A risk identification method, device and equipment

Abstract

The embodiments of this specification disclose a risk identification method, apparatus, and device, which relate to the field of computer technology. The method includes: obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario; performing encoding processing on the business operation data to obtain a data representation corresponding to the business operation data; inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the tasks. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario; based on the risk representation for each task, respectively determine the risk identification result for the target merchant through a pre-trained risk identification network.

Description

Technical Field

[0001] This document relates to the field of computer technology, and in particular, to a risk identification method, apparatus, and device. Background Art

[0002] With the development of informatization and digitalization, and people's increasing attention to their own privacy data, the scale and complexity of merchant risks are constantly rising. Usual risk decisions usually rely heavily on historical cases and expert opinions. Therefore, an automated risk decision-making system can be established, and the premise of establishing an automated risk decision-making system is to be able to better predict the risk occurrence probability of merchants. Since the behavioral manifestations of merchants in different business operation scenarios (such as merchants in offline business operation scenarios (or offline merchants) and merchants in online business operation scenarios (or online merchants), etc.) before the occurrence of risks are often quite different, therefore, how to predict the risk probability of merchants is an important topic that needs to be concerned about currently.

[0003] Generally, the entire life cycle of a merchant can be split into different cycle intervals, and then, corresponding models can be constructed for each cycle interval respectively. However, the above method will consume a large amount of resources for model construction and model maintenance. Therefore, it is necessary to provide a multi-stage merchant risk probability prediction mechanism based on the life cycle to improve the overall accuracy of risk probability prediction and the decision-making ability of the automated decision-making system. Summary of the Invention

[0004] The purpose of the embodiments of this specification is to provide a multi-stage merchant risk probability prediction mechanism based on the life cycle to improve the overall accuracy of risk probability prediction and the decision-making ability of the automated decision-making system.

[0005] To achieve the above technical solution, the embodiments of this specification are implemented as follows:

[0006] A risk identification method provided by the embodiments of this specification, the method includes: obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario. Performing encoding processing on the business operation data to obtain a data representation corresponding to the business operation data. Inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation of each task including the relationship between different tasks in multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include the risk identification service of the target merchant in the first business operation scenario and the risk identification service of the target merchant converted from the first business operation scenario to a second business operation scenario. Based on the risk representation of each task, respectively determining a risk identification result for the target merchant through a pre-trained risk identification network.

[0007] A risk identification device provided by an embodiment of this specification, the device includes: a data acquisition module, which acquires business operation data related to the business operation behavior of a target merchant in a first business operation scenario. An encoding module, which performs encoding processing on the business operation data to obtain a data representation corresponding to the business operation data. A multi-task processing module, which inputs the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include a risk identification service for a target merchant in a first business operation scenario and a risk identification service for a target merchant whose first business operation scenario is converted into a second business operation scenario. A risk identification module, which respectively determines a risk identification result for the target merchant through a pre-trained risk identification network based on the risk representation of each task.

[0008] A risk identification device provided by an embodiment of this specification, the risk identification device includes: a processor; and a memory arranged to store computer-executable instructions, the executable instructions, when executed, cause the processor to: acquire business operation data related to the business operation behavior of a target merchant in a first business operation scenario. Perform encoding processing on the business operation data to obtain a data representation corresponding to the business operation data. Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include a risk identification service for a target merchant in a first business operation scenario and a risk identification service for a target merchant whose first business operation scenario is converted into a second business operation scenario. Based on the risk representation of each task, respectively determine a risk identification result for the target merchant through a pre-trained risk identification network.

[0009] An embodiment of this specification also provides a storage medium for storing computer-executable instructions. When the executable instructions are executed by a processor, the following process is implemented: Obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario. Perform encoding processing on the business operation data to obtain a data representation corresponding to the business operation data. Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include a risk identification service for the target merchant in the first business operation scenario and a risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to a second business operation scenario. Based on the risk representation for each task, respectively determine a risk identification result for the target merchant through a pre-trained risk identification network.

[0010] An embodiment of this specification also provides a computer program product, including a computer program. When the computer program is executed by a processor, the following process is implemented: Obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario. Perform encoding processing on the business operation data to obtain a data representation corresponding to the business operation data. Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include a risk identification service for the target merchant in the first business operation scenario and a risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to a second business operation scenario. Based on the risk representation for each task, respectively determine a risk identification result for the target merchant through a pre-trained risk identification network. Description of the Drawings

[0011] To more clearly illustrate the technical solutions in the embodiments of this specification or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in this specification. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0012] Figure 1 This is an embodiment of a risk identification method in this specification;

[0013] Figure 2 This is another embodiment of a risk identification method in this specification;

[0014] Figure 3 It is a conversion schematic diagram of a business operation scenario in this specification;

[0015] Figure 4 It is an embodiment of another risk identification method in this specification;

[0016] Figure 5 It is an embodiment of another risk identification method in this specification;

[0017] Figure 6 It is a schematic diagram of a risk identification process in this specification;

[0018] Figure 7 It is a schematic diagram of another risk identification process in this specification;

[0019] Figure 8 It is an embodiment of another risk identification method in this specification;

[0020] Figure 9 It is an embodiment of a risk identification device in this specification;

[0021] Figure 10 It is an embodiment of a risk identification device in this specification. Specific implementation manners

[0022] The embodiments of this specification provide a risk identification method, device and equipment.

[0023] In order to enable those skilled in the art to better understand the technical solutions in this specification, the following will clearly and completely describe the technical solutions in the embodiments of this specification with reference to the accompanying drawings in the embodiments of this specification. Obviously, the described embodiments are only a part of the embodiments of this specification, rather than all the embodiments. Based on the embodiments in this specification, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this specification.

[0024] The embodiments of this specification provide a multi-stage risk identification mechanism in the merchant life cycle. With the development of informatization and digitization, the scale and complexity of merchant risks are constantly increasing. Usually, risk decisions highly rely on historical cases and expert opinions, and after imposing penalties on merchants, significant subsequent manual operation and maintenance costs are required, resulting in the current situation of a large decision-making system, low flexibility in operation and maintenance, and high consumption of manpower and time. Therefore, an automated risk decision-making system can be established, and the prerequisite for establishing an automated risk decision-making system is to be able to accurately predict the probability of risk occurrence PRO (Probability of Risk Occurrence) of merchants. Since the behavioral manifestations and probabilities of merchants before risk occurrence often vary greatly in different business operation scenarios, how to predict the risk probability of merchants is an important topic that needs attention currently.

[0025] Generally, the entire life cycle of a merchant can be divided into different cycle intervals, and then corresponding models can be constructed for each cycle interval respectively. However, the above method will consume a large amount of resources for model construction and model maintenance. In addition, business operation data samples related to the business operation behavior of merchants can also be unified and merged, and machine learning models or deep learning models can be directly trained for a certain risk label. This learning method is simple and effective. However, this method ignores the life cycle information of merchants, and the risk bases of merchants in different cycle intervals vary greatly, resulting in poor model learning effects and poor risk learning ability, thus leading to poor model decision-making efficiency. Therefore, the embodiments of this specification propose a multi-stage merchant risk probability prediction mechanism based on the life cycle to improve the overall accuracy of risk probability prediction and the decision-making ability of the automated decision-making system. The specific processing can refer to the specific content in the following embodiments.

[0026] As Figure 1 shown, the embodiments of this specification provide a risk identification method. The execution subject of this method can be a terminal device or a server, etc. The terminal device can be a mobile terminal device such as a mobile phone or a tablet computer, or a computer device such as a laptop computer or a desktop computer, or it can also be an IoT device (specifically such as a smart watch, a vehicle-mounted device, etc.). The server can be an independent server or a server cluster composed of multiple servers. The server can be a background server for financial business or online shopping business, etc., or a background server for a certain application program, etc. In this embodiment, the case where the execution subject is a server is taken as an example for detailed description. For the case where the execution subject is a terminal device, it can refer to the case of the following server for processing, which will not be elaborated here. This method can specifically include the following steps:

[0027] In step S102, obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario.

[0028] Among them, the first business operation scenario can be a scenario that can be converted with another or multiple business operation scenarios. For example, a live broadcast-based business operation scenario, an offline business operation scenario, an online business operation scenario, etc., can be specifically set according to the actual situation. The target merchant can be any merchant. The target merchant can be a merchant selling specified goods or a merchant providing specified services (such as storage services of cloud servers, recharge services, etc.), and can be specifically set according to the actual situation. The embodiments of this specification do not make limitations in this regard. The business operation behavior can include various types. For example, the business operation behavior can include the transaction behavior of the target merchant, the association behavior between different merchants, the association behavior between the target merchant and users, the relevant behavior of the target merchant, etc., and can be specifically set according to the actual situation. The business operation data can include various types. For example, the business operation data can include one or more of the historical transaction data of the target merchant, the current transaction data of the target merchant, the association relationship between the target merchant and different merchants, the association relationship between the target merchant and users, the account information of the target merchant, the transaction risk information of the target merchant, the situation of the target merchant being complained about, etc. The business operation data can also be composed of multiple different types of data. For example, the business operation data can include one or more of text-type data, image-type data, video-type data, tabular data (i.e., Tabular Data), etc., and can be specifically set according to the actual situation. The embodiments of this specification do not make limitations in this regard.

[0029] In implementation, when it is necessary to detect or identify whether a certain merchant (i.e., the target merchant) in the first business operation scenario has a specified risk (such as fraud risk, illegal financial transaction risk, privacy leakage risk, etc.), the business operation data related to the business operation behavior of the target merchant recorded in a specified time period or the time period before the current moment can be obtained. Specifically, for example, the historical transaction data of the target merchant in the most recent month or year (which can include transaction identifiers, account information of users, transaction amounts, and transaction times), the association relationship between the target merchant and different merchants, the association relationship between the target merchant and different users, the account information of the target merchant, the transaction risk information of the target merchant, the situation of the target merchant being complained about, and other business operation data can be obtained.

[0030] Alternatively, when a user needs to execute a certain service (such as a resource trading service (such as a payment service, a transfer service, etc.)) provided by a certain merchant (i.e., the target merchant) in the first business operation scenario, the terminal device can record the relevant data involved in the process of the user executing the service. The recorded data may include business operation data related to the business operation behavior of the target merchant in the first business operation scenario. In order to identify whether there is a preset risk in the above business processing, the server can obtain the business operation data related to the business operation behavior of the target merchant from the data recorded in the terminal device, or the terminal device can also send a risk identification request for the above service to the server. The server can respond to the risk identification request and obtain the business operation data related to the business operation behavior of the target merchant, etc.

[0031] In step S104, the business operation data is encoded to obtain the data representation corresponding to the business operation data.

[0032] Among them, the data representation can be presented in the form of a matrix, can also be presented in the form of a vector, and can also be presented in other forms, etc., which can be specifically set according to the actual situation.

[0033] In implementation, the business operation data may contain various different types of data. For different types of data, corresponding encoders or deep learning network models can be set, and the specified type of data in the business operation data is encoded through the corresponding encoder or deep learning network model to obtain the data representation corresponding to each type of data. For example, for text data, an encoder or a deep learning network model can be pre-trained, or a trained encoder or deep learning network model can be directly obtained from a specified database, and the above text data is encoded through the trained encoder or deep learning network model to obtain the data representation corresponding to the text data; for image data, an image encoder or an image coding model can be pre-trained, or a trained image encoder or image coding model can be directly obtained from a specified database, and the above image data is encoded through the trained image encoder or image coding model to obtain the data representation corresponding to the image data, etc.

[0034] In addition, for a certain type of data, which may include multiple different categories of data, corresponding encoders or deep learning network models can be set for each category of data respectively. For example, for data in the category of "transaction identifier", data in the category of "account information of the target merchant", and data in the category of "transaction amount", corresponding encoders or deep learning network models can be set respectively, which can be specifically set according to the actual situation.

[0035] Data representations corresponding to different types of data can be fused (e.g., the data representations corresponding to different types of data can be concatenated, or the data representations corresponding to different types of data can be fused and calculated through a specified algorithm, etc.) to obtain the data representation corresponding to the business operation data. Alternatively, the data representation corresponding to the business operation data can also be determined by weighted summation. Specifically, corresponding weights can be set in advance for the data representations corresponding to different types of data. Then, the data representations corresponding to different types of data are multiplied by the corresponding weights, and the data representations corresponding to different types of data multiplied by the corresponding weights are added together. The result obtained is used as the data representation corresponding to the business operation data. Or, other algorithms can be set in advance, and the data representations corresponding to different types of data are processed through this algorithm. The result obtained can be used as the data representation corresponding to the business operation data, etc. It can be specifically set according to the actual situation, and the embodiments of this specification do not make any limitations in this regard.

[0036] In step S106, the data representation corresponding to the business operation data is input into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks in multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and gating networks corresponding to the tasks. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant that is converted from the first business operation scenario to the second business operation scenario.

[0037] Among them, multi-task learning network is a machine learning method that adopts an inductive transfer mechanism. Its main goal is to improve the generalization ability of the network by using the domain-specific information hidden in the training signals of multiple related tasks. The multi-task learning network achieves this goal by training multiple tasks in parallel using shared representations, that is, the multi-task learning network can be a network that learns multiple tasks in the same system. These tasks may have similar input spaces, output spaces, or structures. The goal of the multi-task learning network is to improve the generalization ability and performance of the network model by learning multiple tasks. To achieve the above processing, multiple expert models and the corresponding gating network for the task can be set in the multi-task learning network. Specifically, for example, four expert models and three gating networks can be set, or three expert models and three gating networks can be set, or three expert models and two gating networks can be set, etc. It can be specifically set according to the actual situation. The internal structure of the expert model is no longer simply automatically generated based on algorithm fitting data, but is constructed based on rules with business meanings. The decision path passed by each judgment result can be intuitively displayed, and the triggered rules are obvious at a glance. After the expert model is developed, it is equivalent to providing a standard model template for such scenarios, specifying the fields and features required by the model, including the model file. When applied to the same scenario later, only data ETL (i.e., extraction, transformation, and loading) and feature calculation need to be performed according to the constraints of the model, and the cross-platform model file can also be quickly deployed and applied. The gating network can assign corresponding weights to the output data of each expert model. The risk representation can be a representation for the target merchant that can be input into the risk identification network to obtain the corresponding risk identification result. The risk representation can include various forms. The risk representation can be presented in the form of a vector or matrix, etc. (in this case, the vectors or matrices corresponding to each risk can be preset in advance), or it can also be presented in the form of a numerical value, etc. It can be specifically set according to the actual situation, and the embodiments of this specification do not make any limitations in this regard. The second business operation scenario can be a business operation scenario different from the first business operation scenario. For example, if the first business operation scenario is an offline business operation scenario, the second business operation scenario can be a live broadcast-based business operation scenario, or if the first business operation scenario is a live broadcast-based business operation scenario, the second business operation scenario can be an offline business operation scenario, etc. It can be specifically set according to the actual situation.

[0038] In implementation, the multi-task learning network can be processed in the following manner: Construct a shared feature representation: First, a shared feature representation needs to be constructed, which can be a low-dimensional representation of the input data or a high-dimensional representation obtained through a certain feature extraction algorithm; Learn task-specific output layers: In a neural network, the output layers for each task need to be learned, and these output layers can be different tasks such as classification, regression, semantic annotation, etc.; Train the multi-task learning network. When training the multi-task learning network, the loss function for each task needs to be considered, and the above loss functions can be added together to obtain a total loss function. By optimizing the total loss function, multiple tasks can be learned simultaneously; Evaluate the model performance: When evaluating the model performance, the performance of each task can be evaluated through a test set, and the average performance of the multi-task learning network can be calculated.

[0039] In practical applications, the multi-task learning network can include multiple expert models and a gating network. The data representations corresponding to the business operation data can be respectively input into the multiple expert models in the multi-task learning network. After each expert model processes the data representation, corresponding output data can be obtained. At the same time, the data representation corresponding to the business operation data can be input into the gating network. Through the gating network, corresponding weights can be assigned to each expert model. Then, the product between the output data of the expert model and the corresponding weights can be calculated in a weighted summation manner. After that, the results obtained from the products can be added together. Through the above method, the relationships between different tasks among multiple tasks for the target merchant can be determined, and thus the result of the weighted summation can be obtained. Based on the obtained result of the weighted summation, the risk representation for each task can be determined.

[0040] In step S108, based on the risk representation for each task, the risk identification result for the target merchant is respectively determined through a pre-trained risk identification network.

[0041] Among them, the risk identification network can be constructed in various ways. For example, the risk identification network can be constructed through specified network layers (such as fully connected layers, convolutional layers, activation function layers, etc.), or can be constructed through a multi-layer perceptron, etc. It can be specifically set according to the actual situation.

[0042] In implementation, the risk characterizations of each task can be separately input into the sub-network corresponding to each task in the risk identification network, and the risk identification results output by the sub-network corresponding to each task can be obtained respectively. The risk identification result for the target merchant can be determined based on the risk identification results output by the sub-network corresponding to each task. For example, if the first business operation scenario is an offline business operation scenario and the second business operation scenario can be a live broadcast-based business operation scenario, based on this, two tasks can be set, namely the risk identification task in the offline business operation scenario and the risk identification task in the business operation scenario that is converted from the offline business operation scenario to a live broadcast-based business operation scenario. Through the above processing, the risk identification result of the target merchant in the offline business operation scenario and the risk identification result of the target merchant in the business operation scenario that is converted to a live broadcast-based business operation scenario can be determined. After obtaining the risk identification result for the target merchant in the above manner, the risk identification result for the target merchant can be provided to subsequent systems or services for processing. For example, the risk identification result for the target merchant can be provided to an automated decision-making system to execute corresponding risk strategies for the target merchant. Specifically, such as restricting the collection amount of the target merchant or freezing the account of the target merchant, etc., which can be specifically set according to the actual situation.

[0043] It should be noted that the conversion of the target merchant between the first business operation scenario and the second business operation scenario can be a complete conversion. For example, if the target merchant is a merchant in the first business operation scenario, only convert the target merchant to the second business operation scenario; or, it can also be a part of a certain conversion process. For example, if the target merchant is a merchant in the first business operation scenario, convert the target merchant to the second business operation scenario, and then convert the target merchant back to the first business operation scenario (or the third business operation scenario, etc.).

[0044] An embodiment of this specification provides a risk identification method. By obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario, then, the business operation data can be encoded to obtain a data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation for each task, the risk identification result for the target merchant can be determined respectively through a pre-trained risk identification network. In this way, using a multi-task learning network to learn for different life cycles of the target merchant can greatly improve the risk identification efficiency of merchants in multiple stages of the life cycle, and at the same time greatly reduce the maintenance difficulty of the risk identification system. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring a preset risk without conversion (i.e., in the first business operation scenario) and the prediction results of the merchant occurring a preset risk after being converted to the second business operation scenario can be learned simultaneously. Since the behavior performance and probability of the merchant before occurring a risk in the first business operation scenario are often quite different from those of the merchant in the second business operation scenario, the above method can greatly improve the model effect.

[0045] In practical applications, during the process of risk identification for a target merchant, since the behavior performance and probability before a preset risk occurs for the target merchant in the offline business operation scenario and the online business operation scenario are often quite different, if the entire life cycle of the target merchant is directly merged and the subsequent risk occurrence probability is directly predicted, the learning difficulty is relatively large and the learning effect is often not good. Based on this, the first business operation scenario is the offline business operation scenario, and the second business operation scenario is the online business operation scenario, or the first business operation scenario is the online business operation scenario, and the second business operation scenario is the offline business operation scenario. Among them, the offline business operation scenario can be the operation scenario through an offline physical store, and the online business operation scenario can be the operation scenario through an online store in an online trading platform. In the above life cycle of the target merchant, the merchant in the online business operation scenario is often registered through a small program that converts from the offline business operation scenario to the online business operation scenario, and the behavior performance and probability before a preset risk occurs for the target merchant in the offline business operation scenario and the online business operation scenario are often quite different.

[0046] In practical applications, the specific processing method of the above step S108 can be various. The following provides an optional processing method. For example, Figure 2 as shown, it specifically may include the processing of the following steps S1082 to S1086.

[0047] In step S1082, the risk characteristics corresponding to the risk identification service of the target merchant in the first business operation scenario are input into the corresponding risk identification network, and the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario is obtained.

[0048] In step S1084, based on the risk characteristics corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario, through the risk identification network corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario, the risk occurrence probability corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario is determined.

[0049] In step S1086, based on the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario, the risk identification result for the target merchant is determined.

[0050] In implementation, the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario can be used as the risk identification result for the target merchant. Or, the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario can be fused and calculated, and the calculation result obtained is used as the risk identification result for the target merchant, etc. Specifically, it can be set according to the actual situation, and the embodiments of this specification do not make any limitations in this regard.

[0051] In practical applications, the life cycle of the target merchant in the offline business operation scenario can be as Figure 3As shown, that is, the target merchant may generate a preset risk as an entity in the offline business operation scenario. If the target merchant switches from the offline business operation scenario to the online business operation scenario, the target merchant may also generate a preset risk after the conversion. For the life cycle of the target merchant in the offline business operation scenario, the risk identification service for the target merchant that switches from the first business operation scenario to the second business operation scenario includes a first sub-service for the conversion of the business operation scenario (i.e., the first sub-service can be a sub-service for the conversion from the first business operation scenario to the second business operation scenario (e.g., the first sub-service can be the conversion from the offline business operation scenario to the online business operation scenario)) and a second sub-service for the target merchant that switches the business operation scenario (i.e., the second sub-service can be a sub-service for the conversion to the second business operation scenario (e.g., the second sub-service can be a sub-service for the conversion from the offline business operation scenario to the online business operation scenario)). The risk characteristics corresponding to the risk identification service for the target merchant that switches from the first business operation scenario to the second business operation scenario include a first sub-characteristic corresponding to the first sub-service (which can include one or more of the trading behavior of the target merchant, the association behavior between the target merchant and different merchants, the association behavior between the target merchant and users, the relevant behavior of the target merchant, etc.) and a second sub-characteristic corresponding to the second sub-service (which can include one or more of the trading behavior of the target merchant, the association behavior between the target merchant and different merchants, the association behavior between the target merchant and users, the relevant behavior of the target merchant, etc.). The risk identification network corresponding to the risk identification service for the target merchant that switches from the first business operation scenario to the second business operation scenario includes a first sub-network corresponding to the first sub-service and a second sub-network corresponding to the second sub-service. Based on this, the specific processing method of the above step S1084 can be various. The following provides an optional processing method, such as Figure 4 as shown, which may specifically include the processing of the following steps S10842 to S10846.

[0052] In step S10842, the first sub-characteristic corresponding to the first sub-service is input into the first network to obtain the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario.

[0053] In step S10844, the second sub-characteristic corresponding to the second sub-service is input into the second sub-network to obtain the risk occurrence probability corresponding to the risk identification service of the target merchant that switches to the second business operation scenario.

[0054] In step S10846, based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant that switches to the second business operation scenario, the risk occurrence probability corresponding to the risk identification service of the target merchant that switches from the first business operation scenario to the second business operation scenario is determined.

[0055] In implementation, the conversion probability of converting a target merchant from a first business operation scenario to a second business operation scenario can be multiplied by the risk occurrence probability corresponding to the risk identification service of the target merchant converted to the second business operation scenario to obtain a corresponding result, and the obtained corresponding result can be determined as the risk occurrence probability corresponding to the risk identification service of the target merchant converted from the first business operation scenario to the second business operation scenario. Alternatively, a corresponding algorithm can be preset according to the actual situation, and through this algorithm, the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant converted to the second business operation scenario are calculated, and the calculation result is used as the risk occurrence probability corresponding to the risk identification service of the target merchant converted from the first business operation scenario to the second business operation scenario, etc. It can be specifically set according to the actual situation, and the embodiments of this specification do not limit this.

[0056] In practical applications, business operation data includes tabular data, and the data representations corresponding to the business operation data include the data representations corresponding to various different table features included in the tabular data. The tabular data (TabularData) therein can be data in table format, and the table data can include categorical table features and numerical table features, as shown in Table 1 for example.

[0057] Table 1

[0058]

[0059] In the above Table 1, "account information of the target merchant", "account information of the user", "transaction bill identifier", "transaction amount", and "transaction time", etc. can be categorical table features, and the remaining data in Table 1 such as "10000000", "0010005", "0058296", "32000003", "42000204", "58701258", "500", "3000", "1000", "10:50:42", etc. can be numerical table features. In business such as resource trading business, information recommendation business, data mining business, etc., the most commonly used data modality is tabular (Tabular) data. In addition, in risk identification and risk prevention and control, tabular data can also be the most commonly used data form. Based on this, tabular data of a user in the process of executing a target business can be obtained.

[0060] Based on the above content, the specific processing method of the above step S106 can be various. The following provides an optional processing method, as Figure 5 shown, and specifically can include the processing of the following steps S1062 to S1066.

[0061] In step S1062, data representations corresponding to various different table features included in the business operation data are concatenated to obtain a concatenated data representation.

[0062] In implementation, data representations corresponding to various different table features may include data representations corresponding to categorical table features and data representations corresponding to numerical table features. Data representations corresponding to categorical table features and data representations corresponding to numerical table features can both be presented in the form of a matrix, or in the form of a vector, or in other forms, etc., which can be specifically set according to the actual situation. For categorical table features, since categorical table features are usually text or corresponding codes and are generally relatively fixed, therefore, through the EmbeddingLook-Up operation, categorical table features can be converted into corresponding Embedding values to project into a high-dimensional feature space to enhance their expressive ability. Specifically, a mapping relationship between categorical table features and embedding vectors can be preset, and this mapping relationship can be stored in a specified database or storage device. When table data is obtained, categorical table features can be extracted from it, and the embedding vectors corresponding to the categorical table features in the table data can be found from the above mapping relationship, so that the categorical table features in the table data can be converted into corresponding embedding vectors, and the obtained embedding vectors after conversion can be used as the data representations corresponding to the categorical table features. For numerical table features, usually, after raw value input or binning or bucketing processing (which can include multiple types, such as equal-frequency binning processing, equal-width binning processing, etc., which can be specifically set according to the actual situation), Embedding processing is performed to obtain data representations corresponding to numerical table features. After that, data representations corresponding to various different table features included in the business operation data are concatenated to obtain a concatenated data representation.

[0063] In step S1064, the concatenated data representation is respectively input into each expert model in the multi-task learning network to obtain output data of each expert model.

[0064] Among them, multiple expert models (i.e., Experts) can be parallel network models. Each network model learns for the entire sample set, but as the model training progresses, each expert model usually becomes specialized and thus becomes an "expert" for processing specific types of data or specific tasks.

[0065] In step S1066, the spliced data representation and the output data of each expert model are input into the gating network corresponding to the task, to obtain the weights assigned by the gating network to each expert model, and based on the output data of each expert model and the weight corresponding to each expert model, to determine the risk representation of each task that includes the relationship between different tasks among multiple tasks for the target merchant.

[0066] Among them, the task of the gating network is to dynamically assign the weights of the expert models to each data point. The gating network decides, based on the input data, which expert models' output data should dominate in the final prediction.

[0067] In implementation, the spliced data representation and the output data of each expert model can be input into the gating network corresponding to the task. The gating network corresponding to the task decides which expert models' output data should dominate in the final prediction, and then obtains the weights assigned by the gating network to each expert model. The output data of a certain expert model can be multiplied by the weight corresponding to the expert model to obtain the corresponding result. Through the above method, the results of multiplying the output data of each expert model by their corresponding weights can be obtained. The calculated results can be added together, and the obtained result is used as the risk representation of each task.

[0068] In practical applications, the number of gating networks in the above step S1066 is determined by the number of multiple tasks, and each task corresponds to a gating network. The corresponding structure can be like the MMOE network structure, as Figure 6 shown. Then, in the above step S1066, there can be various specific processing methods for inputting the spliced data representation and the output data of each expert model into the gating network corresponding to the task to obtain the weights assigned by the gating network to each expert model. Here is another optional processing method, which can specifically include the following: The spliced data representation and the output data of each expert model are respectively input into the gating network corresponding to each task to obtain the weights assigned by each gating network to each expert model.

[0069] In addition, the multiple expert models in the above step S106 include the expert models corresponding to each task and the shared expert models shared by multiple tasks. The corresponding structure can be like the PLE network structure, as Figure 7 shown. Based on this, there can be various specific processing methods for the above step S106. Here is another optional processing method, which can specifically include the processing of the following step A2 and step A4.

[0070] In step A2, the data representations corresponding to the business operation data are respectively input into the expert models corresponding to each task and the shared expert model shared by multiple tasks, and the output data of the expert models corresponding to each task and the output data of the shared expert model are obtained.

[0071] In implementation, the multiple expert models can include two types. One is an expert model corresponding to each task, and the other is a shared expert model shared by multiple tasks. The shared expert model can include one or more, and can be specifically selected according to the actual situation. To simplify the processing process, the shared expert model can include one. At this time, all tasks can share one shared expert model. The shared expert model and the expert models corresponding to each task are basically the same in data processing as the expert models mentioned above, and there is no obvious difference. Therefore, the specific processing process of the above step A2 can refer to the relevant content mentioned above and will not be elaborated here.

[0072] In step A4, for the first task among multiple tasks, the data representation corresponding to the business operation data is input into the gating network corresponding to the first task, and the weights assigned by the first gating network to the output data of the expert model corresponding to the first task and the output data of the shared expert model are obtained; based on the output data of the expert model corresponding to the first task and the corresponding weights, and the output data of the shared expert model and the corresponding weights, the risk representation of the first task is determined to determine the risk representations of each task.

[0073] Among them, the first task can be any one of the multiple tasks.

[0074] In implementation, for the first task, the data representation corresponding to the business operation data can be input into the gating network corresponding to the first task, and the weights assigned by the first gating network to the output data of the expert model corresponding to the first task and the output data of the shared expert model are obtained. Then, the output data of the expert model corresponding to the first task can be multiplied by the corresponding weights to obtain the corresponding result, the output data of the shared expert model can be multiplied by the corresponding weights to obtain the corresponding result, and the above two results can be added, and the obtained result is used as the risk representation of the first task. Through the above method, the risk representations of other tasks among multiple tasks can be obtained, and then the risk representations of each task can be obtained.

[0075] The training methods of the above multi-task learning network and risk identification network can be various. The following provides an optional processing method, which can specifically refer to the processing in steps B2 to B10 below.

[0076] In step B2, business operation data samples related to the business operation behavior of the first merchant in the first business operation scenario for model training are obtained.

[0077] Among them, the first merchant can be any merchant. In addition, the first merchant can include the target merchant.

[0078] In step B4, the business operation data sample is encoded to obtain the sample data representation corresponding to the business operation data sample.

[0079] For the specific processing procedures of the above steps B2 and B4, reference can be made to the relevant content mentioned above, which will not be elaborated here.

[0080] In step B6, the sample data representation corresponding to the business operation data sample is input into the multi-task learning network to obtain the risk representation sample of each task including the relationship between different tasks in multiple tasks for the first merchant. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant when the first business operation scenario is converted to the second business operation scenario.

[0081] For the specific processing procedure of the above step B6, reference can be made to the relevant content of the foregoing step S106, which will not be elaborated here.

[0082] In practical applications, the processing in step B6 may further include: the multiple expert models are expert models shared by multiple tasks. The sample data representation corresponding to the business operation data sample is respectively input into each expert model in the multi-task learning network to obtain the output data sample of each expert model; the sample data representation corresponding to the business operation data sample and the output data sample of each expert model are input into the gating network corresponding to the task to obtain the weight assigned by the gating network to each expert model; based on the output data sample of each expert model and the sample weight corresponding to each expert model, the risk representation sample of each task is determined.

[0083] In addition, the number of gating networks is determined by the number of multiple tasks, and each task corresponds to a gating network. Inputting the sample data representation corresponding to the business operation data sample and the output data sample of each expert model into the gating network corresponding to the task respectively to obtain the sample weight assigned by the gating network to each expert model includes: inputting the sample data representation corresponding to the business operation data sample and the output data sample of each expert model into the gating network corresponding to each task respectively to obtain the sample weight assigned by each gating network to each expert model.

[0084] In addition, the multiple expert models include expert models corresponding to each task and a shared expert model shared by multiple tasks. The processing in step B6 may further include: respectively inputting the sample data representations corresponding to the business operation data samples into the expert models corresponding to each task and the shared expert model shared by multiple tasks to obtain the output data samples of the expert models corresponding to each task and the output data samples of the shared expert model; for the first task among multiple tasks, inputting the sample data representation corresponding to the business operation data sample into the gating network corresponding to the first task to obtain the weights assigned by the first gating network to the output data samples of the expert model corresponding to the first task and the output data samples of the shared expert model; based on the output data samples of the expert model corresponding to the first task and the corresponding sample weights, and the output data samples of the shared expert model and the corresponding sample weights, determining the risk representation sample of the first task, so as to determine the risk representation samples of each task.

[0085] In step B8, based on the risk representations of each task, the risk identification results of the first merchant are respectively determined through the risk identification network.

[0086] For the specific processing process of the above step B8, reference can be made to the relevant content of the foregoing step S108, which will not be elaborated here.

[0087] In step B10, based on the risk identification results of the first merchant and a preset loss function, the multi-task learning network and the risk identification network are jointly trained to obtain the trained multi-task learning network and the trained risk identification network.

[0088] Among them, the loss function can include various types, for example, the mean square error loss function or the cross entropy loss function, etc., which can be specifically set according to the actual situation, and the embodiments of this specification do not make any limitations in this regard.

[0089] In practical applications, the above loss function includes a first sub-loss function and a second sub-loss function. The first sub-loss function is constructed through the cross entropy loss function based on the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk label corresponding to the target merchant in the first business operation scenario (that is, whether the target merchant has a preset risk in the identity of the first business operation scenario (such as the offline business operation scenario)). The second sub-loss function is constructed through the cross entropy loss function based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant in the second business operation scenario after conversion, and the risk label corresponding to the risk identification service of the target merchant in the second business operation scenario after conversion from the first business operation scenario (that is, whether the target merchant has a preset risk in the identity of the converted second business operation scenario (such as the online business operation scenario)).

[0090] Among them, the risk identification service for target merchants that converts from the first business operation scenario to the second business operation scenario may include a first sub-service and a second sub-service. Based on this, the label corresponding to the first sub-service may be whether the target merchant performs the conversion from the first business operation scenario to the second business operation scenario, and the label corresponding to the second sub-service may be whether the target merchant has a preset risk in the identity of the converted second business operation scenario. Based on this, the output data of the first sub-service can be multiplied by the output data of the second sub-service, a conditional probability can be calculated, and then the cross-entropy loss can be calculated with the label corresponding to the second sub-service, thereby obtaining the above loss function.

[0091] The following describes in detail a risk identification method provided by an embodiment of this specification in combination with a specific application scenario. The first business operation scenario may be an offline business operation scenario, the second business operation scenario may be an online business operation scenario, and the business operation data may include tabular data. The tabular data may include historical transaction data between the target merchant and different users (specifically, it may include, for example, user behavior information, account information of both trading parties, transaction amount, transaction time, and transaction location, etc.), account information of the target merchant, historical risk information of the target merchant, etc. The multi-task learning network includes four expert models and three gating networks.

[0092] Such as Figure 8 As shown, an embodiment of this specification provides a risk identification method. The execution subject of this method may be a terminal device or a server, etc. The terminal device may be a mobile terminal device such as a mobile phone or a tablet computer, or may also be a computer device such as a laptop computer or a desktop computer, or may also be an IoT device (specifically, such as a smart watch, a vehicle-mounted device, etc.). The server may be an independent server or a server cluster composed of multiple servers. The server may be a background server for financial services or online shopping services, etc., or may also be a background server for a certain application program, etc. In this embodiment, the case where the execution subject is a server is taken as an example for detailed description. For the case where the execution subject is a terminal device, reference may be made to the following treatment of the server case, which will not be elaborated here. This method may specifically include the following steps:

[0093] In step S802, obtain a tabular data sample related to the business operation behavior of the first merchant in the offline business operation scenario for model training.

[0094] In step S804, perform encoding processing on the tabular data sample to obtain a sample data representation corresponding to the tabular data sample.

[0095] In step S806, the sample data representation corresponding to the tabular data sample is input into the multi-task learning network to obtain a risk representation sample for each task that includes the relationships between different tasks among multiple tasks for the first merchant. The multiple tasks at least include the risk identification service for the target merchant in the offline business operation scenario and the risk identification service for the target merchant that has been converted from the offline business operation scenario to the online business operation scenario.

[0096] In step S808, based on the risk representation sample for each task, the risk identification result of the first merchant is determined through the risk identification network respectively.

[0097] In step S810, based on the risk identification result of the first merchant and the preset loss function, the multi-task learning network and the risk identification network are jointly trained to obtain the trained multi-task learning network and the trained risk identification network.

[0098] Among them, the loss function includes a first sub-loss function and a second sub-loss function. The first sub-loss function is constructed by the cross-entropy loss function based on the risk occurrence probability corresponding to the risk identification service for the target merchant in the offline business operation scenario and the risk label corresponding to the target merchant in the offline business operation scenario. The second sub-loss function is constructed by the cross-entropy loss function based on the conversion probability of the target merchant from the offline business operation scenario to the online business operation scenario, the risk occurrence probability corresponding to the risk identification service for the target merchant that has been converted to the online business operation scenario, and the risk label corresponding to the risk identification service for the target merchant that has been converted from the offline business operation scenario to the online business operation scenario.

[0099] In step S812, tabular data related to the business operation behavior of the target merchant in the offline business operation scenario is obtained.

[0100] In step S814, the tabular data is encoded to obtain the data representation corresponding to the tabular data.

[0101] In step S816, the data representations corresponding to various different table features included in the tabular data are concatenated to obtain the concatenated data representation.

[0102] In step S818, the concatenated data representation is input into each expert model in the multi-task learning network respectively to obtain the output data of each expert model.

[0103] In step S820, the spliced data representation and the output data of each expert model are respectively input into the gating network corresponding to each task, to obtain the weights assigned by each gating network to each expert model, and based on the output data of each expert model and the weight corresponding to each expert model, determine the risk representation of each task that includes the relationship between different tasks among multiple tasks for the target merchant.

[0104] In step S822, the risk representation corresponding to the risk identification service of the target merchant in the offline business operation scenario is input into the corresponding risk identification network, to obtain the risk occurrence probability corresponding to the risk identification service of the target merchant in the offline business operation scenario.

[0105] In step S824, the representation corresponding to the business operation scenario conversion (i.e., the above-mentioned first sub-business) that converts from the offline business operation scenario to the online business operation scenario is input into the corresponding risk identification network, to obtain the conversion probability of the target merchant converting from the offline business operation scenario to the online business operation scenario.

[0106] In step S826, the representation corresponding to the risk identification service of the target merchant that has been converted to the online business operation scenario (i.e., the above-mentioned second sub-business) is input into the corresponding risk identification network, to obtain the risk occurrence probability corresponding to the risk identification service of the target merchant that has been converted to the online business operation scenario.

[0107] In step S828, based on the conversion probability of the target merchant converting from the offline business operation scenario to the online business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant that has been converted to the online business operation scenario, determine the risk occurrence probability corresponding to the risk identification service of the target merchant that has been converted from the offline business operation scenario to the online business operation scenario.

[0108] In step S830, based on the risk occurrence probability corresponding to the risk identification service of the target merchant in the offline business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant that has been converted from the offline business operation scenario to the online business operation scenario, determine the risk identification result for the target merchant.

[0109] It should be noted that the conversion of the target merchant between the offline business operation scenario and the online business operation scenario can be a complete conversion. For example, if the target merchant is a merchant in the offline business operation scenario, only the target merchant is converted to the online business operation scenario. Or, it can also be a part of a certain conversion process. For example, if the target merchant is a merchant in the offline business operation scenario, the target merchant is converted to the online business operation scenario and then converted back to the offline business operation scenario. Then, the conversion of the target merchant between the offline business operation scenario and the online business operation scenario can be the conversion of the offline business operation scenario in the first half to the online business operation scenario, or the conversion of the online business operation scenario in the second half to the offline business operation scenario.

[0110] The embodiment of this specification provides a risk identification method. By obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario, then, the business operation data can be encoded to obtain a data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant converted from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation for each task, the risk identification result for the target merchant can be determined respectively through a pre-trained risk identification network. In this way, by using a multi-task learning network to learn for different life cycles of the target merchant, the risk identification efficiency of multi-stage merchants in the life cycle is greatly improved, and at the same time, the maintenance difficulty of the risk identification system is greatly reduced. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring a preset risk without conversion (i.e., in the first business operation scenario) and the prediction results of the merchant occurring a preset risk after being converted to the second business operation scenario can be learned simultaneously. Since the behavior performance of the merchant before the risk occurs in the first business operation scenario is often quite different from that of the merchant in the second business operation scenario, the above method can greatly improve the model effect.

[0111] The above is the risk identification method provided by the embodiment of this specification. Based on the same idea, the embodiment of this specification also provides a risk identification device, as Figure 9 shown.

[0112] The risk identification device includes: a data acquisition module 901, an encoding module 902, a multi-task processing module 903, and a risk identification module 904, where:

[0113] The data acquisition module 901 acquires business operation data related to the business operation behavior of a target merchant in a first business operation scenario.

[0114] The encoding module 902 performs encoding processing on the business operation data to obtain a data representation corresponding to the business operation data.

[0115] The multi-task processing module 903 inputs the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include a risk identification service for a target merchant in a first business operation scenario and a risk identification service for a target merchant whose first business operation scenario is converted to a second business operation scenario.

[0116] The risk identification module 904 respectively determines a risk identification result for the target merchant through a pre-trained risk identification network based on the risk representation for each task.

[0117] In an embodiment of this specification, the first business operation scenario is an offline business operation scenario, and the second business operation scenario is an online business operation scenario, or the first business operation scenario is an online business operation scenario, and the second business operation scenario is an offline business operation scenario.

[0118] In an embodiment of this specification, the risk identification module 904 includes:

[0119] A first prediction unit inputs the risk representation corresponding to the risk identification service for the target merchant in the first business operation scenario into the corresponding risk identification network to obtain a risk occurrence probability corresponding to the risk identification service for the target merchant in the first business operation scenario.

[0120] A second prediction unit determines a risk occurrence probability corresponding to the risk identification service for the target merchant whose first business operation scenario is converted to a second business operation scenario through a risk identification network corresponding to the risk identification service for the target merchant whose first business operation scenario is converted to a second business operation scenario based on the risk representation corresponding to the risk identification service for the target merchant whose first business operation scenario is converted to a second business operation scenario.

[0121] A risk identification unit determines a risk identification result for the target merchant based on the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk occurrence probability corresponding to the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario.

[0122] In the embodiments of this specification, the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario includes a first sub-service of the business operation scenario conversion and a second sub-service of the target merchant whose business operation scenario is converted. The risk characteristics corresponding to the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario include a first sub-characteristic corresponding to the first sub-service and a second sub-characteristic corresponding to the second sub-service. The risk identification network corresponding to the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario includes a first sub-network corresponding to the first sub-service and a second sub-network corresponding to the second sub-service.

[0123] The second prediction unit inputs the first sub-characteristic corresponding to the first sub-service into the first network to obtain the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario; inputs the second sub-characteristic corresponding to the second sub-service into the second sub-network to obtain the risk occurrence probability corresponding to the risk identification service of the target merchant converted to the second business operation scenario; and determines the risk occurrence probability corresponding to the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the second sub-service of the target merchant converted to the second business operation scenario.

[0124] In the embodiments of this specification, the business operation data includes tabular data, and the data characteristics corresponding to the business operation data include the data characteristics corresponding to various different table features included in the tabular data.

[0125] The multi-task processing module 903 includes:

[0126] A splicing unit splices the data characteristics corresponding to various different table features included in the business operation data to obtain a spliced data characteristic.

[0127] A first processing unit inputs the spliced data characteristic into each expert model in the multi-task learning network to obtain the output data of each expert model.

[0128] A second processing unit inputs the spliced data representation and the output data of each expert model into the gating network corresponding to the task, obtains the weights assigned by the gating network to each expert model, and determines the risk representation of each task including the relationship between different tasks among multiple tasks for the target merchant based on the output data of each expert model and the weight corresponding to each expert model.

[0129] In the embodiments of the present specification, the number of the gating networks is determined by the number of the multiple tasks, and each task corresponds to a gating network. The second processing unit inputs the spliced data representation and the output data of each expert model into the gating network corresponding to each task respectively, and obtains the weights assigned by each gating network to each expert model.

[0130] In the embodiments of the present specification, the apparatus further includes:

[0131] A sample acquisition module, which acquires business operation data samples related to the business operation behavior of a first merchant in a first business operation scenario for model training;

[0132] A sample encoding module, which performs encoding processing on the business operation data samples to obtain the sample data representation corresponding to the business operation data samples;

[0133] A sample prediction module inputs the sample data representation corresponding to the business operation data samples into the multi-task learning network, and obtains a risk representation sample of each task including the relationship between different tasks among multiple tasks for the first merchant. The multiple tasks at least include the risk identification service of the target merchant in the first business operation scenario and the risk identification service of the target merchant when the first business operation scenario is converted to the second business operation scenario;

[0134] A sample risk identification module respectively determines the risk identification result of the first merchant through the risk identification network based on the risk representation sample of each task;

[0135] A training module jointly trains the multi-task learning network and the risk identification network based on the risk identification result of the first merchant and a preset loss function, and obtains a trained multi-task learning network and a trained risk identification network.

[0136] In the embodiments of this specification, the loss function includes a first sub-loss function and a second sub-loss function. The first sub-loss function is constructed by a cross-entropy loss function based on the risk occurrence probability corresponding to the risk identification service of the target merchant in the first business operation scenario and the risk label corresponding to the target merchant in the first business operation scenario. The second sub-loss function is constructed by a cross-entropy loss function based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario, the risk occurrence probability corresponding to the risk identification service of the target merchant converted to the second business operation scenario, and the risk label corresponding to the risk identification service of the target merchant converted from the first business operation scenario to the second business operation scenario.

[0137] The embodiments of this specification provide a risk identification device. By obtaining business operation data related to the business operation behavior of a target merchant in the first business operation scenario, then, the business operation data can be encoded to obtain a data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the tasks. The multiple tasks at least include the risk identification service of the target merchant in the first business operation scenario and the risk identification service of the target merchant converted from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation for each task, the risk identification result for the target merchant can be determined respectively through a pre-trained risk identification network. In this way, by using a multi-task learning network to learn for different life cycles of the target merchant, the risk identification efficiency of multi-stage merchants in the life cycle is greatly improved, and at the same time, the maintenance difficulty of the risk identification system is greatly reduced. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring a preset risk without conversion (i.e., in the first business operation scenario) and the prediction results of the merchant occurring a preset risk after being converted to the second business operation scenario can be learned simultaneously. Since the behavior performance of the merchant before the risk occurs in the first business operation scenario is often quite different from that in the second business operation scenario, the above method can greatly improve the model effect.

[0138] The above is the risk identification device provided by the embodiments of this specification. Based on the same idea, the embodiments of this specification also provide a risk identification device, as Figure 10 shown.

[0139] The risk identification device may be the terminal device or server provided in the above embodiments, etc.

[0140] Risk identification devices can vary significantly due to differences in configuration or performance. They can include one or more processors 1001 and a memory 1002. One or more stored application programs or data can be stored in the memory 1002. Among them, the memory 1002 can be short-term storage or persistent storage. The application programs stored in the memory 1002 can include one or more modules (not shown in the figure), and each module can include a series of computer-executable instructions for the risk identification device. Further, the processor 1001 can be set to communicate with the memory 1002 and execute a series of computer-executable instructions in the memory 1002 on the risk identification device. The risk identification device can also include one or more power supplies 1003, one or more wired or wireless network interfaces 1004, one or more input / output interfaces 1005, and one or more keyboards 1006.

[0141] Specifically, in this embodiment, the risk identification device includes a memory and one or more programs. One or more of the programs are stored in the memory, and one or more of the programs can include one or more modules. Each module can include a series of computer-executable instructions for the risk identification device and is configured to be executed by one or more processors. The one or more programs include computer-executable instructions for performing the following:

[0142] Obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario;

[0143] Perform encoding processing on the business operation data to obtain a data representation corresponding to the business operation data;

[0144] Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include the risk identification service for the target merchant in a first business operation scenario and the risk identification service for the target merchant that is converted from the first business operation scenario to a second business operation scenario;

[0145] Based on the risk representation for each task, respectively determine the risk identification result for the target merchant through a pre-trained risk identification network.

[0146] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the embodiment of the risk identification device, since it is basically similar to the method embodiment, the description is relatively simple, and for the relevant parts, reference can be made to the partial description of the method embodiment.

[0147] An embodiment of this specification provides a risk identification device. By obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario, then, the business operation data can be encoded to obtain a data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the task. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation of each task, the risk identification result for the target merchant can be determined respectively through a pre-trained risk identification network. In this way, by using a multi-task learning network to learn for different life cycles of the target merchant, the risk identification efficiency of multi-stage merchants in the life cycle is greatly improved, and at the same time, the maintenance difficulty of the risk identification system is greatly reduced. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring preset risks without conversion (i.e., in the first business operation scenario) and the prediction results of the merchant occurring preset risks after being converted to the second business operation scenario can be learned simultaneously. Since the behavior performance of the merchant before the risk occurs in the first business operation scenario is often quite different from that of the merchant in the second business operation scenario, the above method can greatly improve the model effect.

[0148] Further, based on the above Figures 1 to 8 method shown, one or more embodiments of this specification also provide a storage medium for storing computer-executable instruction information. In a specific embodiment, the storage medium can be a USB flash drive, an optical disc, a hard disk, etc. When the computer-executable instruction information stored in the storage medium is executed by a processor, the following process can be realized:

[0149] Obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario;

[0150] Encode the business operation data to obtain a data representation corresponding to the business operation data;

[0151] Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the tasks. The multiple tasks at least include a risk identification service for the target merchant in a first business operation scenario and a risk identification service for the target merchant that is transformed from the first business operation scenario to a second business operation scenario.

[0152] Based on the risk representation for each task, respectively determine a risk identification result for the target merchant through a pre-trained risk identification network.

[0153] Each embodiment in this specification is described in a progressive manner. For the same or similar parts between the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the above-mentioned storage medium embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment.

[0154] An embodiment of this specification provides a storage medium. By obtaining business operation data related to the business operation behavior of a target merchant in a first business operation scenario, then, the business operation data can be encoded to obtain a data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task including the relationship between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and gating networks corresponding to the tasks. The multiple tasks at least include a risk identification service for the target merchant in the first business operation scenario and a risk identification service for the target merchant whose business operation scenario is converted from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation for each task, the risk identification result for the target merchant can be determined respectively through a pre-trained risk identification network. In this way, using a multi-task learning network to learn for different life cycles of the target merchant can greatly improve the risk identification efficiency of merchants in multiple stages of the life cycle, and at the same time greatly reduce the maintenance difficulty of the risk identification system. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring a preset risk without conversion (i.e., in the first business operation scenario) and the prediction results of the merchant occurring a preset risk after being converted to the second business operation scenario can be learned simultaneously. Since the behavior performance of the merchant before the risk occurs in the first business operation scenario is often quite different from that of the merchant in the second business operation scenario, the above method can greatly improve the model effect.

[0155] Further, based on the above Figures 1 to 8 method shown, one or more embodiments of this specification also provide a computer program product, including a computer program. When the computer program in this computer program product is executed by a processor, the following process can be implemented:

[0156] Obtain business operation data related to the business operation behavior of a target merchant in a first business operation scenario;

[0157] Encode the business operation data to obtain a data representation corresponding to the business operation data;

[0158] Input the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the tasks. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant that is transformed from the first business operation scenario to the second business operation scenario;

[0159] Based on the risk representation for each task, respectively determine the risk identification result for the target merchant through a pre-trained risk identification network.

[0160] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. The key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the above-mentioned embodiment of a computer program product, since it is basically similar to the method embodiment, the description is relatively simple, and reference can be made to the corresponding part of the method embodiment for related content.

[0161] The embodiments of this specification provide a computer program product. By obtaining business operation data related to the business operation behavior of a target merchant in the first business operation scenario, then, the business operation data can be encoded to obtain the data representation corresponding to the business operation data. After that, the data representation corresponding to the business operation data can be input into a pre-trained multi-task learning network to obtain a risk representation for each task that includes the relationships between different tasks among multiple tasks for the target merchant. The multi-task learning network includes multiple expert models and a gating network corresponding to the tasks. The multiple tasks at least include the risk identification service for the target merchant in the first business operation scenario and the risk identification service for the target merchant that is transformed from the first business operation scenario to the second business operation scenario. Finally, based on the risk representation for each task, respectively determine the risk identification result for the target merchant through a pre-trained risk identification network. In this way, by using a multi-task learning network to learn for different life cycles of the target merchant, the risk identification efficiency of multi-stage merchants in the life cycle can be greatly improved, and at the same time, the maintenance difficulty of the risk identification system can be greatly reduced. In addition, the subsequent merchant risk decision system depends on the accuracy of merchant risk prediction or risk identification. The improvement of merchant risk prediction accuracy can improve the merchant decision-making efficiency. Moreover, through the multi-task learning method, the prediction results of directly occurring a preset risk without transformation (i.e., in the first business operation scenario) and the prediction results of the merchant occurring a preset risk after being transformed into the second business operation scenario can be learned simultaneously. Since the behavior performance of the merchant before the risk occurs in the first business operation scenario is often quite different from that in the second business operation scenario, the above method can greatly improve the model effect.

[0162] The foregoing describes specific embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the acts or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain implementations, multitasking and parallel processing are also possible or may be advantageous.

[0163] In the 1990s, it was obvious to distinguish whether an improvement to a technology was a hardware improvement (e.g., improvement to the circuit structure of diodes, transistors, switches, etc.) or a software improvement (improvement to the method flow). However, with the development of technology, many improvements to method flows today can be regarded as direct improvements to the hardware circuit structure. Almost all designers obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement to a method flow cannot be implemented with a hardware entity module. For example, a programmable logic device (PLD) (e.g., a field programmable gate array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. The designer can program by himself to "integrate" a digital system on a piece of PLD without asking the chip manufacturer to design and fabricate a dedicated integrated circuit chip. Moreover, nowadays, instead of manually fabricating integrated circuit chips, this programming is mostly implemented using "logic compiler" software, which is similar to the software compiler used in program development and writing. The original code before compilation also has to be written in a specific programming language, which is called a hardware description language (HDL), and there is not only one kind of HDL, but many 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), etc. The most commonly used ones currently are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also be aware that by simply making a little logical programming of the method flow with the above-mentioned several hardware description languages and programming it into the integrated circuit, it is easy to obtain the hardware circuit that implements the logical method flow.

[0164] The controller can be implemented in any suitable manner. For example, the controller can take the form of, for example, a microprocessor or a processor and a computer-readable medium storing computer-readable program code (such as software or firmware) executable by the (micro)processor, logic gates, switches, an application specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of the controller include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that, in addition to implementing the controller in the form of pure computer-readable program code, it is entirely possible to make the controller implement the same function in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, and embedded microcontrollers by logically programming the method steps. Therefore, such a controller can be considered a hardware component, and the devices included therein for implementing various functions can also be regarded as the structures within the hardware component. Or even, the devices for implementing various functions can be regarded as either software modules for implementing the method or structures within the hardware component.

[0165] The systems, devices, modules, or units illustrated in the above embodiments can be specifically implemented by a computer chip or an entity, or by a product with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, a laptop computer, a cellular phone, 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.

[0166] For the convenience of description, when describing the above devices, they are described separately as various units according to their functions. Of course, when implementing one or more embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0167] Those skilled in the art should understand that the embodiments of this specification can be provided as a method, a system, or a computer program product. Therefore, one or more embodiments of this specification can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, one or more embodiments of this specification can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0168] Embodiments of this specification are described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this specification. It should be understood that each flow and / or block in the flowchart and / or block diagram, and combinations of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable serial and parallel devices for fraud cases to generate a machine, such that the instructions executed by the processors of the computer or other programmable serial and parallel devices for fraud cases generate means for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0169] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable serial and parallel devices for fraud cases to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0170] These computer program instructions can also be loaded onto a computer or other programmable serial and parallel devices for fraud cases, such that a series of operation steps are executed on the computer or other programmable devices to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable devices provide steps for implementing the functions specified in the Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0171] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0172] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0173] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media 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 disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined in this article, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0174] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0175] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, one or more embodiments of this specification may be in the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Furthermore, one or more embodiments of this specification may be in the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0176] One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0177] Each embodiment in this specification is described in a progressive manner. For the identical or similar parts among the embodiments, reference can be made to each other, and the differences between each embodiment and other embodiments are emphasized. In particular, for the system embodiments, since they are basically similar to the method embodiments, the descriptions are relatively simple, and for the relevant parts, reference can be made to the descriptions in the method embodiments.

[0178] The above description is only for the embodiments of this specification and is not intended to limit this document. For those skilled in the art, various modifications and changes can be made to this specification. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this specification shall be included within the scope of the claims of this specification.

Claims

1. A risk identification method, the method comprising: Acquiring business operation data related to business operation behavior of a target merchant in a first business operation scenario, where the first business operation scenario is a scenario that is converted to or from another one or more business operation scenarios, and the business operation data includes historical transaction data of the target merchant, account information of the target merchant, and transaction risk information of the target merchant; Encoding the business operation data to obtain a data representation corresponding to the business operation data; Inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation of each task including the relationship between different tasks among the multiple tasks for the target merchant, the multi-task learning network includes multiple expert models and a gating network corresponding to the tasks, the multiple tasks at least include a risk identification business of the target merchant in a first business operation scenario and a risk identification business of the target merchant converted from the first business operation scenario to a second business operation scenario, the gating network is connected to the output end of each expert model respectively, and the gating network assigns a corresponding weight to the output data of each expert model by determining the expert model whose output data dominates the final prediction; Based on the risk characterization of each task, a risk identification result for the target merchant is determined respectively through a pre-trained risk identification network.

2. According to the method of claim 1, the first business operation scenario is an offline business operation scenario, and the second business operation scenario is an online business operation scenario, or the first business operation scenario is an online business operation scenario, and the second business operation scenario is an offline business operation scenario.

3. The method according to claim 1 or 2, wherein the risk characterization of each task is based on, and the risk identification result for the target merchant is determined by a pre-trained risk identification network, respectively, comprising: Inputting the risk characterization corresponding to the risk identification business of the target merchant in the first business operation scenario into the corresponding risk identification network to obtain the risk occurrence probability corresponding to the risk identification business of the target merchant in the first business operation scenario; Based on the risk characterization corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario, determining the risk occurrence probability corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario through the risk identification network corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario; Based on the risk occurrence probability corresponding to the risk identification business of the target merchant in the first business operation scenario and the risk occurrence probability corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario, the risk identification result for the target merchant is determined.

4. According to the method of claim 3, the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario includes a first sub-business of the business operation scenario conversion and a second sub-business of the target merchant of the converted business operation scenario, the risk characterization corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario includes a first sub-characterization corresponding to the first sub-business and a second sub-characterization corresponding to the second sub-business, and the risk identification network corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario includes a first sub-network corresponding to the first sub-business and a second sub-network corresponding to the second sub-business, The risk characterization corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario is based on the risk characterization corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario, and determining the risk occurrence probability corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario through the risk identification network corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario, including: Inputting a first sub-representation corresponding to the first sub-business into a first network to obtain a conversion probability of the target merchant from the first business operation scenario to the second business operation scenario; Inputting the second sub-representation corresponding to the second sub-business into the second sub-network to obtain the risk occurrence probability corresponding to the risk identification business of the target merchant converted into the second business operation scenario; Based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the second sub-business of the target merchant converted to the second business operation scenario, the risk occurrence probability corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario is determined.

5. The method according to claim 1, wherein the business operation data comprises tabular data, and the data representation corresponding to the business operation data comprises data representation corresponding to a plurality of different tabular features contained in the tabular data, The step of inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation of each task including the relationship between different tasks in the multiple tasks for the target merchant includes: splicing data representations corresponding to a plurality of different table features contained in the business operation data to obtain a spliced ​​data representation; Inputting the concatenated data representations into each expert model in the multi-task learning network respectively to obtain output data of each expert model; The concatenated data representation and the output data of each expert model are input into the gating network corresponding to the task to obtain the weight assigned by the gating network to each expert model, and based on the output data of each expert model and the weight corresponding to each expert model, the risk representation of each task including the relationship between different tasks in the multiple tasks for the target merchant is determined.

6. The method according to claim 5, wherein the number of the gating networks is determined by the number of the plurality of tasks, and each task corresponds to a gating network, and the step of inputting the concatenated data representation and the output data of each expert model into the gating network corresponding to the task to obtain the weight assigned by the gating network to each expert model comprises: The concatenated data representation and the output data of each expert model are respectively input into the gating network corresponding to each task to obtain the weight assigned by each gating network to each expert model.

7. The method according to claim 1, further comprising: Acquire a business operation data sample related to a business operation behavior of a first merchant in a first business operation scenario for model training; Encoding the business operation data sample to obtain a sample data representation corresponding to the business operation data sample; Inputting the sample data representation corresponding to the business operation data sample into the multi-task learning network to obtain a risk representation sample of each task including a relationship between different tasks in a plurality of tasks for the first merchant, wherein the plurality of tasks at least include a risk identification business of a target merchant in a first business operation scenario and a risk identification business of a target merchant converted from the first business operation scenario to the second business operation scenario; Based on the risk characterization samples of each task, determining the risk identification result of the first merchant through the risk identification network respectively; Based on the risk identification result of the first merchant and a preset loss function, the multi-task learning network and the risk identification network are jointly trained to obtain a trained multi-task learning network and a trained risk identification network.

8. According to the method described in claim 7, the loss function includes a first sub-loss function and a second sub-loss function, the first sub-loss function is based on the risk occurrence probability corresponding to the risk identification business of the target merchant in the first business operation scenario and the risk label corresponding to the target merchant in the first business operation scenario, and is constructed through a cross-entropy loss function, the second sub-loss function is based on the conversion probability of the target merchant from the first business operation scenario to the second business operation scenario and the risk occurrence probability corresponding to the risk identification business of the target merchant converted to the second business operation scenario, and the risk label corresponding to the risk identification business of the target merchant converted from the first business operation scenario to the second business operation scenario, and is constructed through a cross-entropy loss function.

9. A risk identification device, comprising: A data acquisition module, which acquires business operation data related to business operation behavior of a target merchant in a first business operation scenario, where the first business operation scenario is a scenario that is converted to or from another one or more business operation scenarios, and the business operation data includes historical transaction data of the target merchant, account information of the target merchant, and transaction risk information of the target merchant; An encoding module, which performs encoding processing on the business operation data to obtain a data representation corresponding to the business operation data; a multi-task processing module, inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network, obtaining a risk representation of each task including the relationship between different tasks among the multiple tasks for the target merchant, wherein the multi-task learning network includes multiple expert models and a gating network corresponding to the tasks, wherein the multiple tasks include at least a risk identification business of the target merchant in a first business operation scenario and a risk identification business of the target merchant converted from the first business operation scenario to a second business operation scenario, wherein the gating network is connected to an output end of each expert model, and the gating network assigns a corresponding weight to the output data of each expert model by determining the expert model whose output data dominates the final prediction; The risk identification module determines the risk identification result for the target merchant through a pre-trained risk identification network based on the risk characterization of each task.

10. A risk identification device, comprising: processor; as well as a memory arranged to store computer executable instructions which, when executed, cause the processor to: Acquire business operation data related to business operation behavior of a target merchant in a first business operation scenario, where the first business operation scenario is a scenario that is converted to or from another one or more business operation scenarios, and the business operation data includes historical transaction data of the target merchant, account information of the target merchant, and transaction risk information of the target merchant; Encoding the business operation data to obtain a data representation corresponding to the business operation data; Inputting the data representation corresponding to the business operation data into a pre-trained multi-task learning network to obtain a risk representation of each task including the relationship between different tasks among the multiple tasks for the target merchant, the multi-task learning network includes multiple expert models and a gating network corresponding to the tasks, the multiple tasks at least include a risk identification business of the target merchant in a first business operation scenario and a risk identification business of the target merchant converted from the first business operation scenario to a second business operation scenario, the gating network is connected to the output end of each expert model respectively, and the gating network assigns a corresponding weight to the output data of each expert model by determining the expert model whose output data dominates the final prediction; Based on the risk characterization of each task, a risk identification result for the target merchant is determined respectively through a pre-trained risk identification network.

Images