CN114819614B

CN114819614B - Data processing method, device, system and equipment

Info

Publication number: CN114819614B
Application number: CN202210429444.1A
Authority: CN
Inventors: 王立
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2022-04-22
Filing date: 2022-04-22
Publication date: 2024-10-15
Anticipated expiration: 2042-04-22
Also published as: CN114819614A; WO2023202496A1

Abstract

The embodiment of the specification provides a data processing method, a device, a system and equipment, wherein the method comprises the following steps: receiving a risk identification result of a server side aiming at a first user, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server side based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user feature data; performing initialization training on a preset second risk identification model based on second characteristic data of the first user and a risk identification result of the first user to obtain a second risk identification model after the initialization training; and carrying out parameter updating processing on the initialized and trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model.

Description

Data processing method, device, system and equipment

Technical Field

The present document relates to the field of data processing technologies, and in particular, to a data processing method, device, system, and apparatus.

Background

Along with the rapid development of computer technology, the user quantity of enterprises such as online banks and the like providing resource transfer services for users is rapidly increased, the data risk faced by the enterprises is also increased, and in order to reduce the data risk, the enterprises can perform risk identification processing by constructing a risk identification model. For example, an enterprise may train a risk identification model based on locally stored user sample data and risk tags, and perform risk identification processing based on the trained risk identification model.

However, enterprises such as online banks may have situations of small data amount of user sample data and risk labels and poor data quality, which may result in poor risk recognition effect and low accuracy of risk recognition of the risk recognition model obtained by training, and therefore, a solution capable of improving the risk recognition effect and accuracy of risk recognition of the risk recognition model is needed.

Disclosure of Invention

An object of the embodiments of the present disclosure is to provide a data processing method, apparatus, system and device, so as to provide a solution capable of improving a risk recognition effect and a risk recognition accuracy of a risk recognition model.

In order to achieve the above technical solution, the embodiments of the present specification are implemented as follows:

In a first aspect, a data processing method includes: receiving a risk identification result of a server side aiming at a first user, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server side based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user feature data; performing initialization training on a preset second risk identification model based on second characteristic data of the first user and a risk identification result of the first user to obtain a second risk identification model after the initialization training; and carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model, so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

In a second aspect, an embodiment of the present disclosure provides a data processing method, including: under the condition that a target user is detected to trigger execution of a target service, acquiring first characteristic data of the target user; inputting first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of the first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user are obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data; and determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

In a third aspect, an embodiment of the present disclosure provides a data processing system, including a server and a client, where: the client is used for sending the user identification of the first user to the server; the server side is configured to obtain a risk identification result of the first user based on the user identifier, where the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by using the server side based on a first risk identification model that is trained in advance, and the first risk identification model that is trained by using the server side based on a preset first number of user feature data; the client is configured to perform initialization training on a preset second risk recognition model based on the second feature data of the first user and a risk recognition result of the first user returned by the server, obtain an initialized trained second risk recognition model, and perform parameter updating processing on the initialized second risk recognition model based on the feature data of the second user and the risk recognition result of the second user, obtain a pre-trained second risk recognition model, so as to perform risk recognition processing on the user based on the pre-trained second risk recognition model.

In a fourth aspect, embodiments of the present disclosure provide a data processing apparatus, including: the risk identification module is used for receiving a risk identification result of a first user aiming at a server side, wherein the risk identification result of the first user is obtained by performing risk identification on first characteristic data of the first user based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user characteristic data; the first training module is used for carrying out initialization training on a preset second risk recognition model based on the second characteristic data of the first user and the risk recognition result of the first user to obtain an initialized second risk recognition model; and the second training module is used for carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

In a fifth aspect, embodiments of the present specification provide a data processing apparatus, including: the data acquisition module is used for acquiring first characteristic data of the target user under the condition that the target user is detected to trigger the execution of the target service; the result acquisition module is used for inputting the first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on the second feature data of the first user, the risk recognition result of the first user, the feature data of the second user and the risk recognition result of the second user, the risk recognition result of the first user is obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data; and the risk determining module is used for determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

In a sixth aspect, embodiments of the present specification provide a data processing apparatus, including: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to: receiving a risk identification result of a server side aiming at a first user, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server side based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user feature data; performing initialization training on a preset second risk identification model based on second characteristic data of the first user and a risk identification result of the first user to obtain a second risk identification model after the initialization training; and carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model, so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

In a seventh aspect, embodiments of the present specification provide a data processing apparatus, the data processing apparatus including: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to: under the condition that a target user is detected to trigger execution of a target service, acquiring first characteristic data of the target user; inputting first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of the first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user are obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data; and determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

In an eighth aspect, the present description provides a storage medium for storing computer-executable instructions that when executed implement the following: receiving a risk identification result of a server side aiming at a first user, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server side based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user feature data; performing initialization training on a preset second risk identification model based on second characteristic data of the first user and a risk identification result of the first user to obtain a second risk identification model after the initialization training; and carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model, so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

In a ninth aspect, embodiments of the present disclosure provide a storage medium for storing computer-executable instructions that, when executed, implement the following: under the condition that a target user is detected to trigger execution of a target service, acquiring first characteristic data of the target user; inputting first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of the first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user are obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data; and determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

Drawings

In order to more clearly illustrate the embodiments of the present description or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some of the embodiments described in the present description, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1A is a flowchart illustrating an embodiment of a data processing method according to the present disclosure;

FIG. 1B is a schematic diagram illustrating a data processing method according to the present disclosure;

FIG. 2 is a schematic diagram illustrating a processing procedure of another data processing method according to the present disclosure;

FIG. 3 is a schematic diagram illustrating an initialization training and parameter updating process of a second risk identification model according to the present disclosure;

FIG. 4A is a flowchart illustrating yet another embodiment of a data processing method according to the present disclosure;

FIG. 4B is a schematic diagram illustrating a processing procedure of another data processing method according to the present disclosure;

FIG. 5 is a schematic diagram illustrating a processing procedure of another data processing method according to the present disclosure;

FIG. 6 is a schematic diagram of a data processing system of the present specification;

FIG. 7 is a schematic diagram of another embodiment of a data processing apparatus according to the present disclosure;

FIG. 8 is a schematic diagram of another embodiment of a data processing apparatus according to the present disclosure;

Fig. 9 is a schematic diagram of a data processing apparatus of the present specification.

Detailed Description

The embodiment of the specification provides a data processing method, a device, a system and equipment.

In order to make the technical solutions in the present specification better understood by those skilled in the art, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only some embodiments of the present specification, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are intended to be within the scope of the present disclosure.

Example 1

As shown in fig. 1A and 1B, an embodiment of the present disclosure provides a data processing method, where an execution body of the method may be a client, and the client may be a server or a terminal device, where the server may be an independent server or a server cluster formed by a plurality of servers, and the terminal device may be a device such as a personal computer or a mobile terminal device such as a mobile phone or a tablet computer. The method specifically comprises the following steps:

In S102, a risk identification result of the server for the first user is received.

The risk recognition result of the first user may be obtained by performing risk recognition on first feature data of the first user based on a first risk recognition model trained in advance by the service end, the first user may be a historical user using any service provided by the client, the first feature data of the first user may be a model stored by the service end and including attribute information (such as gender, age, region, etc.) of the first user, behavior information (such as transfer behavior, call behavior, etc.) of the first user, for example, the first user may be any one or more users transferring money using a resource transfer service provided by the client in the last month, the first risk recognition model trained in advance may be obtained by the service end based on a preset first number of user feature data (including attribute information, behavior information, etc. feature data of the user), the first risk recognition model may be a model constructed based on a preset deep learning algorithm and used for performing risk recognition, the preset first number may be greater than a preset sample number threshold, for example, the first risk recognition model trained in advance may be a service end may be 1000 pieces of user feature data, and the risk recognition network (Deep Neural Networks) is obtained by constructing a deep neural network recognition algorithm.

In implementation, with the rapid development of computer technology, the user quantity of enterprises, such as online banks, for providing resource transfer services for users increases sharply, and the data risk faced by the enterprises is also increasing, so that in order to reduce the data risk, the enterprises can perform risk identification processing by constructing a risk identification model. For example, an enterprise may train a risk identification model based on locally stored user sample data and risk tags, and perform risk identification processing based on the trained risk identification model. However, enterprises such as online banks may have situations of small data amount of user sample data and risk labels and poor data quality, which may result in poor risk recognition effect and low accuracy of risk recognition of the risk recognition model obtained by training, and therefore, a solution capable of improving the risk recognition effect and accuracy of risk recognition of the risk recognition model is needed. For this reason, the embodiments of the present specification provide a technical solution that can solve the above-mentioned problems, and specifically, reference may be made to the following.

Taking a client as an example of a client used by a mechanism capable of providing a resource transfer service for a user, because the private data of the user owned by different mechanisms are different, and the private data of the user owned by each mechanism may have the problems of smaller data volume and poorer data quality, in order to enable a locally used risk identification model (namely a second risk identification model) to achieve a better risk identification effect, data interaction can be performed through a server with larger data volume and stronger data processing capability, so that model parameters of the locally used risk identification model are updated.

The client may construct a first sample data set for training the second risk identification model, where the first sample data set may include user identifiers of a plurality of users and corresponding feature data, the client may send the user identifiers in the first sample data set to the server, the server may determine, according to the user identifiers, whether the server has a user corresponding to the user identifiers and the feature data, and if the user corresponding to the user identifiers and the feature data exist, determine the user as a first user, determine the feature data of the user as first feature data of the first user, and then the server may input the first feature data of the first user into the first risk identification model trained in advance, obtain a risk identification result for the first user, and the server may return the risk identification result for the first user to the client.

The client may update and filter the first sample data set based on the received risk identification result of the first user, so as to obtain a target sample data set composed of the second feature data of the first user and the risk identification result of the first user.

For example, the client may construct a first sample data set based on the user data acquired in the preset model training period, for example, the client may construct a first sample data set based on the user data of the resource transfer service in the last half month, specifically, the constructed first sample data set may include the feature data of the user 1 and the user identification of the user 1, the feature data of the user 2 and the user identification of the user 2, the client may send the user identifications of the user 1 and the user 2 to the server, if the server only stores the feature data of the user 2, the server may determine the user 2 as the first user, and the server may determine a risk identification result for the user 2 based on the stored first feature data of the user 2 and the pre-trained first risk identification model, and return the risk identification result of the user 2 to the client. The client may determine the feature data of the user 2 in the first sample data set as the second feature data of the user 2, and construct the target sample data set according to the second feature data of the user 2 and the risk identification result of the user 2 determined by the server.

In S104, based on the second feature data of the first user and the risk identification result of the first user, performing an initialization training on the preset second risk identification model to obtain an initialized second risk identification model.

The second risk recognition model may be a model for risk recognition constructed based on a preset deep learning algorithm, and a model structure of the second risk recognition model may be the same as or different from that of the first risk recognition model, for example, the first risk recognition model may be a model constructed based on a decision tree algorithm, the second risk recognition model may be a model constructed based on a decision tree algorithm, or the second risk recognition model may be a model constructed based on a neural network algorithm.

In implementation, data in the target sample data set may be input into a preset second risk identification model for initialization training, so as to obtain a second risk identification model after the initialization training.

Therefore, the risk recognition result of the first user is obtained based on the first risk recognition model of the server, and the first risk recognition model of the server is obtained by training based on the preset first number of user characteristic data, so that the risk recognition effect of the first risk recognition model of the first user is good, the second risk recognition model of the first user is initialized and trained based on the risk recognition result of the first user, and the risk recognition effect of the second risk recognition model can be improved under the assistance of the first risk recognition model. In addition, in the training process, the private data of the server can also realize the purpose that the data cannot go out of the domain, and the data security of the private data of the server can be protected.

In S106, based on the feature data of the second user and the risk recognition result of the second user, performing parameter updating processing on the initialized trained second risk recognition model to obtain a pre-trained second risk recognition model, so as to perform risk recognition processing on the user based on the pre-trained second risk recognition model.

The second user may include the first user, and the risk identification result of the second user may be a risk identification result determined according to the feature data of the second user by a manual method or the like.

In implementation, parameters of the second risk recognition model after the initialization training can be finely adjusted based on the feature data of the second user and the risk recognition result of the second user, so that the obtained risk recognition effect of the pre-trained second risk recognition model is more consistent with the risk recognition requirement of the client.

As described in S102 above, the client may construct a first sample data set, and determine a target sample data set including the second feature data of the first user and the risk identification result of the first user according to the risk identification result of the first user and the first sample data set returned by the server. In addition, the client may construct a second sample data set including feature data of the second user and risk recognition results of the second user according to the first sample data set, for example, the client may obtain risk recognition results corresponding to the feature data of the user in the first sample data set (the risk recognition results may be risk recognition results determined by the client based on a manual manner or the like), and construct the second sample data set based on the feature data of the user in the first sample data set and the determined risk recognition results, where the user in the second sample data set is the second user.

And inputting the feature data of the second user and the risk identification result of the second user into the initialized and trained second risk identification model for parameter updating processing, so that the pre-trained second risk identification model can be obtained.

In addition, in the case that the model structures of the first risk recognition model and the second risk recognition model are different, the second risk recognition model may learn the risk recognition capability of the first risk recognition model with the aid of the risk recognition result of the first user. The second risk recognition model only depends on publicly interactable information (namely a risk recognition result of the first user) in the initialization training process, and can be trained and deployed under the condition of completely isolating data without introducing and depending on other additional data in the parameter updating processing process and the subsequent risk recognition process, and also does not depend on a complex underlying data interaction frame, so that the risk recognition effect of the second risk recognition model is improved on the basis of protecting the safety of private data.

The embodiment of the specification provides a data processing method, which is used for receiving a risk identification result of a first user by a server, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server based on a pre-trained first risk identification model, the pre-trained first risk identification model is obtained by training the server based on preset first number of user feature data, performing initialization training on a preset second risk identification model based on second feature data of the first user and the risk identification result of the first user, obtaining an initialized trained second risk identification model, performing parameter updating processing on the initialized second risk identification model based on the feature data of the second user and the risk identification result of the second user, and obtaining a pre-trained second risk identification model so as to perform risk identification processing on the user based on the pre-trained second risk identification model. Therefore, the server only returns the risk identification result of the first user to the client, the first characteristic data of the first user is not required to be sent to the client, the private data of the server can be prevented from being out of the domain, the safety of the private data is guaranteed, meanwhile, the client can learn the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user, and the parameter update is carried out on the initialized second risk identification model based on the locally stored characteristic data and the risk identification result of the second user, so that the pre-trained second risk identification model can better meet the local risk identification requirement of the client, and the risk identification effect of the pre-trained second risk identification model is improved.

Example two

As shown in fig. 2, the embodiment of the present disclosure provides a data processing method, where an execution body of the method may be a client, and the client may be a server or a terminal device, where the server may be an independent server, or may be a server cluster formed by a plurality of servers, and the terminal device may be a device such as a personal computer, or may be a mobile terminal device such as a mobile phone, a tablet computer, or the like. The method specifically comprises the following steps:

The risk identification result of the first user may be obtained by performing risk identification on first feature data of the first user based on a first risk identification model trained in advance by the service end, and the first risk identification model trained in advance may be obtained by performing training on a first number of user feature data preset by the service end.

In S202, an initialization training is performed on a preset second risk recognition model based on the second feature data of the first user, the risk recognition result of the first user, and the first loss function, so as to obtain an initialized second risk recognition model.

The risk identification result of the first user may include a risk score, and the first loss function may be a mean square error loss function.

In S204, based on the feature data of the second user, the risk classification label of the second user, and the second loss function, parameter updating processing is performed on the initialized and trained second risk recognition model, so as to obtain a pre-trained second risk recognition model.

Wherein the risk identification result of the second user may comprise a risk classification tag and the second loss function may be a cross entropy loss function.

In implementation, since the data size and the data processing capability of the server may be better than those of the client, the first risk identification model constructed by the server may be different from the model structure of the second risk identification model, and the obtained risk identification structure may be different, for example, the risk identification result of the first user obtained by the first risk identification model constructed by the server may be a more complex continuous risk score (for example, the risk score may be any value of 0-1), and the risk identification result of the second user determined locally by the client may be a simpler two-class risk classification tag (for example, a 0-1 tag, i.e., a tag of 0 may represent a risk-free user and a tag of 1 may represent a risk user), so different loss functions may be set for the initialization training process and the parameter updating process.

For example, as shown in fig. 3, the second risk recognition model may include a feature extraction layer and a full connection layer, taking the second risk recognition model as an example of a model constructed based on a convolutional neural network (Convolutional Neural Networks, CNN), the second risk recognition model may include a convolutional layer, a pooling layer, and a full connection layer, where the convolutional layer and the pooling layer are the feature extraction layers.

The client can input second feature data of the first user into the second risk recognition model to obtain a predicted risk score of the first user, and perform initialization training on the second risk recognition model according to the risk score (namely, the risk recognition result of the first user) of the first user determined by the server based on the first feature data of the first user and the first risk recognition model and a mean square error loss function to obtain a second risk model after the initialization training.

Inputting the feature data of the second user into the initialized and trained second risk model to obtain a predicted risk type of the second user, and carrying out parameter updating processing on the initialized and trained second risk recognition model based on the risk classification labels and the cross entropy loss functions of the second user to obtain a pre-trained second risk recognition model.

In addition, in order to maintain the stability of the second risk model, the adjustment amplitude of the parameters of the feature extraction layer and the full connection layer of the second risk model can be determined according to the different sample sizes, and the specific processing procedure can be seen from the following steps one to two:

step one, determining a first parameter updating amplitude aiming at a characteristic extraction layer and a second parameter updating amplitude aiming at a full connection layer based on the data quantity of a second user and a preset quantity threshold value.

Wherein the first parameter update amplitude is smaller than the second parameter update amplitude.

In implementation, for example, in the case that the number of the second users is smaller than the preset number threshold, parameters of the feature extraction layer of the second risk model after the initialization training may be maintained not to be updated, and only parameters of the full-connection layer of the second risk identification model after the initialization training are subjected to parameter updating processing; under the condition that the number of the second users is not smaller than a preset number threshold, the feature extraction layer of the second risk model after the initialization training can be controlled through the regular loss function to update based on the first parameter updating amplitude, and the parameters of the full-connection layer of the second risk identification model after the initialization training are subjected to normal parameter updating processing based on the first parameter updating amplitude, wherein in order to maintain the stability of the second risk identification model, the first parameter updating amplitude can be smaller than the second parameter updating amplitude.

And secondly, carrying out parameter updating processing on the initialized and trained second risk recognition model based on the first parameter updating amplitude, the second parameter updating amplitude, the characteristic data of the second user, the risk classification label of the second user and the second loss function to obtain a pre-trained second risk recognition model.

In the implementation, assuming that the preset number threshold is 100, under the condition that the number of the second users is less than 100, parameters of a feature extraction layer of the second risk recognition model after initialization training can be maintained not to be updated, and parameters of the full-connection layer are updated only based on the second parameter updating amplitude, so that a pre-trained second risk recognition model is obtained; under the condition that the number of the second users is not less than 100 (i.e. the training sample data size of the client is larger), the parameters of the feature extraction layer of the second risk identification model after the initialization training can be finely adjusted based on the first parameter updating amplitude, and the parameters of the full-connection layer are updated based on the second parameter updating amplitude, so that the pre-trained second risk identification model is obtained.

In addition, the method for determining the first parameter updating magnitude and the second parameter updating magnitude may be various, and may be different according to the actual application scenario, which is not specifically limited in the embodiment of the present disclosure,

In S206, based on the preset data processing period, a detection process is performed on whether the second risk recognition model trained in advance meets the preset risk recognition requirement.

In implementation, for example, whether the pre-trained second risk recognition model meets the preset risk recognition requirement may be detected every half month, and specific detection processing methods may be various, for example, the accuracy of risk recognition based on the pre-trained second risk recognition model may be acquired in the near half month, if the accuracy is smaller than the preset accuracy threshold, the pre-trained second risk recognition model may be considered to not meet the preset risk recognition requirement, or in a preset data processing period, whether the pre-trained second risk recognition model meets the preset risk recognition requirement may be determined according to the change situation of the usage scenario of the pre-trained second risk recognition model, for example, when the usage scenario of the pre-trained second risk recognition model is converted from the resource transfer scenario to the identity recognition scenario, the pre-trained second risk recognition model may be considered to not meet the preset risk recognition requirement.

The method for detecting whether the pre-trained second risk recognition model meets the preset risk recognition requirement is an optional and realizable detection method, and in an actual application scene, there may be a plurality of different detection methods, and the detection methods may be different according to the actual application scene, which is not particularly limited in the embodiment of the present disclosure.

In S208, in the case where it is detected that the pre-trained second risk recognition model does not meet the preset risk recognition requirement, updating the pre-trained second risk recognition model based on the feature data of the third user.

In practice, the above-mentioned processing manner of S208 may be varied, and the following provides an alternative implementation manner, which can be seen from the following steps one to four:

Step one, inputting the characteristic data of the third user into a pre-trained second risk identification model to obtain a risk identification result of the third user.

And secondly, determining the second risk recognition model after the model structure is changed as a third risk recognition model under the condition that the model structure of the pre-trained second risk recognition model is changed.

Thirdly, initializing and training the third risk identification model based on the feature data of the third user and the risk identification result of the third user to obtain a third risk identification model after initializing and training.

In implementation, since the update speed of the fraud means of the malicious third party is faster, in order to improve the accuracy of risk identification, the model structure of the second risk identification model may be updated (for example, the pre-trained second risk model may be a risk identification model constructed based on a neural network algorithm, and the second risk identification model after the model structure change may be a risk identification model constructed based on a decision tree), so that, in order to enable the second risk identification model after the model structure change to retain the risk identification capability of the second risk identification model before the model structure change, the third risk identification model (the second risk identification model after the model structure change) may be subjected to initialization training based on the feature data of the third user and the risk identification result of the third user, so as to obtain the third risk identification model after the initialization training, and thus, the third risk identification model after the initialization training may learn the risk identification capability of the pre-trained second risk identification model.

And step four, carrying out parameter updating processing on the initialized trained third risk recognition model based on the characteristic data of the fourth user and the risk recognition result of the fourth user to obtain a pre-trained third risk recognition model, and determining the pre-trained third risk recognition model as a pre-trained second risk recognition model.

In implementation, the process of performing the parameter updating process on the third risk identification model after the initialization training may refer to the process of performing the parameter updating process on the second risk identification model after the initialization training, which is not described herein.

Example III

As shown in fig. 4A and fig. 4B, the embodiment of the present disclosure provides a data processing method, where an execution body of the method may be a client, and the client may be a server or a terminal device, where the server may be an independent server or a server cluster formed by a plurality of servers, and the terminal device may be a device such as a personal computer or a mobile terminal device such as a mobile phone or a tablet computer. The method specifically comprises the following steps:

in S402, in a case where it is detected that the target user triggers execution of the target service, first feature data of the target user is acquired.

The target service may be any service that the client can provide for the user, for example, the target service may be a resource transfer service, an authentication service, etc., and the first feature data of the target user may include attribute information (such as gender, age, region, etc.) of the first user, behavior information (such as transfer behavior, call behavior, etc.), etc.

In an implementation, taking a client as an example of a client used by a mechanism capable of providing a resource transfer service for a user, a target service may be a resource transfer service, a target user may trigger to start a resource transfer application installed in the client and trigger to start the resource transfer service (i.e., the target service), and the client may acquire attribute information and resource transfer behavior information (such as a resource transfer number, a resource transfer time, a resource transfer object, etc.) of the target user when detecting that the target user triggers to execute the resource transfer service.

In S404, the first feature data of the target user is input into a pre-trained second risk recognition model, so as to obtain a target risk recognition result of the target user.

The pre-trained second risk recognition model may be obtained by training based on second feature data of the first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user may be obtained by performing risk recognition on the first feature data of the first user based on the pre-trained first risk recognition model by the service end, and the pre-trained first risk recognition model may be obtained by training based on a preset first number of user feature data by the service end.

In implementation, attribute information of the target user, resource transfer behavior information (such as the number of resource transfers, resource transfer time, resource transfer objects, etc.) and other first feature data can be input into a pre-trained second risk identification model to obtain a target risk identification result of the target user.

In S406, it is determined whether there is a risk of triggering execution of the target service based on the target risk identification result.

In implementation, if it is determined that the target service is triggered to be executed and has risk based on the target risk identification result, preset alarm information can be output to the target user, and execution of the target service is stopped.

The embodiment of the specification provides a data processing method, under the condition that a target user is detected to trigger execution of target service, first feature data of the target user is obtained, the first feature data of the target user is input into a pre-trained second risk recognition model, a target risk recognition result of the target user is obtained, the pre-trained second risk recognition model is obtained by training based on the second feature data of the first user, the risk recognition result of the first user, the feature data of the second user and the risk recognition result of the second user, the risk recognition result of the first user is obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, the pre-trained first risk recognition model is obtained by training a service end based on a preset first number of user feature data, and whether the trigger execution of the target service is at risk is determined based on the target risk recognition result. When the client trains the second risk model, the client only uses the first risk recognition result provided by the server except the second feature data of the first user, the feature data of the second user and the risk recognition result of the second user, so that the private data of the server can not go out of the domain, the safety of the private data is ensured, and therefore, the client can learn the risk recognition capability of the first risk recognition model under the assistance of the first risk recognition result, and the obtained pre-trained second risk recognition model accords with the local risk recognition requirement of the client more according to the feature data of the second user and the risk recognition result of the second user, thereby improving the risk recognition effect of the pre-trained second risk recognition model and the accuracy of triggering the target service to the target user.

Example IV

As shown in fig. 5, the embodiment of the present disclosure provides a data processing method, where an execution body of the method may be a client, and the client may be a server or a terminal device, where the server may be an independent server, or may be a server cluster formed by a plurality of servers, and the terminal device may be a device such as a personal computer, or may be a mobile terminal device such as a mobile phone, a tablet computer, or the like. The method specifically comprises the following steps:

In S502, the user identifier of the target user is sent to the server, and a first risk identification result of the target user returned by the server is received.

The first risk identification result of the target user may be obtained by performing risk identification on the second feature data of the target user by the service end based on a first risk identification model trained in advance.

In S504, it is determined whether there is a risk of triggering execution of the target service based on the first risk identification result and the target risk identification result.

In implementation, since the data size and the data processing capability of the server may be better than those of the client, the client may also send the user identifier of the target user to the server, and receive the first risk identification result of the target user returned by the server.

In addition, the model result output by the first risk identification model constructed by the server may be different from the model result of the second risk identification model, for example, the first risk identification result of the first risk identification model constructed by the server may be a relatively complex continuous risk score (for example, the risk score may be any value of 0-1), while the target risk identification result locally determined by the client may be a relatively simple two-class risk classification tag (for example, a0, 1 tag, i.e., a tag of 0 may represent a risk-free user and a tag of 1 may represent a risk user), so that, in combination with the risk score of the target user and the risk classification tag, it may be determined whether the risk exists in triggering the execution of the target service. The specific risk determination method may be selected according to different actual application scenarios, which is not specifically limited in the embodiment of the present disclosure.

Example five

Embodiments of the present specification provide a data processing system, comprising: server side and customer side, wherein:

the client can be used for sending the user identification of the first user to the server.

The server side can be used for acquiring a risk identification result of the first user based on the user identification, the risk identification result of the first user can be obtained by performing risk identification on first feature data of the first user based on a pre-trained first risk identification model, and the pre-trained first risk identification model can be obtained by performing training on the first feature data of the first user based on the pre-set first number of user by the server side.

The client can be used for carrying out initialization training on a preset second risk recognition model based on second characteristic data of the first user and a risk recognition result of the first user returned by the server to obtain an initialized trained second risk recognition model, carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model, and carrying out risk recognition processing on the user based on the pre-trained second risk recognition model.

For example, as shown in fig. 6, there may be multiple servers and clients in the data processing system, taking risk user identification in a resource transfer scenario as an example, assuming that the clients and the multiple servers respectively store feature data (such as resource transfer behavior information) of multiple users, the clients may respectively send user identifiers of multiple first users stored locally to the multiple servers, and the servers may acquire the first feature data of the first users stored locally according to the received user identifiers of the first users, determine risk identification results of the first users according to a first risk identification model trained in advance by the servers, and return the risk identification results to the clients.

The client performs initialization training and parameter updating processing on the preset second risk identification model based on the second feature data of the first user, the risk identification result of the first user returned by the server, the feature data of the second user and the risk identification result of the second user (the specific processing procedure can be referred to the processing procedure in the first embodiment to the second embodiment) to obtain a pre-trained second risk identification model.

In addition, there may be multiple servers in the data processing system, the client may receive risk identification results of multiple servers for the same first user, for example, as shown in fig. 6, the client may receive risk identification results of the first user 1 determined by the server 1 based on the locally stored first feature data 1 of the first user and the first risk identification model 1 trained in advance, the client may also receive risk identification results of the first user 1 determined by the server 2 based on the locally stored first feature data 2 of the first user and the first risk identification model 2 trained in advance, and the client may determine risk identification results of the first user 1 based on the risk identification results of the first user 1 returned by the server 1 and the risk identification results of the first user 1 returned by the server 2.

For example, the client may determine the risk recognition result of the first user according to the preset weight value corresponding to the server and the risk recognition result of the first user returned by the server, and specifically, for example, assuming that the preset weight of the server 1 is 0.8, the risk recognition result of the first user 1 returned by the server 1 is a risk score of 0.8, the preset weight of the server 2 is 0.7, the risk recognition result of the first user 1 returned by the server 1 is a risk score of 0.75, the risk recognition result of the first user may be (0.8×0.8+0.7×0.75)/2=0.58.

The method for determining the risk identification result of the first user is an optional and implementable method, and a plurality of different determining methods may be available in the actual application scenario, for example, the risk identification result of the first user may be determined according to the wind control priority of the server, and the different determining methods may be selected according to the different actual application scenarios, which is not particularly limited in the embodiment of the present disclosure.

In addition, the client can acquire the first characteristic data of the target user under the condition that the target user is detected to trigger the execution of the target service, and input the first characteristic data of the target user into a pre-trained second risk identification model to obtain a target risk identification result of the target user.

The client may send the user identifier of the target user to the server, and receive the first risk identification result of the target user returned by the server. The first risk identification result of the target user may be obtained by performing risk identification on the second feature data of the target user by the service end based on a first risk identification model trained in advance.

Finally, the client may determine whether there is a risk of triggering to execute the target service based on the first risk identification result and the target risk identification result.

Therefore, the second risk recognition model can migrate the risk recognition result of the server side through the distillation learning mode, the risk recognition energy mechanics of the first risk recognition model of the server side can be applied to the client, the problem of privacy data interaction in practical application is avoided, in addition, the client can conduct real-time risk recognition through the pre-trained second risk recognition model, and prevention and control loopholes existing in the mode of conducting risk recognition through data interaction (such as conducting risk recognition through blacklist interaction) are avoided.

The embodiment of the specification provides a data processing system, because the server only returns the risk identification result of the first user to the client, the private data of the server can be realized without sending the first characteristic data of the first user to the client, the security of the private data is ensured, meanwhile, the client can learn the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user, and the parameter update is performed on the initialized second risk identification model based on the locally stored characteristic data and the risk identification result of the second user, so that the pre-trained second risk identification model can better meet the local risk identification requirement of the client, and the risk identification effect of the pre-trained second risk identification model is improved.

Example six

The data processing method provided in the embodiment of the present disclosure is based on the same concept, and the embodiment of the present disclosure further provides a data processing device, as shown in fig. 7.

The data processing apparatus includes: a result receiving module 701, a first training module 702 and a second training module 703, wherein:

the result receiving module 701 is configured to receive a risk identification result of a server for a first user, where the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user based on a first risk identification model that is trained in advance by the server, and the first risk identification model is obtained by training the server based on a preset first number of user feature data;

The first training module 702 is configured to perform an initialization training on a preset second risk recognition model based on the second feature data of the first user and the risk recognition result of the first user, so as to obtain an initialized second risk recognition model;

And the second training module 703 is configured to perform parameter updating processing on the initialized and trained second risk recognition model based on the feature data of the second user and the risk recognition result of the second user, so as to obtain a pre-trained second risk recognition model, so as to perform risk recognition processing on the user based on the pre-trained second risk recognition model.

In this embodiment of the present disclosure, the risk identification result of the first user includes a risk score, and the first training module 702 is configured to:

performing initialization training on the preset second risk recognition model based on the second characteristic data of the first user, the risk recognition result of the first user and a first loss function to obtain a second risk recognition model after the initialization training;

the risk identification result of the second user includes a risk classification label, and the second training module 1103 is configured to:

Based on the feature data of the second user, the risk classification label of the second user and a second loss function, carrying out parameter updating processing on the initialized and trained second risk recognition model to obtain the pre-trained second risk recognition model;

The first loss function is a mean square error loss function and the second loss function is a cross entropy loss function.

In this embodiment of the present disclosure, the second risk recognition model includes a feature extraction layer and a full connection layer, and the second training module 703 is configured to:

Determining a first parameter update amplitude for the feature extraction layer and a second parameter update amplitude for the fully connected layer based on the data volume of the second user and a preset number threshold, the first parameter update amplitude being smaller than the second parameter update amplitude;

and carrying out parameter updating processing on the initialized and trained second risk recognition model based on the first parameter updating amplitude, the second parameter updating amplitude, the characteristic data of the second user, the risk classification label of the second user and a second loss function to obtain the pre-trained second risk recognition model.

In an embodiment of the present disclosure, the apparatus further includes:

the model detection module is used for detecting whether the pre-trained second risk identification model meets the preset risk identification requirement or not based on a preset data processing period;

And the model updating module is used for updating the pre-trained second risk recognition model based on the characteristic data of the third user under the condition that the pre-trained second risk recognition model is detected to not meet the preset risk recognition requirement.

In an embodiment of the present disclosure, the model update module is configured to:

inputting the characteristic data of the third user into the pre-trained second risk identification model to obtain a risk identification result of the third user;

under the condition that the model structure of the pre-trained second risk identification model changes, determining the second risk identification model after the model structure changes as a third risk identification model;

Performing initialization training on the third risk recognition model based on the characteristic data of the third user and the risk recognition result of the third user to obtain a third risk recognition model after the initialization training;

And carrying out parameter updating processing on the initialized and trained third risk recognition model based on the characteristic data of the fourth user and the risk recognition result of the fourth user to obtain a pre-trained third risk recognition model, and determining the pre-trained third risk recognition model as the pre-trained second risk recognition model.

The embodiment of the specification provides a data processing device, which is used for receiving a risk identification result of a first user by a server, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server based on a pre-trained first risk identification model, the pre-trained first risk identification model is obtained by training the server based on preset first number of user feature data, performing initialization training on a preset second risk identification model based on second feature data of the first user and the risk identification result of the first user, obtaining an initialized trained second risk identification model, performing parameter updating processing on the initialized second risk identification model based on the feature data of the second user and the risk identification result of the second user, and obtaining a pre-trained second risk identification model so as to perform risk identification processing on the user based on the pre-trained second risk identification model. Therefore, the server only returns the risk identification result of the first user to the client, the first characteristic data of the first user is not required to be sent to the client, the private data of the server can be prevented from being out of the domain, the safety of the private data is guaranteed, meanwhile, the client can learn the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user, and the parameter update is carried out on the initialized second risk identification model based on the locally stored characteristic data and the risk identification result of the second user, so that the pre-trained second risk identification model can better meet the local risk identification requirement of the client, and the risk identification effect of the pre-trained second risk identification model is improved.

Example seven

Based on the same concept, the embodiment of the present disclosure further provides a data processing apparatus, as shown in fig. 8.

The data processing apparatus includes: a data acquisition module 801, a result acquisition module 802, and a risk determination module 803, wherein:

A data obtaining module 801, configured to obtain first feature data of a target user when it is detected that the target user triggers execution of a target service;

The result obtaining module 802 is configured to input the first feature data of the target user into a pre-trained second risk recognition model, to obtain a target risk recognition result of the target user, where the pre-trained second risk recognition model is obtained by training based on the second feature data of the first user, the risk recognition result of the first user, the feature data of the second user, and the risk recognition result of the second user, the risk recognition result of the first user is obtained by performing risk recognition on the first feature data of the first user by using a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training the service end based on a preset first number of user feature data;

and the risk determining module 803 is configured to determine whether there is a risk in triggering execution of the target service based on the target risk identification result.

In the embodiment of the present specification, the risk determining module 803 is configured to:

The user identification of the target user is sent to the server, and a first risk identification result of the target user returned by the server is received, wherein the first risk identification result of the target user is obtained by the server performing risk identification on second characteristic data of the target user based on the pre-trained first risk identification model;

And determining whether the target service is triggered to be executed or not to be at risk based on the first risk identification result and the target risk identification result.

The embodiment of the specification provides a data processing device, under the condition that a target user is detected to trigger execution of target service, first feature data of the target user is obtained, the first feature data of the target user is input into a pre-trained second risk recognition model, a target risk recognition result of the target user is obtained, the pre-trained second risk recognition model is obtained by training based on the second feature data of the first user, the risk recognition result of the first user, the feature data of the second user and the risk recognition result of the second user, the risk recognition result of the first user is obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, the pre-trained first risk recognition model is obtained by training a service end based on a preset first number of user feature data, and whether the trigger execution of the target service is at risk is determined based on the target risk recognition result. When the client trains the second risk model, the client only uses the first risk recognition result provided by the server except the second feature data of the first user, the feature data of the second user and the risk recognition result of the second user, so that the private data of the server can not go out of the domain, the safety of the private data is ensured, and therefore, the client can learn the risk recognition capability of the first risk recognition model under the assistance of the first risk recognition result, and the obtained pre-trained second risk recognition model accords with the local risk recognition requirement of the client more according to the feature data of the second user and the risk recognition result of the second user, thereby improving the risk recognition effect of the pre-trained second risk recognition model and the accuracy of triggering the target service to the target user.

Example eight

Based on the same idea, the embodiment of the present disclosure further provides a data processing apparatus, as shown in fig. 9.

The data processing apparatus may vary widely in configuration or performance, may include one or more processors 901 and memory 902, and may store one or more storage applications or data in memory 902. Wherein the memory 902 may be transient storage or persistent storage. The application programs stored in the memory 902 may include one or more modules (not shown) each of which may include a series of computer executable instructions for use in a data processing apparatus. Still further, the processor 901 may be arranged to communicate with a memory 902 and execute a series of computer executable instructions in the memory 902 on a data processing device. The data processing device may also include one or more power supplies 903, one or more wired or wireless network interfaces 904, one or more input output interfaces 905, and one or more keyboards 906.

In particular, in this embodiment, the data processing apparatus includes a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the data processing apparatus, and the one or more programs configured to be executed by the one or more processors comprise instructions for:

Receiving a risk identification result of a server side aiming at a first user, wherein the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by the server side based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user feature data;

Performing initialization training on a preset second risk identification model based on second characteristic data of the first user and a risk identification result of the first user to obtain a second risk identification model after the initialization training;

And carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model, so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

Optionally, the risk recognition result of the first user includes a risk score, and the initializing the preset second risk recognition model based on the feature data of the first user and the risk recognition result of the first user to obtain an initialized second risk recognition model includes:

the risk identification result of the second user includes a risk classification label, and the parameter updating process is performed on the initialized and trained second risk identification model based on the feature data of the second user and the risk identification result of the second user to obtain a pre-trained second risk identification model, which includes:

Optionally, the second risk recognition model includes a feature extraction layer and a full connection layer, and the parameter updating process is performed on the initialized and trained second risk recognition model based on the feature data of the second user, the risk classification label of the second user, and a second loss function, so as to obtain the pre-trained second risk recognition model, which includes:

Optionally, the method further comprises:

based on a preset data processing period, detecting whether the pre-trained second risk identification model meets a preset risk identification requirement or not;

And under the condition that the pre-trained second risk identification model does not meet the preset risk identification requirement, updating the pre-trained second risk identification model based on the characteristic data of a third user.

Optionally, the updating the pre-trained second risk recognition model based on the feature data of the third user and the risk recognition result of the third user includes:

In addition, the data processing apparatus includes a memory, and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer-executable instructions for the data processing apparatus, and configured to be executed by the one or more processors, the one or more programs comprising computer-executable instructions for:

Under the condition that a target user is detected to trigger execution of a target service, acquiring first characteristic data of the target user;

Inputting first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of the first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user are obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, and the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data;

And determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

Optionally, the determining, based on the risk identification result of the target user, whether there is a risk in triggering to execute the target service includes:

The embodiment of the specification provides data processing equipment, because the server only returns the risk identification result of the first user to the client, the private data of the server can be realized without sending the first characteristic data of the first user to the client, the security of the private data is ensured, meanwhile, the client can learn the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user, and the parameter update is carried out on the initialized second risk identification model based on the locally stored characteristic data and the risk identification result of the second user, so that the pre-trained second risk identification model can better meet the local risk identification requirement of the client, and the risk identification effect of the pre-trained second risk identification model is improved.

Example nine

The embodiments of the present disclosure further provide a computer readable storage medium, where a computer program is stored, where the computer program when executed by a processor implements each process of the embodiments of the data processing method, and the same technical effects can be achieved, and for avoiding repetition, a detailed description is omitted herein. The computer readable storage medium is, for example, a Read-Only Memory (ROM), a random access Memory (Random Access Memory RAM), a magnetic disk or an optical disk.

The embodiment of the specification provides a computer readable storage medium, because the server only returns the risk identification result of the first user to the client, the private data of the server can be realized without sending the first characteristic data of the first user to the client, so that the security of the private data is ensured, meanwhile, the client can learn the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user, and update parameters of the initialized second risk identification model based on the locally stored characteristic data and the risk identification result of the second user, so that the pre-trained second risk identification model can better meet the local risk identification requirement of the client, and the risk identification effect of the pre-trained second risk identification model is improved.

The foregoing describes specific embodiments of the present disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims can be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

In the 90 s of the 20 th century, improvements to one technology could clearly be distinguished as improvements in hardware (e.g., improvements to circuit structures such as diodes, transistors, switches, etc.) or software (improvements to the process flow). However, with the development of technology, many improvements of the current method flows can be regarded as direct improvements of hardware circuit structures. Designers almost always obtain corresponding hardware circuit structures by programming improved method flows into hardware circuits. Therefore, an improvement of a method flow cannot be said to be realized by a hardware entity module. For example, a programmable logic device (Programmable Logic Device, PLD) (e.g., field programmable gate array (Field Programmable GATE ARRAY, FPGA)) is an integrated circuit whose logic functions are determined by user programming of the device. A designer programs to "integrate" a digital system onto a PLD without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Moreover, nowadays, instead of manually manufacturing integrated circuit chips, such programming is mostly implemented with "logic compiler (logic compiler)" software, which is similar to the software compiler used in program development and writing, and the original code before being compiled is also written in a specific programming language, which is called hardware description language (Hardware Description Language, HDL), but HDL is not just one, but a plurality of kinds, such as ABEL(Advanced Boolean Expression Language)、AHDL(Altera Hardware Description Language)、Confluence、CUPL(Cornell University Programming Language)、HDCal、JHDL(Java Hardware Description Language)、Lava、Lola、MyHDL、PALASM、RHDL(Ruby Hardware Description Language), and VHDL (Very-High-SPEED INTEGRATED Circuit Hardware Description Language) and Verilog are currently most commonly used. It will also be apparent to those skilled in the art that a hardware circuit implementing the logic method flow can be readily obtained by merely slightly programming the method flow into an integrated circuit using several of the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer readable medium storing computer readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, application SPECIFIC INTEGRATED Circuits (ASICs), programmable logic controllers, and embedded microcontrollers, examples of controllers include, but are not limited to, the following microcontrollers: ARC625D, atmel AT91SAM, microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic of the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller in a pure computer readable program code, it is well possible to implement the same functionality by logically programming the method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers, etc. Such a controller may thus be regarded as a kind of hardware component, and means for performing various functions included therein may also be regarded as structures within the hardware component. Or even means for achieving the various functions may be regarded as either software modules implementing the methods or structures within hardware components.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. One typical implementation is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing one or more embodiments of the present description.

It will be appreciated by those skilled in the art that embodiments of the present description may be provided as a method, system, or computer program product. Accordingly, one or more embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, one or more embodiments of the present description can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

Embodiments of the present description are described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the specification. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

One or more embodiments of the present specification may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. One or more embodiments of the present description may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for system embodiments, since they are substantially similar to method embodiments, the description is relatively simple, as relevant to see a section of the description of method embodiments.

The foregoing is merely exemplary of the present disclosure and is not intended to limit the disclosure. Various modifications and alterations to this specification will become apparent to those skilled in the art. Any modifications, equivalent substitutions, improvements, or the like, which are within the spirit and principles of the present description, are intended to be included within the scope of the claims of the present description.

Claims

1. A data processing method, comprising:

Initializing a preset second risk recognition model based on second characteristic data of the first user and a risk recognition result of the first user to obtain an initialized second risk recognition model, wherein the second risk recognition model is a risk recognition model of a client, and the second risk recognition model is a model obtained by learning the risk recognition capability of the first risk recognition model under the assistance of the risk recognition result of the first user;

2. The method of claim 1, wherein the risk recognition result of the first user includes a risk score, and the initializing the preset second risk recognition model based on the second feature data of the first user and the risk recognition result of the first user to obtain an initialized second risk recognition model includes:

3. The method of claim 2, wherein the second risk recognition model includes a feature extraction layer and a full connection layer, and the performing parameter update processing on the initialized trained second risk recognition model based on feature data of the second user, a risk classification label of the second user, and a second loss function to obtain the pre-trained second risk recognition model includes:

4. The method of claim 1, the method further comprising:

5. The method according to claim 4, wherein the updating the pre-trained second risk recognition model based on the feature data of the third user and the risk recognition result of the third user includes:

6. A data processing method, comprising:

Inputting first feature data of the target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of a first user, risk recognition results of the first user, feature data of the second user and risk recognition results of the second user, the risk recognition results of the first user are obtained by training the first feature data of the first user based on a pre-trained first risk recognition model by a service end, the pre-trained first risk recognition model is obtained by training the service end based on a preset first number of user feature data, the second risk recognition model is a risk recognition model of a client, and the second risk recognition model is obtained by learning the risk recognition capability of the first risk recognition model with the aid of the risk recognition results of the first user;

7. The method of claim 6, the determining whether there is a risk to trigger execution of the target service based on the risk identification result of the target user, comprising:

8. A data processing system comprising a server and a client, wherein:

The client is used for sending the user identification of the first user to the server;

The server side is configured to obtain a risk identification result of the first user based on the user identifier, where the risk identification result of the first user is obtained by performing risk identification on first feature data of the first user by using the server side based on a first risk identification model that is trained in advance, and the first risk identification model that is trained by using the server side based on a preset first number of user feature data;

The client is configured to perform an initialization training on a preset second risk recognition model based on second feature data of the first user and a risk recognition result of the first user returned by the server, obtain an initialized trained second risk recognition model, perform parameter updating processing on the initialized second risk recognition model based on feature data of the second user and the risk recognition result of the second user, and obtain a pre-trained second risk recognition model, so as to perform risk recognition processing on the user based on the pre-trained second risk recognition model, where the second risk recognition model is a model obtained by learning a risk recognition capability of the first risk recognition model with the aid of the risk recognition result of the first user.

9. A data processing apparatus comprising:

The risk identification module is used for receiving a risk identification result of a first user aiming at a server side, wherein the risk identification result of the first user is obtained by performing risk identification on first characteristic data of the first user based on a pre-trained first risk identification model, and the pre-trained first risk identification model is obtained by training the server side based on a preset first number of user characteristic data;

the first training module is used for carrying out initialization training on a preset second risk identification model based on the second characteristic data of the first user and the risk identification result of the first user to obtain an initialized second risk identification model, wherein the second risk identification model is a risk identification model of a client, and the second risk identification model is a model obtained by learning the risk identification capability of the first risk identification model under the assistance of the risk identification result of the first user;

And the second training module is used for carrying out parameter updating processing on the initialized trained second risk recognition model based on the characteristic data of the second user and the risk recognition result of the second user to obtain a pre-trained second risk recognition model so as to carry out risk recognition processing on the user based on the pre-trained second risk recognition model.

10. A data processing apparatus comprising:

the data acquisition module is used for acquiring first characteristic data of the target user under the condition that the target user is detected to trigger the execution of the target service;

The method comprises the steps of inputting first feature data of a target user into a pre-trained second risk recognition model to obtain a target risk recognition result of the target user, wherein the pre-trained second risk recognition model is obtained by training based on second feature data of a first user, the risk recognition result of the first user, feature data of the second user and the risk recognition result of the second user, the risk recognition result of the first user is obtained by performing risk recognition on the first feature data of the first user by a service end based on the pre-trained first risk recognition model, the pre-trained first risk recognition model is obtained by training by the service end based on a preset first number of user feature data, the second risk recognition model is a risk recognition model of a client, and the second risk recognition model is obtained by learning the risk recognition capability of the first risk recognition model with the aid of the risk recognition result of the first user;

And the risk determining module is used for determining whether the risk exists in triggering execution of the target service based on the target risk identification result.

11. A data processing apparatus, the data processing apparatus comprising:

A processor; and

A memory arranged to store computer executable instructions that, when executed, cause the processor to:

12. A data processing apparatus, the data processing apparatus comprising:

A processor; and

13. A storage medium for storing computer-executable instructions that when executed implement the following:

14. A storage medium for storing computer-executable instructions that when executed implement the following: