other organizations in a decentralized environment, and to conduct joint training of telecommunications fraud detection models in a safe and reliable manner, thereby realizing the value circulation of telecommunications operators' data, improving the security and reliability of telecommunications fraud detection model training, and effectively protecting the data privacy of each participant while achieving efficient detection of telecommunications fraud.

[0077] like Figure 4 The figure is a flowchart of the steps of registering to join the alliance chain provided by an embodiment of the present invention, referring to Figure 4 As an optional implementation, the telecommunications fraud detection model training method further includes the step of registering to join the alliance chain, and the specific steps are:

[0078] S201. Receive a consortium chain registration request from a user, where the consortium chain registration request includes the user's identity and current timestamp;

[0079] S202, verifying the identity and current timestamp through the alliance chain;

[0080] S203: When the verification is successful, the user is determined to be a member of the alliance chain, and a second public key and a second private key are allocated to the user.

[0081] Specifically, suppose there are N users in the system, with u1, u2, L, u N Indicates that user u i ,i∈[1,N] selects an identity UID i , get the current system timestamp ut i , then {UID i ,ut i}Send to the alliance chain to apply for registration. When the alliance link receives the user's registration application, it verifies ut i and UID i If it is valid, the registration request is accepted and the user is assigned a system public-private key pair (pk i ,sk i ), otherwise the user's registration request is rejected.

[0082] like Figure 5 FIG. 1 is a flowchart of step S101 provided in an embodiment of the present invention, referring to FIG. Figure 5 As an optional implementation method, a model training task issued by a model demander is obtained, task participants and a federal committee are determined according to the model training task, and corresponding first public key and private key secret shares are allocated to the task participants and the federal committee, which specifically includes:

[0083] S1011. Obtain the model training task issued by the model demander, where the model training task includes an initial global model and task incentives;

[0084] S1012. Determine multiple task participants from members of the alliance chain according to the model training task, and select multiple task participants to form a federal committee according to the reputation value, and then generate a homomorphic key, where the homomorphic key includes a first public key and a corresponding first private key;

[0085] S1013. Divide the first private key into a plurality of private key secret shares according to the number of task participants, and distribute the first public key and the private key secret shares to corresponding task participants.

[0086] Specifically, the telecommunications fraud detection model demander publishes a model learning task Task and selects the initialization model parameter Δw 0 , and pay a certain amount of tokens b T Uploaded to the alliance chain as a task incentive. Participants apply to join the training task based on their own data resource information and pay a certain amount of tokens b u As a deposit, if the user makes a wrong move, a certain percentage of tokens will be deducted as punishment.

[0087] Assume that the model demander selects k task participants u1,u2,L,u k The model demander selects M = k / 2 members from the federated learning team to form a federated committee and generates a homomorphic key (pk m ,sk m ). Then use the private key sk m Divide into k secret shares and share them with task participants, whose set is S = {s1, s2, L, s k}, each participant ui Will get the public key pk m and secret share s i .

[0088] In some optional embodiments, the method for generating a homomorphic key is:

[0089] Randomly select two large prime numbers p and q, and let n = pq, where p = 2p'+1, q = 2q'+1, Let m = p'q', any The private key sk m =βm, randomly selected Let g = (1 + n) a b n modn 2 , θ=amβmodn, generate the public key.

[0090] In some optional embodiments, the method for dividing the shared secret shares is:

[0091] Let a0 = βm, a i Randomly select from {0,1,L,nm-1} and set sk m Divide it into k secret shares for sharing, and the secret share calculation formula is as follows:

[0092]

[0093] Then s i Distribute to the corresponding participants u i , each participant u i Will get the public key pk m and secret share s i After the distribution is completed, the private key sk m destroy.

[0094] In some optional embodiments, the method for forming the federation committee and electing the federation learning leader is: sort the participants according to their reputation from high to low, and select the top 1 / 2 task participants with the highest reputation to form the federation committee. Initially, the leader of the federation committee is randomly selected from the members of the federation committee. If the leader fails or quits midway, the next federation learning leader will be selected from the federation committee in a round-robin manner.

[0095] like Figure 6 FIG. 1 is a flow chart of step S102 provided in an embodiment of the present invention, referring to FIG. Figure 6 As an optional implementation, the current global model is trained according to the local data set to obtain a local local model, which specifically includes:

[0096] S1021. Divide the local data set into a training set and a test set;

[0097] S1022. According to the training set, the current global model is trained by a gradient boosting tree algorithm to obtain a current local model;

[0098] S1023. Test the current local model according to the test set to obtain a local local model.

[0099] Specifically, the gradient boosting tree machine learning algorithm is selected to train the local model. The gradient boosting tree algorithm uses one or more of the F1 value, accuracy, recall, and AUC indicators to evaluate the effect of the model. The local data of the participants is divided into a training set and a test set, with the ratio of the test set to the training set being 3:7.

[0100] In some optional embodiments, the LightGBM gradient boosting tree algorithm can be used to train the local model, and the core parameters of the LightGBM algorithm are set as follows: (boosting_type, max_depth, num_leaves, min_child_samples, feature_fraction, bagging_fraction, min_child_weight, bagging_freq), where boosting_type is fixed to gbdt, parameter max_depth = 8, and the remaining parameters are selected according to the actual situation and meet the condition num_leaves ≤ 2 max_depth , parameter optimization and iteration are continuously performed during the training process.

[0101] In some optional embodiments, the XGBoost gradient boosting tree algorithm can also be used to train the local model. The core parameters of the XGBoost algorithm are set to: (objective, booster, eval_metric, learning_rate, max_depth, subsample, colsample_bytree, reg_alpha, reg_lambda), where the objective parameter is set to binary:logistic, the booster parameter is set to gbtree, the parameter max_depth=8, and the remaining parameters are selected according to the actual situation, and the parameters are continuously optimized and iterated during the training process.

[0102] In some optional embodiments, the CatBoost gradient boosting tree algorithm can also be used to train the local model. The core parameters of the CatBoost algorithm are set to: (bootstrap_type, learning_rate, depth, l2_leaf_reg, subsample, early_stopping_rounds, od_type), where the parameter bootstrap_type is set to Bernoulli, the parameter depth = 8, and the remaining parameters are selected according to the actual situation, and the parameters are continuously optimized and iterated during the training process.

[0103] like Figure 7 FIG. 1 is another step flow chart of step S102 provided in an embodiment of the present invention, referring to FIG. Figure 7 As an optional implementation, the local partial model is encrypted according to the first public key to obtain an encrypted partial model, and the encrypted partial model is uploaded to the alliance chain, which specifically includes:

[0104] S1024. Encrypt the local partial model according to the first public key to obtain a first partial model;

[0105] S1025, signing the first local model according to the identity identifier, the local training duration, and the second private key to obtain an encrypted local model;

[0106] S1026. Upload the encrypted local model to the alliance chain, so that the alliance chain verifies the encrypted local model according to the second public key, and receives the encrypted local model when the verification passes.

[0107] Specifically, task participant u i Download the global model parameter Δw of round T from the consortium chain T , and use it as the input model for the next round of local training. Task participant u i Use the local data set for training and get the intermediate gradient parameters of the local model after training. Random Selection Encrypted to get ciphertext Then use the private key to calculate the signature and will Upload to the alliance chain. When the alliance chain receives the participant u i The uploaded message first verifies the participant's identity UID i and local training duration Is it valid? If it is valid, verify the validity of the signature information through the following equation;

[0108]

[0109] If the equality holds, the upload request is accepted, otherwise the upload is rejected.

[0110] like Figure 8 FIG. 1 is a flow chart of step S103 provided in an embodiment of the present invention, referring to FIG. Figure 8 As an optional implementation, the encrypted local models are aggregated to obtain an encrypted global model, and then the encrypted global model is decrypted according to the private key secret share to obtain an optimized current global model, which specifically includes:

[0111] S1031, randomly selecting a member from the federal committee as a leader of the federal committee, and performing aggregation calculation on multiple encrypted local models through the leader of the federal committee to obtain an encrypted global model;

[0112] S1032. Aggregate verification of the global model of encryption by all members of the Federal Committee;

[0113] S1033. When the aggregate verification succeeds, the corresponding decryption share is calculated according to the private key secret share, and then all members of the federal committee collaboratively decrypt the encrypted global model according to the corresponding decryption share to obtain the optimized current global model.

[0114] Specifically, the leader of the federal committee downloads the local model parameter ciphertext of all participants from the alliance chain, and then uses the incentive mechanism to calculate the participant u i Contribution reward, if u i Falsely reporting the amount of data will result in a certain amount of deposit being frozen as a fine. After calculating the data contribution rewards of all participants, according to the homomorphic properties of the Paillier algorithm for addition and multiplication, the leader will aggregate the local model parameters to form the global model parameter ciphertext C(Δw T+1 ), the system will allocate a certain amount of computing contribution rewards to the leader.

[0115] The leader of the Federal Committee encrypts the parameters of the local fraud detection model of k task participants based on the additive homomorphism of the homomorphic encryption algorithm. Perform aggregate calculations to form global model parameter ciphertext Each member of the Federal Council calculates C'(Δw T+1 ), when at least M members have calculated C'(Δw T+1 )=C(Δw T+1 ) when the aggregation result is verified.

[0116] Then, the members of the Federal Committee use their respective secret shares s j Calculate decryption share When at least t committee members give the correct decryption share, the global plaintext model M(ΔwT ) and upload it to the consortium chain.

[0117] Assuming that the number of members of the Federal Committee is t, if the effective decryption share c received by the Federal Committee members i If there are less than t shares, joint decryption is impossible; if the number of valid shares received is not less than t, let S′ be the set of t valid shares, and the plaintext can be calculated as:

[0118]

[0119] in

[0120] like Fig. 9 FIG. 1 is another flow chart of the steps of a telecommunications fraud detection model training method based on blockchain and federated learning provided by an embodiment of the present invention, referring to FIG. Fig. 9 As an optional implementation, the telecommunications fraud detection model training method further includes the following steps:

[0121] S104, calculating the reward value and reputation value incentive of each task participant according to the preset incentive mechanism;

[0122] S105. Reward each task participant according to the reward value, and update the reputation value of each task participant according to the reputation value incentive.

[0123] Specifically, the embodiment of the present invention uses an incentive mechanism to attract participants to actively join the collaborative training process. The specific incentive mechanism is divided into two parts: reputation value incentive and reward value.

[0124] The first part of the incentive mechanism is that the credit value incentive calculation method for task participants who complete the training task is:

[0125]

[0126] Among them, i represents the i-th task participant, j represents the j-th training task, represents the reputation reward obtained by the i-th task participant for completing the j-th training task; the weight η reflects the data contribution of the participant's local training, represents the task duration (the training time required for participants with larger local datasets) The size of the training data set is positively correlated with the training time). j ) represents the accuracy of the global model obtained by all participants in the jth training, rpt i represents the total credit value reward of the i-th task participant, which is initially set to 0; m represents the total number of task trainings participated in by the i-th task participant.

[0127] The second part of the incentive mechanism is to calculate the reward value obtained by participants after completing the training task as follows:

[0128]

[0129] in represents the reward value of the i-th model participant completing the j-th task, b j represents the basic reward for the jth task, δ i represents the reputation reward coefficient of the i-th task participant; n represents the number of participants in the collaborative training task, and RWD represents the total reward value of the current task set by the system.

[0130] The above is a description of the method flow of the embodiment of the present invention. Figure 3 The following is a timing diagram of a telecommunications fraud detection model training method based on blockchain and federated learning provided by an embodiment of the present invention. Figure 3 The implementation process of the embodiment of the present invention is further described.

[0131] 1) User registration. When a user applies to join the alliance chain, the alliance chain will review the user's qualifications. Users who pass the review will obtain the user's public and private keys (pk i ,sk i ), and successfully registered to join the system, avoiding attacks and damages from malicious nodes. Suppose there are N users in the alliance chain system, namely u1, u2, L, u N The users in the alliance chain mainly include four types: telecommunications fraud detection model demanders MD, task participants TP, alliance chain CBC, and federal committee FC.

[0132] 2) Construction of telecommunications fraud training tasks. The telecommunications fraud detection model demander MD publishes a model training task Task and selects the initialization model parameter Δw 0 , and pay a certain amount of tokens b T Uploaded to the alliance chain as a task incentive. Task participants TP apply to join the training task based on their own data resources and pay a certain amount of tokens b u As a deposit, if the user makes a wrong move, a certain percentage of tokens will be deducted as punishment.

[0133] Telecom fraud detection model demander MD selects k task participants u1,u2,L,u k To form a federated learning team, MD selects M from the federated learning team based on the reputation value of TP (where ) members form the Federal Committee FC and generate a homomorphic key (pk m ,sk m ). Then use the private key sk mDivide into k secret shares and share them with TP, that is, S = {s1, s2, L, s k}, each participant u i Will get the public key pk m and secret share s i .

[0134] 3) Local telecommunications fraud detection model is uploaded to the blockchain. Task participants i Download the global model parameters Δw of the T-th round of training from the alliance chain CBC T , and use it as the input model for the next round of local training. i Use the local data set for training, and get the intermediate gradient parameters of the local fraud detection model after training. Encrypt the local model gradient parameters to obtain ciphertext Then calculate the signature and will Upload to the alliance chain. When the alliance chain receives the participant u i The uploaded message first verifies the participant's identity UID i and local training duration Is it valid? If valid, verify the validity of the signature information through the following equation:

[0135]

[0136] If the equality holds, the upload request is accepted, otherwise the upload is rejected.

[0137] 4) Global model aggregation. The leader of the federal committee downloads the local model parameter ciphertext of all participants from the alliance chain, and then uses the incentive mechanism to calculate the participant u i Contribution reward, if u i Falsely reporting the amount of data will result in a certain amount of deposit being frozen as a fine. After calculating the data contribution rewards of all participants, the leader will aggregate the local model parameters to form the global model parameter ciphertext C(Δw T+1 ), the system will allocate a certain amount of computing contribution reward to the leader. Each member of the Federal Committee calculates C'(Δw T+1 ), if and only if at least M members have calculated C'(Δw T+1 )=C(Δw T+1 ) when the aggregation result is verified. Then the members of the Federal Committee use the secret shares they hold s j Calculate the decryption share c j , when at least t committee members give the correct decryption share, the global plaintext model M(Δw T ) and upload it to the alliance chain CBC.

[0138] 5) Repeat the above steps 3) to 4) until the model converges or reaches the preset accuracy. Finally, the model demander MD downloads the final telecommunications fraud detection model from the alliance chain.

[0139] It can be recognized that the present invention uses federated learning to jointly model and train multi-source fraud data, and uses blockchain to replace the central server to give federated learning anti-tampering and anti-single point failure attack characteristics, so as to achieve the management of the training process. Collaborative training participants use local data sets to train telecommunications fraud detection models. The privacy data of telecommunications operators, public security departments, financial institutions, etc. do not need to leave the local database. Homomorphic encryption and secret sharing algorithms are used to protect the privacy of intermediate model parameters of training participants. Model aggregation and collaborative decryption are performed through the federal committee to provide privacy protection for collaborative training. An incentive mechanism is established to ensure the fairness of collaborative training, and task participants are encouraged to join the collaborative training of federated learning to improve the accuracy of telecommunications fraud detection model training. Combining blockchain technology with federated learning and cryptography mechanisms can form a good complement, which is conducive to multi-party collaboration between telecommunications operators and public security departments, financial institutions, Internet companies and other organizations in a decentralized environment, and to conduct joint anti-fraud model training safely and reliably, so as to realize the value circulation of telecommunications operator data, effectively protect the data privacy of each participant, and realize effective detection of telecommunications fraud.

[0140] like Fig.10 The diagram is a schematic diagram of the structure of a telecommunications fraud detection model training device based on blockchain and federated learning provided by an embodiment of the present invention, referring to Fig.10 The embodiment of the present invention provides a telecommunications fraud detection model training device based on blockchain and federated learning, including:

[0141] The task publishing module is used to obtain the model training tasks published by the model demander, determine the task participants and the federal committee according to the model training tasks, and allocate the corresponding first public key and private key secret shares to the task participants and the federal committee;

[0142] The local model training module is used to download the current global model from the alliance chain through the task participants, train the current global model according to the local data set to obtain the local local model, and then encrypt the local local model according to the first public key to obtain the encrypted local model, and upload the encrypted local model to the alliance chain;

[0143] The global model aggregation module is used to download the encrypted local model from the alliance chain through the federal committee, aggregate the encrypted local model to obtain the encrypted global model, and then decrypt the encrypted global model according to the private key secret share to obtain the optimized current global model, and upload the optimized current global model to the alliance chain.

[0144] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0145] The embodiment of the present invention also provides an electronic device, which includes: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory, and when the program is executed by the processor, the above-mentioned telecommunications fraud detection model training method based on blockchain and federated learning is realized. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0146] like Fig.11 FIG. 1 is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention, referring to FIG. Fig.11 , an embodiment of the present invention provides an electronic device, including:

[0147] The processor 1101 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention;

[0148] The memory 1102 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1102 can store an operating system and other applications. When the technical solution provided in the embodiments of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1102, and the processor 1101 calls and executes the telecommunications fraud detection model training method based on blockchain and federated learning in the embodiments of the present invention;

[0149] Input / output interface 1103, used to implement information input and output;

[0150] The communication interface 1104 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0151] A bus 1105 that transmits information between various components of the device (e.g., the processor 1101, the memory 1102, the input / output interface 1103, and the communication interface 1104);

[0152] The processor 1101 , the memory 1102 , the input / output interface 1103 and the communication interface 1104 are connected to each other in communication within the device via the bus 1105 .

[0153] like Fig.12 FIG. 1 is a schematic diagram of the structure of a storage medium provided in an embodiment of the present invention, referring to FIG. Fig.12 An embodiment of the present invention further provides a storage medium, which is a computer-readable storage medium used for computer-readable storage. The storage medium stores one or more programs 1201, and the one or more programs 1201 can be executed by one or more processors to implement the above-mentioned telecommunications fraud detection model training method based on blockchain and federated learning.

[0154] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0155] The embodiment of the present invention also discloses a computer program product or a computer program, wherein the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes Figure 1 The method shown.

[0156] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the above-mentioned boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.

[0157] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified to the contrary, one or more of the above-mentioned functions and / or features can be integrated into a single physical device and / or software module, or one or more functions and / or features can be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.

[0158] If the above functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the above methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0159] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0160] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and editable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the above-mentioned program is printed, since the above-mentioned program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or processing in other suitable ways as necessary, and then stored in a computer memory.

[0161] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0162] In the above description of this specification, the description with reference to the terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0163] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

[0164] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A telecommunications fraud detection model training method based on blockchain and federated learning, characterized in that: The following steps are involved: Obtain a model training task issued by a model demander, determine a task participant and a federal committee according to the model training task, and allocate corresponding first public key and private key secret shares to the task participant and the federal committee; The task participant downloads the current global model from the alliance chain, trains the current global model according to the local data set to obtain a local partial model, encrypts the local partial model according to the first public key to obtain an encrypted partial model, and uploads the encrypted partial model to the alliance chain; The encrypted local model is downloaded from the alliance chain by the federal committee, the encrypted local model is aggregated to obtain an encrypted global model, and then the encrypted global model is decrypted according to the private key secret share to obtain the optimized current global model, and the optimized current global model is uploaded to the alliance chain.

2. According to claim 1, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The telecommunications fraud detection model training method also includes the step of registering to join the alliance chain, and the specific steps are: Receive a consortium chain registration request from a user, where the consortium chain registration request includes an identity identifier and a current timestamp of the user; Verifying the identity and the current timestamp through the alliance chain; When the verification is successful, the user is determined to be a member of the alliance chain, and a second public key and a second private key are allocated to the user.

3. According to claim 2, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The step of obtaining a model training task issued by a model demander, determining a task participant and a federal committee according to the model training task, and allocating corresponding first public keys and private key secret shares to the task participants and the federal committee specifically includes: Acquire the model training task published by the model demander, where the model training task includes an initial global model and task incentives; Determine a plurality of task participants from members of the alliance chain according to the model training task, and select a plurality of task participants to form the federal committee according to reputation values, and then generate a homomorphic key, wherein the homomorphic key includes the first public key and a corresponding first private key; The first private key is divided into a plurality of the private key secret shares according to the number of the task participants, and the first public key and the private key secret shares are distributed to the corresponding task participants.

4. According to claim 1, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The training of the current global model according to the local data set to obtain the local local model specifically includes: Dividing the local data set into a training set and a test set; According to the training set, the current global model is trained by a gradient boosting tree algorithm to obtain a current local model; The current local model is tested according to the test set to obtain the local local model.

5. According to claim 2, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The step of encrypting the local partial model according to the first public key to obtain an encrypted partial model, and uploading the encrypted partial model to the alliance chain specifically includes: Encrypting the local partial model according to the first public key to obtain a first partial model; Sign the first local model according to the identity identifier, the local training duration, and the second private key to obtain the encrypted local model; The encrypted local model is uploaded to the alliance chain, so that the alliance chain verifies the encrypted local model according to the second public key, and receives the encrypted local model when the verification passes.

6. According to claim 3, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The step of aggregating the encrypted local models to obtain an encrypted global model, and then decrypting the encrypted global model according to the private key secret share to obtain the optimized current global model specifically includes: Randomly selecting a member from the federal committee as a federal committee leader, and performing aggregation calculation on a plurality of the encrypted local models by the federal committee leader to obtain the encrypted global model; Aggregate validation of the cryptographic global model by all members of the federal committee; When the aggregate verification succeeds, the corresponding decryption share is calculated according to the private key secret share, and then all members of the federal committee collaboratively decrypt the encrypted global model according to the corresponding decryption share to obtain the optimized current global model.

7. According to claim 6, a telecommunications fraud detection model training method based on blockchain and federated learning is characterized in that: The telecommunications fraud detection model training method also includes the following steps: Calculate the reward value and reputation value incentive of each task participant according to the preset incentive mechanism; Each of the task participants is rewarded according to the reward value, and the reputation value of each of the task participants is updated according to the reputation value incentive.

8. A telecommunications fraud detection model training device based on blockchain and federated learning, characterized in that: include: A task publishing module, used to obtain a model training task published by a model demander, determine a task participant and a federal committee according to the model training task, and allocate corresponding first public keys and private key secret shares to the task participants and the federal committee; A local model training module, used for downloading the current global model from the alliance chain through the task participant, training the current global model according to the local data set to obtain a local local model, further encrypting the local local model according to the first public key to obtain an encrypted local model, and uploading the encrypted local model to the alliance chain; A global model aggregation module is used to download the encrypted local model from the alliance chain through the federal committee, aggregate the encrypted local model to obtain an encrypted global model, and then decrypt the encrypted global model according to the private key secret share to obtain the optimized current global model, and upload the optimized current global model to the alliance chain.

9. An electronic device, characterized in that: The electronic device includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory. When the program is executed by the processor, the steps of the telecommunications fraud detection model training method based on blockchain and federated learning are realized as described in any one of claims 1 to 7.

10. A storage medium, the storage medium being a computer-readable storage medium, used for computer-readable storage, characterized in that: The storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the telecommunications fraud detection model training method based on blockchain and federated learning as described in any one of claims 1 to 7.

Images