Data processing method, device and equipment

Abstract

The embodiments of this specification provide a data processing method, apparatus and equipment, the method comprising: obtaining a target feature vector to be identified, the target feature vector being determined by the target user's interactive content for a target business, wherein the content restored based on the target feature vector is different from the interactive content; identifying the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model being trained based on a feature vector corresponding to a target risk speech, the feature vector corresponding to the target risk speech being obtained by screening a first feature vector corresponding to historical interactive content with risks in the target business based on a pre-trained risk speech screening model; based on the identification result, determining whether there is a risk speech in the interactive content to determine whether there is a risk in triggering the execution of the target business.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of data processing technology, and in particular, to a data processing method, device and equipment. Background Art

[0002] With the rapid development of the Internet industry, network risks have also increased. In risk control scenarios, we can use the pre-built prevention and control vocabulary to detect whether there are words in the interaction content between users that match the risk words in the prevention and control vocabulary to determine whether there are risks in the interaction content.

[0003] However, when the amount of risk word data is large and the update speed is fast, the above method has a high pressure to update the prevention and control word list. It may be impossible to update the prevention and control word list in a timely and accurate manner, resulting in the inability to carry out risk prevention and control. Therefore, a solution is needed that can timely and accurately determine whether there are risk words in the interactive content in a risk control scenario, so as to improve the efficiency and accuracy of risk prevention and control. Summary of the invention

[0004] The purpose of the embodiments of this specification is to provide a solution that can timely and accurately determine whether there are risky words in the interactive content in a risk control scenario, so as to improve the efficiency and accuracy of risk prevention and control.

[0005] In order to implement the above technical solution, the embodiments of this specification are implemented as follows:

[0006] In a first aspect, an embodiment of the present specification provides a data processing method, including: obtaining a target feature vector to be identified, the target feature vector being determined by the target user's interactive content for a target business, wherein the content restored based on the target feature vector is different from the interactive content; performing identification processing on the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model being trained based on a feature vector corresponding to a target risk speech, the feature vector corresponding to the target risk speech being obtained by screening a first feature vector corresponding to historical interactive content with risks in the target business based on a pre-trained risk speech screening model; based on the identification result, determining whether there is a risk speech in the interactive content to determine whether there is a risk in triggering the execution of the target business.

[0007] In a second aspect, an embodiment of the present specification provides a data processing device, the device comprising: a vector acquisition module, used to acquire a target feature vector to be identified, the target feature vector is determined by the target user's interactive content for a target business, wherein the content restored based on the target feature vector is different from the interactive content; a vector recognition module, used to identify the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, the feature vector corresponding to the target risk speech is obtained by screening a first feature vector corresponding to historical interactive content with risks in the target business based on a pre-trained risk speech screening model; a risk detection module, used to determine whether there is a risk speech in the interactive content based on the recognition result, so as to determine whether there is a risk in triggering the execution of the target business.

[0008] In a third aspect, an embodiment of the present specification provides a data processing device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein the executable instructions, when executed, cause the processor to: obtain a target feature vector to be identified, wherein the target feature vector is determined by the target user's interactive content for a target business, wherein the content restored based on the target feature vector is different from the interactive content; perform identification processing on the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, wherein the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, wherein the feature vector corresponding to the target risk speech is obtained by screening a first feature vector corresponding to historical interactive content with risks in the target business based on a pre-trained risk speech screening model; based on the identification result, determine whether there is a risk speech in the interactive content to determine whether there is a risk in triggering the execution of the target business.

[0009] In a fourth aspect, an embodiment of the present specification provides a storage medium, which is used to store computer-executable instructions, and the executable instructions implement the following process when executed: obtaining a target feature vector to be identified, the target feature vector is determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content; identifying the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening a first feature vector corresponding to historical interactive content with risks in the target business based on a pre-trained risk speech screening model; based on the identification result, determining whether there is a risk speech in the interactive content to determine whether there is a risk in triggering the execution of the target business. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0011] Figure 1A A flowchart of a data processing method embodiment of this specification;

[0012] Figure 1B A schematic diagram of a processing process of a data processing method embodiment of this specification;

[0013] Figure 2 A schematic diagram of a processing process of another data processing method embodiment of the present specification;

[0014] Figure 3 This is a schematic diagram of a risk speech screening model for this manual;

[0015] Figure 4 This is a schematic diagram of a risk speech screening model for this manual;

[0016] Figure 5A A flowchart of another data processing method embodiment of the present specification;

[0017] Figure 5B A schematic diagram of a processing process of another data processing method embodiment of the present specification;

[0018] Figure 6 A schematic diagram of the structure of a data processing device according to an embodiment of the present specification;

[0019] Figure 7 This is a structural schematic diagram of a data processing device in this specification. DETAILED DESCRIPTION

[0020] The embodiments of this specification provide a data processing method, device and equipment.

[0021] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0022] Embodiment 1

[0023] like Figure 1A and Figure 1B As shown, an embodiment of the present specification provides a data processing method, and the execution subject of the method can be a server, and the server can be an independent server or a server cluster composed of multiple servers.

[0024] The method may specifically include the following steps:

[0025] In S102, a target feature vector to be identified is obtained.

[0026] Among them, the target feature vector can be determined by the target user's interaction content for the target business. The target business can be any business involving user privacy, property safety, etc. For example, the target business can be a resource transfer business, an instant messaging business, etc. The target user's interaction content for the target business can be the interaction content between the target user and the resource receiving user for the resource transfer business, etc. The target feature vector can be a vector corresponding to the interaction content generated based on a preset vector generation model, and the content restored based on the target feature vector is different from the interaction content, that is, the target feature preset vector is restored, and the target user's interaction content for the target business cannot be obtained.

[0027] In practice, with the rapid development of the Internet industry, network risks have also increased. In risk control scenarios, we can use the pre-built prevention and control vocabulary to detect whether there are words in the interaction content between users that match the risk words in the prevention and control vocabulary to determine whether there are risks in the interaction content.

[0028] However, when the amount of risk word data is large and the update speed is fast, the above method has a high pressure to update the prevention and control word list. It may be impossible to update the prevention and control word list in a timely and accurate manner, resulting in the inability to perform risk prevention and control. Therefore, a solution is needed that can timely and accurately determine whether there are risk words in the interactive content in a risk control scenario to improve the efficiency and accuracy of risk prevention and control. To this end, the embodiments of this specification provide a technical solution that can solve the above problem, and the details can be found in the following content.

[0029] Taking the target service as a resource transfer service in a resource management application installed in a terminal device (i.e., a terminal device or a server) as an example, the target user can trigger the start of the resource management application and trigger the execution of the resource transfer service in the resource management application. The target user can interact with user 1 for the resource transfer service, and the terminal device can obtain the interaction content between the target user and user 1, and generate a target feature vector corresponding to the interaction content.

[0030] In addition, the terminal device may also send the generated target feature vector corresponding to the interactive content to the server, that is, the server may obtain the target feature vector sent by the terminal device.

[0031] Since the content restored by the target feature vector is different from the interactive content, the privacy of the target user can be protected.

[0032] In addition, to improve data processing efficiency, after obtaining the target user's interactive content for the target service, the terminal device can generate a target feature vector based on the content in the interactive content other than the content corresponding to the target user. For example, taking the above target service as the resource transfer service as an example, when the target user triggers the resource transfer service, he can interact with user 1, and the interactive content obtained by the terminal device can be as follows:

[0033] User 1: Please transfer the money before 12:00 today.

[0034] Target user: OK, is 12:10 ok?

[0035] User 1: Yes.

[0036] Target user: OK, is this to transfer to account 1?

[0037] User 1: No, please go to Account 2.

[0038] The terminal device can generate a target feature vector based on the content corresponding to user 1 in the above interactive content, and send the target feature vector to the server for risk speech recognition processing.

[0039] Alternatively, the terminal device may also match the target user's interactive content for the target business based on preset risk keywords, and if the match is successful, generate a target feature vector based on the interactive content that is associated with the content matching the risk keyword.

[0040] For example, assuming that the risk keyword includes "transfer", the terminal device can generate a target feature vector based on the interaction content containing "transfer" in the above interaction content.

[0041] In S104, the target feature vector is identified based on the pre-trained risk speech recognition model to obtain a recognition result for the target feature vector.

[0042] Among them, the risk speech recognition model can be trained based on the feature vector corresponding to the target risk speech. The feature vector corresponding to the target risk speech can be a risk speech screening model based on pre-trained risk speech, which is obtained by screening the first feature vector corresponding to the historical interactive content with risks in the target business. The content restored based on the first feature vector is different from the interactive content.

[0043] During implementation, based on the business identifier of the target business, a first feature vector corresponding to the historical interaction content that is determined to have risks when the target business is triggered can be obtained, and then the first feature vector can be screened based on a pre-trained risk speech screening model to obtain a feature vector corresponding to the target risk speech, and a risk speech recognition model can be trained based on the feature vector corresponding to the target risk speech to obtain a trained risk speech recognition model.

[0044] The target feature vector can be input into a pre-trained risk speech recognition model to obtain a recognition result for the target feature vector. In this way, the risk speech recognition model can be used to identify whether there is risk speech in the target user's interactive content for the target business, and risk detection can be located at the speech level.

[0045] In S106, based on the recognition result, it is determined whether there is risky speech in the interactive content to determine whether there is a risk in triggering the execution of the target business.

[0046] In implementation, when it is determined based on the recognition result that risky language exists in the interactive content, a preset prompt message can be sent to the target user to trigger the target device to execute the target business, so as to prompt the target user that risky language exists in the interactive content.

[0047] In addition, since risky speech is updated quickly, in order to improve the accuracy of model training, when it is determined that there is risky speech in the interactive content based on the recognition results, the interactive content can be determined as historical interactive content, that is, the target feature vector can be used as the first feature vector corresponding to the historical interactive content with risks in the target business, and when the number of target feature vectors reaches a preset incremental threshold, the risky speech screening model is updated.

[0048] If the target feature vector is a feature vector generated by the terminal device based on the target user's interactive content for the target business, then when the server determines, based on the recognition result, that there is risky language in the interactive content, it can send a preset prompt message to the terminal device to remind the target user and prevent the target user from triggering the execution of the target business and causing security losses.

[0049] The embodiment of the present specification provides a data processing method, which obtains a target feature vector to be identified, and the target feature vector can be determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, and the risk speech recognition model is obtained by training the feature vector corresponding to the target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing the first feature vector corresponding to the historical interactive content with risks in the target business based on the pre-trained risk speech screening model, and based on the identification result, it is determined whether there is a risk speech in the interactive content, so as to determine whether there is a risk in triggering the execution of the target business. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through the pre-trained risk speech recognition model, the identification result for the target feature vector can be obtained, so as to determine whether there is a risk speech in the interactive content based on the identification result, so as to improve the efficiency and accuracy of determining the risk speech, that is, to improve the efficiency and accuracy of risk prevention and control.

[0050] Embodiment 2

[0051] like Figure 2 As shown, the embodiment of this specification provides a data processing method, the execution subject of the method can be a server, the server can be an independent server, or a server cluster composed of multiple servers. The method can specifically include the following steps:

[0052] In S202, the interaction content of the target user for the target service is obtained.

[0053] In implementation, the terminal device can desensitize the target user's interactive content for the target business based on a preset desensitizing algorithm, and send the desensitized interactive content to the server, that is, the interactive content received by the server does not contain the target user's privacy data.

[0054] Alternatively, upon receiving an authorization instruction from the target user (ie, the target user authorizes the server to process the interaction content), the terminal device may send the interaction content of the target user for the target service to the server for processing.

[0055] In S204, the interactive content is divided into a plurality of sub-contents, and based on a preset encoding rule, the sub-contents are respectively encoded to obtain a plurality of target feature vectors to be identified.

[0056] Among them, the content restored based on the target feature vector is different from the interactive content.

[0057] In implementation, for example, the content input by the user each time in the interactive content can be used as a sub-content. Specifically, assuming that the interactive content obtained is:

[0058] User 1: Please transfer the money before 12:00 today.

[0059] Target user: OK, is 12:10 ok?

[0060] User 1: Yes.

[0061] Target user: OK, is this to transfer to account 1?

[0062] User 1: No, please go to Account 2.

[0063] If the content input by the user each time in the interactive content is regarded as a sub-content, the above interactive content can be divided into 5 sub-contents, that is, the sub-contents obtained by division can include "Please transfer before 12:00 today", "OK, is 12:10 ok" as one-word content, etc.

[0064] Alternatively, the interactive content may be divided into multiple sub-contents based on preset risk keywords (including preset risk character types, such as numerical type and time type). For example, assuming that the risk keyword includes "account" and time type, the interactive content may be divided into sub-content 1 corresponding to "account" and sub-content 2 corresponding to time type, wherein sub-content 1 may be: "OK, is it transferred to account 1? No, please transfer to account 2", and sub-content 2 may be: "Please transfer before 12:00 today. OK, is 12:10 ok?"

[0065] The above-mentioned method for determining sub-content is an optional and feasible division method. In actual application scenarios, there may be a variety of different division methods, which may vary according to different actual application scenarios. The embodiments of this specification do not make specific limitations on this.

[0066] After obtaining the sub-content, the sub-content can be encoded separately based on the preset encoding rules to obtain the target feature vector corresponding to each sub-content, wherein the preset encoding rule can be any irreversible encoding rule, that is, the content restored based on the target feature vector is different from the interactive content, so as to protect the privacy data of the target user.

[0067] In S206, a historical feature vector is obtained.

[0068] The historical feature vector is determined by the first interactive content, and the content restored based on the historical feature vector is different from the first interactive content. The first interactive content may be risky interactive content.

[0069] In implementation, each first interactive content may be divided into a plurality of first sub-contents based on the division method in S204, and each first sub-content may be encoded based on a preset encoding rule to obtain a plurality of historical feature vectors.

[0070] In S208, the historical feature vectors are input into the risk speech screening model to obtain the target probability vector and attention level of each historical feature vector.

[0071] Among them, the risky speech screening model can be built based on the multiple instance learning algorithm (Multiple Instance Learning, MIL). In this way, the risky speech screening model built based on MIL can be used to realize risk detection at the speech level of the interactive content.

[0072] In implementation, in actual applications, the processing method of the above S208 can be various. An optional implementation method is provided below. For details, please refer to steps 1 to 2.

[0073] Step 1: Input the historical feature vectors into the neural network layer of the risk speech screening model to obtain the target probability vector for each historical feature vector.

[0074] Among them, the neural network layer can be constructed by algorithms such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory Artificial Neural Networks (LSTM), etc., which are used to determine the probability vector of the vector.

[0075] Step 2: Determine the attention level of each historical feature vector based on the preset attention mechanism and the target probability vector of each historical feature vector.

[0076] In implementation, Figure 3 As shown in the figure, taking the neural network layer constructed by bidirectional LSTM as an example, after each historical feature vector (such as vector s1, vector s2, vector s3) passes through the neural network layer and the activation function (such as softmax function), the target probability vector (i.e., p1, p2, p3) of each historical feature vector can be obtained. Based on the attention mechanism, the degree of attention between the historical feature vectors (i.e., β1, β2, β3, ) is obtained. Among them, are sequences corresponding to vector s1, vector s2, and vector 3, respectively, and with the same direction. Sequences with the same direction as vector s1, vector s2, and vector 3 objects, respectively, where, and The larger the target probability vector is, the greater the risk that the first interaction content corresponding to the historical feature vector contains risky words.

[0077] In S210, based on the target probability vector and attention level of each historical feature vector and the risk label of the historical interaction content, it is determined whether to retrain the risk speech screening model, and the risk speech screening model when the training is stopped is obtained.

[0078] In implementation, in actual applications, the processing method of the above S212 can be various. An optional implementation method is provided below. For details, please refer to steps 1 to 2.

[0079] Step 1: Determine the target score of each first interactive content based on the target probability vector and the attention level of the historical feature vector corresponding to each first interactive content.

[0080] In implementation, one or more historical feature vectors corresponding to each first interaction may be obtained, and the sum of the products of the target probability vector and the attention level of the historical feature vectors corresponding to each first interaction may be determined as the target score for each first interaction content.

[0081] There may be multiple methods for determining the target score of each of the first interactive contents. Different determination methods may be selected according to different actual application scenarios, and this specification does not specifically limit this.

[0082] Step 2: Based on the risk label and target score of each first interaction content, determine whether to retrain the risk speech screening model.

[0083] In implementation, the predicted type of each first interactive content (such as risk exists, risk does not exist, other, etc.) can be determined based on the target score and the preset risk score threshold, and then based on the predicted type and risk label of each first interactive content, it can be determined whether to retrain the risk speech screening model.

[0084] In S212, a first feature vector corresponding to historical interaction content with risks in the target business is obtained.

[0085] Among them, the content restored based on the first eigenvector is different from the historical interaction content with risks.

[0086] In implementation, as with the division method in S204 above, each historical interaction content may be divided into a plurality of historical sub-contents, and each historical sub-content may be encoded based on a preset encoding rule to obtain a plurality of first feature vectors.

[0087] In S214, the first feature vector is screened based on the pre-trained risk speech screening model to obtain a third feature vector, and the feature vector corresponding to the target risk speech is determined based on the third feature vector.

[0088] In implementation, in actual applications, the processing method of the above S214 can be various. An optional implementation method is provided below. For details, please refer to steps 1 to 2.

[0089] Step 1: Input the first feature vector into the pre-trained risk speech screening model to obtain the target probability vector and attention level of each first feature vector.

[0090] Step 2: Determine the third eigenvector based on the target probability vector and the attention level of each first eigenvector.

[0091] In implementation, the risk score of each first feature vector may be determined based on the target probability vector and the attention level of each first feature vector, and the third feature vector may be determined based on the risk score of each first feature vector.

[0092] For example, the product of the target probability vector and the attention level of each first eigenvector can be determined as the risk score of each first eigenvector, the first eigenvectors are sorted based on the risk scores, and the third eigenvector is determined based on the sorted first eigenvectors. For example, the first n first eigenvectors with larger risk scores can be determined as the third eigenvector.

[0093] There may be multiple methods for determining each third eigenvector mentioned above, and different determination methods may be selected according to different actual application scenarios, which is not specifically limited in the embodiments of this specification.

[0094] In S216, the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech to obtain a trained risk speech recognition model.

[0095] In implementation, in actual applications, the processing method of the above S216 can be various. An optional implementation method is provided below, and the details can be referred to steps A1 to A2.

[0096] In A1, a predetermined number of third feature vectors are randomly selected from the historical feature vectors.

[0097] In implementation, the third feature vector may be a feature vector of a risk-free label.

[0098] In A2, based on the feature vector corresponding to the target risk speech and the third feature vector, the risk speech recognition model is trained in an unsupervised training manner to obtain a pre-trained risk speech recognition model.

[0099] In implementation, since risk speech is updated quickly, there may be a problem of small amount of sample data with risk labels. Therefore, a risk speech recognition model can be constructed based on the positive unlabeled learning (PUlearning) algorithm, and the risk speech recognition model can be trained in an unsupervised training manner.

[0100] In addition, the risk speech recognition model can also be a model constructed based on a preset classification algorithm, and the risk speech recognition model can be trained based on a supervised training method. An optional implementation method is provided below, and please refer to steps B1 to B2 for details.

[0101] In B1, a predetermined number of fifth eigenvectors are randomly selected from the fourth eigenvectors of the historical eigenvectors.

[0102] The first interactive content corresponding to the fourth eigenvector may be interactive content without risk.

[0103] In B2, based on the feature vector corresponding to the target risk speech and the fifth feature vector, the risk speech recognition model is trained to obtain a pre-trained risk speech recognition model.

[0104] During implementation, the fifth eigenvector can be used as high-confidence white sample data, and the eigenvector corresponding to the target risk speech can be used as black sample data to train the risk speech recognition model to obtain a pre-trained risk speech recognition model, so that the trained risk speech recognition model can hit the interaction content between malicious third parties and non-risk users as much as possible, and hit the interaction content between non-risk users as little as possible.

[0105] In S104, the target feature vector is identified based on the pre-trained risk speech recognition model to obtain a recognition result for the target feature vector.

[0106] In S218, if it is determined based on the recognition result that risky speech exists in the interactive content, a second feature vector corresponding to the risky speech in the target feature vector is determined.

[0107] In the implementation, it is assumed that the target user's interaction content for the target business is:

[0108] User 1: Please transfer the money before 12:00 today.

[0109] Target user: OK, is 12:10 ok?

[0110] User 1: Yes.

[0111] Target user: OK, is this to transfer to account 1?

[0112] User 1: No, please go to Account 2.

[0113] The target feature vector includes target feature vector 1 corresponding to sub-content 1 and target feature vector 2 corresponding to sub-content 2, wherein sub-content 1 may be: "OK, is it transferred to account 1? No, please transfer to account 2", and sub-content 2 may be: "Please transfer before 12:00 today. OK, is 12:10 okay?"

[0114] If, based on the recognition result, it is determined that the target feature vector 1 is a feature vector corresponding to the risky speech, then the target feature vector 1 may be determined as the second feature vector.

[0115] There may be multiple methods for determining the second eigenvector, and different determination methods may be selected according to different actual application scenarios, which is not specifically limited in the embodiments of this specification.

[0116] In S220, the target business is stopped, and risk warning information corresponding to the second feature vector is generated to warn the target user that there are risky words in the interactive content based on the risk warning information.

[0117] During implementation, the server can generate risk warning information corresponding to the second feature vector. For example, the server can generate risk warning information carrying the position information of the second feature vector in the interactive content corresponding to the target feature vector, and send the risk warning information to the terminal device. The terminal device determines the content containing risky words in the interactive content based on the position information in the risk warning information, and merges the content with the risk warning information for output to the target user.

[0118] For example, taking the interactive content in S218 above as an example, assuming that the target feature vector 2 corresponding to the sub-content 2 is the second feature vector, the risk warning information generated by the server may be "This may contain risky words, please pay attention", and the terminal device may output the following information based on the position information of the content corresponding to the target feature vector 2 in the interactive content: Figure 4 The risk warning information shown.

[0119] The embodiment of the present specification provides a data processing method, which obtains a target feature vector to be identified, and the target feature vector can be determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, and the risk speech recognition model is obtained by training the feature vector corresponding to the target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing the first feature vector corresponding to the historical interactive content with risks in the target business based on the pre-trained risk speech screening model, and based on the identification result, it is determined whether there is a risk speech in the interactive content, so as to determine whether there is a risk in triggering the execution of the target business. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through the pre-trained risk speech recognition model, the identification result for the target feature vector can be obtained, so as to determine whether there is a risk speech in the interactive content based on the identification result, so as to improve the efficiency and accuracy of determining the risk speech, that is, to improve the efficiency and accuracy of risk prevention and control.

[0120] Embodiment 3

[0121] like Figure 5A and Figure 5B As shown, the embodiment of this specification provides a data processing method, and the execution subject of the method can be a terminal device, which can be a mobile terminal device such as a mobile phone, a tablet computer, or a terminal device such as a personal computer. The method can specifically include the following steps:

[0122] In S502, a target feature vector corresponding to the interactive content is determined.

[0123] Among them, the content restored based on the target feature vector is different from the interactive content.

[0124] In implementation, the terminal device may divide the interactive content into a plurality of sub-contents, and based on a preset encoding rule, encode the sub-contents respectively to obtain a plurality of target feature vectors to be identified.

[0125] In S504, the target feature vector is sent to the server, so that the server performs recognition processing on the target feature vector based on a pre-trained risk speech recognition model to obtain a recognition result for the target feature vector.

[0126] Among them, the risk speech recognition model can be trained based on the feature vector corresponding to the target risk speech. The feature vector corresponding to the target risk speech is obtained by screening the first feature vector corresponding to the interactive content with risks in the target business based on the pre-trained risk speech screening model.

[0127] In implementation, the terminal device can send the target feature vector to the cloud server, so that the server can identify the target feature vector based on the pre-trained risk speech recognition model to obtain the recognition result for the target feature vector.

[0128] In S506, based on the risk identification result sent by the server, it is determined whether there is any risky speech in the interactive content.

[0129] During implementation, when the terminal device determines whether there is risky language in the interactive content based on the risk identification results sent by the server, it can output preset prompt information to remind the target user that there is risk in executing the target business. Alternatively, the terminal device can also output corresponding prompt information for the content in the interactive content corresponding to the risky language based on the risk identification results, so as to output prompt information to the target user in a targeted manner.

[0130] The embodiments of this specification provide a data processing method. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through a pre-trained risk speech recognition model, a recognition result for the target feature vector can be obtained. Based on the recognition result, it is determined whether there is risk speech in the interactive content, which can improve the efficiency and accuracy of determining risk speech, that is, the efficiency and accuracy of risk prevention and control can be improved.

[0131] Embodiment 4

[0132] The above is a data processing method provided in the embodiment of this specification. Based on the same idea, the embodiment of this specification also provides a data processing device, such as Figure 6 shown.

[0133] The data processing device includes: a vector acquisition module 601, a vector identification module 602 and a risk detection module 603, wherein:

[0134] A vector acquisition module 601 is used to acquire a target feature vector to be identified, wherein the target feature vector is determined by the interaction content of the target user for the target service, wherein the content restored based on the target feature vector is different from the interaction content;

[0135] A vector recognition module 602 is used to perform recognition processing on the target feature vector based on a pre-trained risk speech recognition model to obtain a recognition result for the target feature vector, wherein the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening a first feature vector corresponding to historical interaction content with risks in the target business based on a pre-trained risk speech screening model;

[0136] The risk detection module 603 is used to determine whether there is risky speech in the interactive content based on the identification result, so as to determine whether there is risk in triggering the execution of the target business.

[0137] In the embodiment of this specification, the risk detection module 603 is used to:

[0138] If it is determined based on the recognition result that risky speech exists in the interactive content, a second feature vector corresponding to the risky speech in the target feature vector is determined;

[0139] The target business is stopped, and risk warning information corresponding to the second feature vector is generated, so as to warn the target user of the presence of risky language in the interactive content based on the risk warning information.

[0140] In the embodiment of this specification, the device further includes:

[0141] A vector extraction module, configured to obtain a first feature vector corresponding to historical interaction content with risks in the target business, wherein content restored based on the first feature vector is different from the historical interaction content with risks;

[0142] A vector screening module, configured to screen the first feature vector based on a pre-trained risk speech screening model to obtain a third feature vector, and determine a feature vector corresponding to the target risk speech based on the third feature vector;

[0143] The model training module is used to train the risk speech recognition model based on the feature vector corresponding to the target risk speech to obtain the trained risk speech recognition model.

[0144] In the embodiment of this specification, the risk speech screening model is constructed based on a multi-instance learning algorithm, and the device also includes:

[0145] A first acquisition module, configured to acquire a historical feature vector, wherein the historical feature vector is determined by a first interactive content, wherein content restored based on the historical feature vector is different from the first interactive content;

[0146] A data acquisition module, used for inputting the historical feature vectors into the risk speech screening model to obtain a target probability vector and a degree of attention for each of the historical feature vectors;

[0147] A model acquisition module is used to determine whether to retrain the risk speech screening model based on the target probability vector and attention level of each of the historical feature vectors and the risk label of the historical interaction content, and to obtain the risk speech screening model when training is stopped.

[0148] In the embodiment of this specification, the data acquisition module is used to:

[0149] Inputting the historical feature vectors into the neural network layer of the risk speech screening model to obtain the target probability vector of each of the historical feature vectors;

[0150] Based on a preset attention mechanism and a target probability vector of each of the historical feature vectors, the attention degree of each of the historical feature vectors is determined.

[0151] In the embodiment of this specification, the model acquisition module is used to:

[0152] Determining a target score for each of the first interactive contents based on a target probability vector and an attention level of a historical feature vector corresponding to each of the first interactive contents;

[0153] Based on the risk label and target score of each of the first interactive contents, determine whether to retrain the risk speech screening model.

[0154] In the embodiment of this specification, the vector screening module is used to:

[0155] Inputting the first feature vector into the pre-trained risk speech screening model to obtain a target probability vector and a degree of attention for each of the first feature vectors;

[0156] The third feature vector is determined based on the target probability vector and the attention degree of each of the first feature vectors.

[0157] In the embodiment of this specification, the risk detection module 603 is used to:

[0158] Based on the target probability vector and the attention level of each of the first feature vectors, a risk score of each of the first feature vectors is determined, and based on the risk score of each of the first feature vectors, the third feature vector is determined.

[0159] In the embodiments of this specification, the model training module is used to:

[0160] randomly selecting a predetermined number of third feature vectors from the historical feature vectors;

[0161] Based on the feature vector corresponding to the target risk speech and the third feature vector, the risk speech recognition model is trained in an unsupervised training manner to obtain the pre-trained risk speech recognition model.

[0162] In the embodiments of this specification, the model training module is used to:

[0163] Randomly selecting a predetermined number of fifth feature vectors from the fourth feature vectors of the historical feature vectors, the first interaction content corresponding to the fourth feature vector being interaction content without risk;

[0164] Based on the feature vector corresponding to the target risk speech and the fifth feature vector, the risk speech recognition model is trained to obtain the pre-trained risk speech recognition model.

[0165] In the embodiment of this specification, the vector acquisition module 601 is used to:

[0166] Acquiring the target user's interactive content for the target service;

[0167] The interactive content is divided into a plurality of sub-contents, and based on a preset encoding rule, the sub-contents are respectively encoded to obtain a plurality of target feature vectors to be identified.

[0168] The embodiment of the present specification provides a data processing device, which obtains a target feature vector to be identified, and the target feature vector can be determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, and the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing the first feature vector corresponding to the historical interactive content with risks in the target business based on the pre-trained risk speech screening model, and based on the identification result, it is determined whether there is a risk speech in the interactive content, so as to determine whether there is a risk in triggering the execution of the target business. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through the pre-trained risk speech recognition model, the identification result for the target feature vector can be obtained, so as to determine whether there is a risk speech in the interactive content based on the identification result, so as to improve the determination efficiency and accuracy of the risk speech, that is, to improve the efficiency and accuracy of risk prevention and control.

[0169] Embodiment 5

[0170] Based on the same idea, the embodiment of this specification also provides a data processing device, such as Figure 7 shown.

[0171] The data processing device may have relatively large differences due to different configurations or performances, and may include one or more processors 701 and memory 702, and the memory 702 may store one or more storage applications or data. Among them, the memory 702 may be a short-term storage or a persistent storage. The application stored in the memory 702 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the data processing device. Furthermore, the processor 701 may be configured to communicate with the memory 702 and execute a series of computer executable instructions in the memory 702 on the data processing device. The data processing device may also include one or more power supplies 703, one or more wired or wireless network interfaces 704, one or more input and output interfaces 705, and one or more keyboards 706.

[0172] Specifically in this embodiment, the data processing device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions in the data processing device, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following:

[0173] Acquire a target feature vector to be identified, where the target feature vector is determined by the interaction content of the target user for the target service, wherein the content restored based on the target feature vector is different from the interaction content;

[0174] The target feature vector is identified based on a pre-trained risk speech recognition model to obtain a recognition result for the target feature vector, wherein the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening a first feature vector corresponding to historical interaction content with risks in the target business based on a pre-trained risk speech screening model;

[0175] Based on the identification result, determine whether there is risky language in the interactive content to determine whether there is a risk in triggering the execution of the target business.

[0176] Optionally, the determining whether there is risky speech in the interactive content based on the recognition result to determine whether there is risk in triggering the execution of the target business includes:

[0177] If it is determined based on the recognition result that risky speech exists in the interactive content, a second feature vector corresponding to the risky speech in the target feature vector is determined;

[0178] The target business is stopped, and risk warning information corresponding to the second feature vector is generated, so as to warn the target user of the presence of risky language in the interactive content based on the risk warning information.

[0179] Optionally, before the target feature vector is identified based on the pre-trained risk speech identification model to obtain an identification result for the target feature vector, the method further includes:

[0180] Acquire a first feature vector corresponding to historical interaction content with risks in the target business, wherein content restored based on the first feature vector is different from the historical interaction content with risks;

[0181] The first feature vector is screened based on a pre-trained risk speech screening model to obtain a third feature vector, and a feature vector corresponding to the target risk speech is determined based on the third feature vector;

[0182] The risk speech recognition model is trained based on the feature vector corresponding to the target risk speech to obtain a trained risk speech recognition model.

[0183] Optionally, the risk speech screening model is constructed based on a multi-instance learning algorithm, and before the risk speech screening model based on pre-training screens the first feature vector to obtain a third feature vector, it further includes:

[0184] Acquire a historical feature vector, where the historical feature vector is determined by the first interactive content, wherein the content restored based on the historical feature vector is different from the first interactive content;

[0185] Inputting the historical feature vectors into the risk speech screening model to obtain a target probability vector and a degree of attention for each of the historical feature vectors;

[0186] Based on the target probability vector and attention level of each of the historical feature vectors and the risk label of the historical interaction content, determine whether to retrain the risky speech screening model and obtain the risky speech screening model when the training is stopped.

[0187] Optionally, the step of inputting the historical feature vectors into the risk speech screening model to obtain a target probability vector and a degree of attention for each of the historical feature vectors includes:

[0188] Inputting the historical feature vectors into the neural network layer of the risk speech screening model to obtain the target probability vector of each of the historical feature vectors;

[0189] Based on a preset attention mechanism and a target probability vector of each of the historical feature vectors, the attention degree of each of the historical feature vectors is determined.

[0190] Optionally, determining whether to retrain the risk speech screening model based on the target probability vector and the attention level of each of the historical feature vectors and the risk label of the historical interaction content includes:

[0191] Determining a target score for each of the first interactive contents based on a target probability vector and an attention level of a historical feature vector corresponding to each of the first interactive contents;

[0192] Based on the risk label and target score of each of the first interactive contents, determine whether to retrain the risk speech screening model.

[0193] Optionally, the pre-trained risk speech screening model is used to screen the first feature vector to obtain a third feature vector, including:

[0194] Inputting the first feature vector into the pre-trained risk speech screening model to obtain a target probability vector and a degree of attention for each of the first feature vectors;

[0195] The third feature vector is determined based on the target probability vector and the attention degree of each of the first feature vectors.

[0196] Optionally, determining the third feature vector based on the target probability vector and the attention degree of each of the first feature vectors includes:

[0197] Based on the target probability vector and the attention level of each of the first feature vectors, a risk score of each of the first feature vectors is determined, and based on the risk score of each of the first feature vectors, the third feature vector is determined.

[0198] Optionally, the step of training a risk speech recognition model based on the feature vector corresponding to the target risk speech to obtain a pre-trained risk speech recognition model includes:

[0199] randomly selecting a predetermined number of third feature vectors from the historical feature vectors;

[0200] Based on the feature vector corresponding to the target risk speech and the third feature vector, the risk speech recognition model is trained in an unsupervised training manner to obtain the pre-trained risk speech recognition model.

[0201] Optionally, the step of training a risk speech recognition model based on the feature vector corresponding to the target risk speech to obtain a pre-trained risk speech recognition model includes:

[0202] Randomly selecting a predetermined number of fifth feature vectors from the fourth feature vectors of the historical feature vectors, the first interaction content corresponding to the fourth feature vector being interaction content without risk;

[0203] Based on the feature vector corresponding to the target risk speech and the fifth feature vector, the risk speech recognition model is trained to obtain the pre-trained risk speech recognition model.

[0204] Optionally, obtaining the target feature vector to be identified includes:

[0205] Acquiring the target user's interactive content for the target service;

[0206] The interactive content is divided into a plurality of sub-contents, and based on a preset encoding rule, the sub-contents are respectively encoded to obtain a plurality of target feature vectors to be identified.

[0207] The embodiment of the present specification provides a data processing device, which obtains a target feature vector to be identified, and the target feature vector can be determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, and the risk speech recognition model is obtained by training the feature vector corresponding to the target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing the first feature vector corresponding to the historical interactive content with risks in the target business based on the pre-trained risk speech screening model, and based on the identification result, it is determined whether there is a risk speech in the interactive content, so as to determine whether there is a risk in triggering the execution of the target business. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through the pre-trained risk speech recognition model, the identification result for the target feature vector can be obtained, so as to determine whether there is a risk speech in the interactive content based on the identification result, so as to improve the determination efficiency and accuracy of the risk speech, that is, to improve the efficiency and accuracy of risk prevention and control.

[0208] Embodiment 6

[0209] The embodiments of this specification also provide a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above-mentioned data processing method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. The computer-readable storage medium is, for example, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0210] The embodiment of the present specification provides a computer-readable storage medium, which obtains a target feature vector to be identified, and the target feature vector can be determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, and the risk speech recognition model is obtained by training the feature vector corresponding to the target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing the first feature vector corresponding to the historical interactive content with risks in the target business based on the pre-trained risk speech screening model, and based on the identification result, it is determined whether there is a risk speech in the interactive content, so as to determine whether there is a risk in triggering the execution of the target business. In this way, since the content restored by the target feature vector is different from the interactive content, the security of the target user's privacy information can be guaranteed. In addition, through the pre-trained risk speech recognition model, the identification result for the target feature vector can be obtained, so as to determine whether there is a risk speech in the interactive content based on the identification result, so as to improve the efficiency and accuracy of determining the risk speech, that is, to improve the efficiency and accuracy of risk prevention and control.

[0211] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0212] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one 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 are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0213] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC625D, 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 a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0214] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may 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.

[0215] For the convenience of description, the above devices are described in terms of functions and are divided into various units. 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.

[0216] 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 take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, one or more embodiments of this specification may take 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.) that contain computer-usable program code.

[0217] The embodiments of this specification are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one process or multiple processes in the flowchart and / or one box or multiple boxes in the block diagram.

[0218] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0219] These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0220] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0221] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0222] 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 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 herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0223] 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.

[0224] It should be understood by those skilled in the art 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 take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, one or more embodiments of this specification may take 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 codes.

[0225] 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.

[0226] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0227] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A data processing method, comprising: Acquire a target feature vector to be identified, the target feature vector is determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is generated by the terminal device based on the content in the interactive content other than the content corresponding to the target user, or the terminal device performs matching processing on the target user's interactive content for the target business based on preset risk keywords, and generates a feature vector based on the interactive content that has an associated relationship with the content matching the risk keyword if the match is successful; The target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech, the feature vector corresponding to the target risk speech is obtained by screening the first feature vector corresponding to the historical interaction content with risks in the target business based on a pre-trained risk speech screening model, the risk speech screening model is a model trained based on the target probability vector and the attention degree of multiple historical feature vectors and the risk label of the historical interaction content, the historical feature vector is determined by the first interaction content, and the content restored based on the historical feature vector is different from the first interaction content; Based on the identification result, determine whether there is risky language in the interactive content to determine whether there is a risk in triggering the execution of the target business.

2. The method according to claim 1, wherein determining whether there is risky speech in the interactive content based on the recognition result to determine whether there is risk in triggering the execution of the target business comprises: If it is determined based on the recognition result that risky speech exists in the interactive content, a second feature vector corresponding to the risky speech in the target feature vector is determined; The target business is stopped, and risk warning information corresponding to the second feature vector is generated, so as to warn the target user of the presence of risky language in the interactive content based on the risk warning information.

3. The method according to claim 2, before the target feature vector is identified based on the pre-trained risk speech recognition model to obtain the recognition result for the target feature vector, further comprising: Acquire a first feature vector corresponding to historical interaction content with risks in the target business, wherein content restored based on the first feature vector is different from the historical interaction content with risks; The first feature vector is screened based on a pre-trained risk speech screening model to obtain a third feature vector, and a feature vector corresponding to the target risk speech is determined based on the third feature vector; The risk speech recognition model is trained based on the feature vector corresponding to the target risk speech to obtain a trained risk speech recognition model.

4. The method according to claim 3, wherein the risk speech screening model is constructed based on a multi-instance learning algorithm, and before the risk speech screening model based on pre-trained screening processes the first feature vector to obtain the third feature vector, the method further comprises: Acquire a historical feature vector, where the historical feature vector is determined by the first interactive content, wherein the content restored based on the historical feature vector is different from the first interactive content; Inputting the historical feature vectors into the risk speech screening model to obtain a target probability vector and a degree of attention for each of the historical feature vectors; Based on the target probability vector and attention level of each of the historical feature vectors and the risk label of the historical interaction content, determine whether to retrain the risky speech screening model and obtain the risky speech screening model when the training is stopped.

5. The method according to claim 4, wherein the inputting the historical feature vectors into the risk speech screening model to obtain the target probability vector and the attention level of each historical feature vector comprises: Inputting the historical feature vectors into the neural network layer of the risk speech screening model to obtain the target probability vector of each of the historical feature vectors; Based on a preset attention mechanism and a target probability vector of each of the historical feature vectors, the attention degree of each of the historical feature vectors is determined.

6. The method according to claim 5, wherein determining whether to retrain the risk speech screening model based on the target probability vector and the attention level of each of the historical feature vectors and the risk label of the historical interaction content comprises: Determining a target score for each of the first interactive contents based on a target probability vector and an attention level of a historical feature vector corresponding to each of the first interactive contents; Based on the risk label and target score of each of the first interactive contents, determine whether to retrain the risk speech screening model.

7. The method according to claim 6, wherein the first feature vector is screened based on a pre-trained risk speech screening model to obtain a third feature vector, comprising: Inputting the first feature vector into the pre-trained risk speech screening model to obtain a target probability vector and a degree of attention for each of the first feature vectors; The third feature vector is determined based on the target probability vector and the attention degree of each of the first feature vectors.

8. The method according to claim 7, wherein determining the third feature vector based on the target probability vector and the attention level of each of the first feature vectors comprises: Based on the target probability vector and the attention level of each of the first feature vectors, a risk score of each of the first feature vectors is determined, and based on the risk score of each of the first feature vectors, the third feature vector is determined.

9. The method according to claim 8, wherein the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech to obtain a pre-trained risk speech recognition model, comprising: randomly selecting a predetermined number of third feature vectors from the historical feature vectors; Based on the feature vector corresponding to the target risk speech and the third feature vector, the risk speech recognition model is trained in an unsupervised training manner to obtain the pre-trained risk speech recognition model.

10. The method according to claim 8, wherein the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech to obtain a pre-trained risk speech recognition model, comprising: Randomly selecting a predetermined number of fifth feature vectors from the fourth feature vectors of the historical feature vectors, the first interaction content corresponding to the fourth feature vector being interaction content without risk; Based on the feature vector corresponding to the target risk speech and the fifth feature vector, the risk speech recognition model is trained to obtain the pre-trained risk speech recognition model.

11. The method according to claim 1, wherein obtaining the target feature vector to be identified comprises: Acquiring the target user's interactive content for the target service; The interactive content is divided into a plurality of sub-contents, and based on a preset encoding rule, the sub-contents are respectively encoded to obtain a plurality of target feature vectors to be identified.

12. A data processing device, comprising: A vector acquisition module, used to acquire a target feature vector to be identified, wherein the target feature vector is determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is generated by the terminal device based on the content in the interactive content other than the content corresponding to the target user, or the terminal device performs matching processing on the target user's interactive content for the target business based on preset risk keywords, and in the case of successful matching, generates a feature vector based on the interactive content that has an associated relationship with the content matching the risk keyword; A vector recognition module is used to identify and process the target feature vector based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, wherein the risk speech recognition model is trained based on a feature vector corresponding to a target risk speech, and the feature vector corresponding to the target risk speech is obtained by screening and processing a first feature vector corresponding to historical interaction content with risks in the target business based on a pre-trained risk speech screening model, wherein the risk speech screening model is a model trained based on a target probability vector of multiple historical feature vectors and a degree of attention and a risk label of historical interaction content, wherein the historical feature vector is determined by a first interaction content, and content restored based on the historical feature vector is different from the first interaction content; The risk detection module is used to determine whether there is risky language in the interactive content based on the identification result, so as to determine whether there is a risk in triggering the execution of the target business.

13. A data processing device, comprising: processor; as well as a memory arranged to store computer executable instructions which, when executed, cause the processor to: Acquire a target feature vector to be identified, the target feature vector is determined by the target user's interactive content for the target business, wherein the content restored based on the target feature vector is different from the interactive content, and the target feature vector is generated by the terminal device based on the content in the interactive content other than the content corresponding to the target user, or the terminal device performs matching processing on the target user's interactive content for the target business based on preset risk keywords, and generates a feature vector based on the interactive content that has an associated relationship with the content matching the risk keyword if the match is successful; The target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech, the feature vector corresponding to the target risk speech is obtained by screening the first feature vector corresponding to the historical interaction content with risks in the target business based on a pre-trained risk speech screening model, the risk speech screening model is a model trained based on the target probability vector and the attention degree of multiple historical feature vectors and the risk label of the historical interaction content, the historical feature vector is determined by the first interaction content, and the content restored based on the historical feature vector is different from the first interaction content; Based on the identification result, determine whether there is risky language in the interactive content to determine whether there is a risk in triggering the execution of the target business.

14. A storage medium for storing computer executable instructions, wherein the executable instructions implement the following process when executed by a processor: Obtain a target feature vector to be identified, where the target feature vector is determined by the target user's interaction content with respect to the target service, wherein: The content restored based on the target feature vector is different from the interactive content, and the target feature vector is generated by the terminal device based on the content in the interactive content other than the content corresponding to the target user, or the terminal device matches the interactive content of the target user for the target business based on preset risk keywords, and generates a feature vector based on the interactive content that has an associated relationship with the content matching the risk keyword if the match is successful; The target feature vector is identified and processed based on a pre-trained risk speech recognition model to obtain an identification result for the target feature vector, the risk speech recognition model is trained based on the feature vector corresponding to the target risk speech, the feature vector corresponding to the target risk speech is obtained by screening the first feature vector corresponding to the historical interaction content with risks in the target business based on a pre-trained risk speech screening model, the risk speech screening model is a model trained based on the target probability vector and the attention degree of multiple historical feature vectors and the risk label of the historical interaction content, the historical feature vector is determined by the first interaction content, and the content restored based on the historical feature vector is different from the first interaction content; Based on the identification result, determine whether there is risky language in the interactive content to determine whether there is a risk in triggering the execution of the target business.

Images