WO2021068635A1 - 信息处理方法、装置及电子设备 - Google Patents

信息处理方法、装置及电子设备 Download PDF

Info

Publication number
WO2021068635A1
WO2021068635A1 PCT/CN2020/107889 CN2020107889W WO2021068635A1 WO 2021068635 A1 WO2021068635 A1 WO 2021068635A1 CN 2020107889 W CN2020107889 W CN 2020107889W WO 2021068635 A1 WO2021068635 A1 WO 2021068635A1
Authority
WO
WIPO (PCT)
Prior art keywords
risk score
target
transaction request
risk
sentence
Prior art date
Application number
PCT/CN2020/107889
Other languages
English (en)
French (fr)
Inventor
徐晓辉
Original Assignee
支付宝(杭州)信息技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 支付宝(杭州)信息技术有限公司 filed Critical 支付宝(杭州)信息技术有限公司
Publication of WO2021068635A1 publication Critical patent/WO2021068635A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q20/00—Payment architectures, schemes or protocols
    • G06Q20/38—Payment protocols; Details thereof
    • G06Q20/40—Authorisation, e.g. identification of payer or payee, verification of customer or shop credentials; Review and approval of payers, e.g. check credit lines or negative lists
    • G06Q20/401—Transaction verification
    • G06Q20/4016—Transaction verification involving fraud or risk level assessment in transaction processing

Definitions

  • This specification relates to the field of information processing technology, and more specifically, to information processing methods, information processing devices, and electronic equipment.
  • the embodiments of this specification provide a new technical solution for managing transaction requests.
  • an information processing method including: acquiring transaction information of a target transaction request; enabling a user to input an interactive sentence based on the transaction information; acquiring the user's input for the target transaction request Interactive sentence; based on the first risk score evaluation model, according to the interactive sentence, the first risk score of the target transaction request is obtained, wherein the first risk score evaluation model is a nerve for determining the risk score based on the interactive sentence Network model; and determining the management and control decision of the target transaction request according to the first risk score.
  • obtaining the first risk score of the target transaction request includes: obtaining a preset vector value of an interaction feature vector from the interaction sentence; and inputting the vector value of the interaction feature vector to the first In the risk assessment model, the first risk score of the target transaction request is obtained.
  • the method further includes: obtaining historical interactive sentences of historical transaction requests as a first sample; setting a first risk score corresponding to the first sample; and according to the first sample and the The first risk score is set, and the first neural network model is trained to obtain the first risk score evaluation model.
  • enabling the user to input an interactive statement based on the transaction information includes: determining the target risk type of the target transaction request according to the transaction information; and when the target risk category is a designated risk category, An entry for inputting the interactive sentence is provided for the user to input the interactive sentence.
  • enabling the user to input an interactive statement based on the transaction information includes: determining the target risk type of the target transaction request according to the transaction information; and when the target risk category is a designated risk category, The user is automatically required to input the interactive sentence.
  • determining the target risk type of the target transaction request according to the transaction information includes: obtaining a second risk score of the target transaction request according to the transaction information based on a second risk score evaluation model, where: The second risk score evaluation model is a neural network model for determining a risk score based on transaction information; and according to the second risk score, the target risk type of the target transaction request is determined.
  • the method further includes: acquiring transaction information of historical transaction requests as a second sample; setting a second risk score corresponding to the second sample; and according to the second sample and the set second risk score For the risk score, the second neural network model is trained to obtain the second risk score evaluation model.
  • determining the management and control decision of the target transaction request further includes: obtaining a comprehensive risk score of the target transaction request according to the first risk score and the second risk score of the target transaction request; and according to the comprehensive risk score To determine the management and control decision of the target transaction request.
  • determining the management and control decision of the target transaction request includes: rejecting the target transaction request when the comprehensive risk score is within a preset comprehensive risk score range.
  • the method further includes: within a set time after the first risk score of the target transaction request is obtained, if a new transaction request triggered by the user is detected again to occur, according to the first risk The score determines the management and control decision of the new transaction request.
  • the method further includes: obtaining a target keyword in the interactive sentence; based on a pre-stored comparison table reflecting the correspondence between the keyword and the response sentence, and according to the target keyword, obtaining a comparison with the A target response sentence corresponding to the interactive sentence; and presenting the target response sentence to the user.
  • an information processing device including: a transaction information acquisition module for acquiring transaction information requested by a target transaction; and a sentence input module for enabling user input interaction based on the transaction information Sentence; an interactive sentence acquisition module for acquiring the interactive sentence input by the user for the target transaction request; a first score determination module for obtaining the target based on the first risk score evaluation model and according to the interactive sentence
  • the first risk score of the transaction request wherein the first risk score evaluation model is a neural network model for determining a risk score based on interactive sentences; and a management and control decision determination module is used to determine the first risk score based on the first risk score.
  • the management and control decision of the target transaction request is described.
  • obtaining the first risk score of the target transaction request includes: obtaining a preset vector value of an interaction feature vector from the interaction sentence; and inputting the vector value of the interaction feature vector to the first In the risk assessment model, the first risk score of the target transaction request is obtained.
  • the device further includes: a module for obtaining historical interaction statements of historical transaction requests as a first sample; a module for setting a first risk score corresponding to the first sample; and According to the first sample and the set first risk score, the first neural network model is trained to obtain the module of the first risk score evaluation model.
  • the enabling sentence input module may also be used to: determine the target risk type of the target transaction request according to the transaction information; and provide an input source when the target risk category is a designated risk category. The entry of the interactive sentence for the user to input the interactive sentence.
  • the enabling sentence input module may also be used to: determine the target risk type of the target transaction request according to the transaction information; and when the target risk category is a designated risk category, automatically request the The user inputs the interactive sentence.
  • determining the target risk type of the target transaction request according to the transaction information includes: obtaining a second risk score of the target transaction request according to the transaction information based on a second risk score evaluation model, where: The second risk score evaluation model is a neural network model for determining a risk score based on transaction information; and according to the second risk score, the target risk type of the target transaction request is determined.
  • the device further includes: a module for obtaining transaction information of historical transaction requests as a second sample; a module for setting a second risk score corresponding to the second sample; and a module for setting a second risk score corresponding to the second sample; and The second sample and the set second risk score are used to train the second neural network model to obtain the module of the second risk score evaluation model.
  • determining the management and control decision of the target transaction request further includes: obtaining a comprehensive risk score of the target transaction request according to the first risk score and the second risk score of the target transaction request; and according to the comprehensive risk score To determine the management and control decision of the target transaction request.
  • determining the management and control decision of the target transaction request includes: rejecting the target transaction request when the comprehensive risk score is within a preset comprehensive risk score range.
  • the device further includes: within a set time after the first risk score of the target transaction request is obtained, if the occurrence of a new transaction request triggered by the user is detected again, according to the first risk score A risk score, which determines the module for the management and control decision of the new transaction request.
  • the device further includes: a module for acquiring the target keyword in the interactive sentence; and a module for obtaining the corresponding relationship between the keyword and the response sentence based on a pre-stored comparison table, according to the target key Word, a module for obtaining the target response sentence corresponding to the interactive sentence; and a module for presenting the target response sentence to the user.
  • an electronic device including: a processor and a memory, the memory is used to store executable instructions, and the instructions are used to control the processor to execute according to the first aspect of this specification The method described.
  • Fig. 1 is a block diagram of a hardware configuration of an information processing system that can be used to implement an embodiment.
  • Fig. 2 shows a flowchart of an information processing method according to an embodiment.
  • Fig. 3 shows a schematic diagram of an information processing scene of an embodiment.
  • Fig. 4 shows a schematic diagram of an information processing scene of another embodiment.
  • Fig. 5 shows a flowchart of an example of an information processing method.
  • Fig. 6 shows a block diagram of an information processing device of an embodiment.
  • Fig. 7 shows a block diagram of an electronic device of an embodiment.
  • FIG. 1 is a schematic diagram of the composition structure of an information processing system to which an information processing method according to an embodiment of this specification can be applied.
  • the information processing system 1000 of this embodiment includes a server 1100, a terminal device 1200, and a network 1300.
  • the server 1100 may be, for example, a blade server, a rack server, etc.
  • the server 1100 may also be a server cluster deployed in the cloud, which is not limited here.
  • the server 1100 may include a processor 1110, a memory 1120, an interface device 1130, a communication device 1140, a display device 1150, and an input device 1160.
  • the processor 1110 may be, for example, a central processing unit CPU or the like.
  • the memory 1120 includes, for example, ROM (Read Only Memory), RAM (Random Access Memory), nonvolatile memory such as a hard disk, and the like.
  • the interface device 1130 includes, for example, a USB interface, a serial interface, and the like.
  • the communication device 1140 can perform wired or wireless communication, for example.
  • the display device 1150 is, for example, a liquid crystal display.
  • the input device 1160 may include, for example, a touch screen, a keyboard, and the like.
  • the memory 1120 of the server 1100 is used to store instructions, and the instructions are used to control the processor 1110 to operate to execute the information processing method of any embodiment of this specification.
  • Technicians can design instructions according to the scheme disclosed in this specification. How the instruction controls the processor to operate is well known in the art, so it will not be described in detail here.
  • server 1100 in the embodiment of the present specification may only involve some of the devices, for example, only the processor 1110 and the memory 1120.
  • the terminal device 1200 may include a processor 1210, a memory 1220, an interface device 1230, a communication device 1240, a display device 1250, an input device 1260, an audio output device 1270, an audio input device 1280, and so on.
  • the processor 1210 may be a central processing unit (CPU), a microprocessor MCU, or the like.
  • the memory 1220 includes, for example, ROM (Read Only Memory), RAM (Random Access Memory), nonvolatile memory such as a hard disk, and the like.
  • the interface device 1230 includes, for example, a USB interface, a headphone interface, and the like.
  • the communication device 1240 can perform wired or wireless communication, for example.
  • the display device 1250 is, for example, a liquid crystal display, a touch display, or the like.
  • the input device 1260 may include, for example, a touch screen, a keyboard, and the like.
  • the terminal device 1200 may output audio information through an audio output device 1270, which includes, for example, a speaker.
  • the terminal device 1200 may pick up the voice information input by the user through an audio pickup device 1280, which includes, for example, a microphone.
  • the terminal device 1200 may be any device that can support the operation of the business system, such as a smart phone, a portable computer, a desktop computer, or a tablet computer.
  • the memory 1220 of the terminal device 1200 is used to store instructions, and the instructions are used to control the processor 1210 to operate to support the realization of the information processing method according to any embodiment of this specification.
  • Technicians can design instructions according to the scheme disclosed in this specification. How the instruction controls the processor to operate is well known in the art, so it will not be described in detail here.
  • terminal device 1200 in the embodiment of this specification may only involve some of the devices, for example, only the processor 1210 and the memory are involved. 1220, display device 1250, input device 1260, etc.
  • the communication network 1300 may be a wireless network or a wired network, and may be a local area network or a wide area network.
  • the terminal device 1200 may communicate with the server 1100 through the communication network 1300.
  • the information processing system 1000 shown in FIG. 1 is only explanatory, and is by no means intended to limit this specification, its application, or use.
  • FIG. 1 only shows one server 1100 and one terminal device 1200, it is not meant to limit the respective numbers.
  • the risk identification system 1000 may include multiple servers 1100 and/or multiple terminal devices 1200.
  • Fig. 2 is a schematic flowchart of an information processing method according to an embodiment.
  • the method shown in FIG. 2 may be implemented solely by the server or terminal device alone, or may be implemented jointly by the server and terminal device.
  • the terminal device may be the terminal device 1200 shown in FIG. 1
  • the server may be the server 1100 shown in FIG. 1.
  • the method of this embodiment includes the following steps S202 to S210.
  • Step S202 Obtain transaction information requested by the target transaction.
  • the transaction information may be information input by the user in response to the target transaction request, for example, it may include at least the transaction amount and the account numbers of both parties to the transaction. Based on the accounts of both parties to the transaction, the user attribute information of both parties can also be obtained, which is also used as transaction information. Among them, the user attribute information may include geographic location, age, total deposits, total average monthly expenditures, total average monthly income, and so on.
  • Step S204 based on the transaction information, enable the user to input an interactive sentence.
  • enabling the user to input an interactive sentence may include: determining the target risk type of the target transaction request according to the transaction information; in the case where the target risk type is a specified risk type, Provide an entry for inputting interactive sentences for users to input interactive sentences.
  • determining the target risk type of the target transaction request according to the transaction information includes: obtaining the second risk score of the target transaction request based on the second risk score evaluation model and the transaction information; The risk score determines the target risk type of the target transaction request.
  • the second risk score evaluation model is a neural network model used to determine the risk score based on transaction information.
  • the method may further include the following steps of obtaining a second risk score evaluation model: obtaining transaction information of historical transaction requests as a second sample; setting a second risk score corresponding to the second sample; according to the second sample And the set second risk score, the second neural network model is trained to obtain the second risk score evaluation model.
  • the second risk score of the target transaction request can be targeted.
  • the target risk type of the target transaction request can be obtained.
  • the risk type corresponding to each second risk score range can be preset.
  • multiple non-overlapping continuous second risk score ranges may be preset according to application scenarios or specific requirements.
  • the preset second risk score ranges may include [0, a), [a, b), and [b,1], where b ⁇ a, and a,b ⁇ [0,1], the corresponding risk types are low risk, medium risk, and high risk.
  • the specified risk type may be one or multiple.
  • the specified risk type may be high risk, then it may be considered that the target risk type is the specified risk type when the target risk type of the target transaction request is high risk.
  • the specified risk type may be medium risk and high risk, then, when the target risk type of the target transaction request is medium risk or high risk, the target risk type is considered to be the specified risk type.
  • the target risk type is the specified risk type, as shown in FIG. 3, an entry for inputting an interactive sentence is provided for the user to input the interactive sentence.
  • an interactive interface for inputting interactive sentences may be provided when the user clicks on the entry.
  • examples of interactive sentences can also be provided to the user for the user to choose.
  • the interactive sentence examples can be preset.
  • the interactive sentence examples can be displayed in the interactive interface. If the user clicks on one of the interactive sentence examples, The example of the clicked interactive sentence can be directly input into the aforementioned input box for the user to edit and/or click to send the interactive sentence to the electronic device that executes this embodiment; or directly submit the clicked interactive sentence
  • An example is provided for the electronic device implementing this embodiment to obtain.
  • enabling the user to input an interactive sentence may further include: determining the target risk type of the target transaction request according to the transaction information; in the case where the target risk type is a specified risk type , The user is automatically required to input interactive sentences.
  • the interactive interface is directly displayed without user operation, so as to automatically require the user to input an interactive sentence.
  • examples of interactive sentences may also be displayed in the interactive interface for the user to select.
  • examples of interactive sentences may also be displayed in the interactive interface for the user to select.
  • Step S206 Obtain the interactive sentence input by the user in response to the target transaction request.
  • a corresponding response sentence may be provided based on the interactive sentence input by the user to the target transaction request.
  • the method may further include: obtaining the target keyword in the interactive sentence; based on a pre-stored comparison table reflecting the correspondence between the keyword and the response sentence, and according to the target keyword, obtaining the target response sentence corresponding to the interactive sentence ; Present the target response sentence to the user.
  • a comparison table reflecting the correspondence between keywords and response sentences may be stored in advance, and according to the comparison table, the response sentence corresponding to each keyword can be determined. According to the target keyword in the interactive sentence, the comparison table is searched, and the response sentence corresponding to the target keyword can be obtained as the target response sentence corresponding to the interactive sentence.
  • the specified response sentence may be presented to the user.
  • the specified response sentence may be used to prompt the user to re-enter other interactive sentences. The details can be shown in Figure 3 and Figure 4.
  • the corresponding response sentence is provided based on the interactive sentence input by the user to the target transaction request, which can provide a targeted anti-fraud guidance service.
  • the acquired interactive sentence input by the user for the target transaction request may be a sentence input by the user one or more times.
  • Step S208 based on the first risk score evaluation model and the interactive sentence, obtain the first risk score of the target transaction request.
  • the first risk assessment model is a neural network model for determining risk scores based on interactive sentences.
  • obtaining the first risk score of the target transaction request may include: obtaining the vector value of the preset interaction feature vector from the interaction sentence; inputting the vector value of the interaction feature vector into the first In the risk assessment model, the first risk score of the target transaction request is obtained.
  • the method may further include: obtaining historical interactive sentences of historical transaction requests as a first sample; setting a first risk score corresponding to the first sample; And the set first risk score, the first neural network model is trained to obtain the first risk score evaluation model.
  • the interactive sentence of the historical transaction request may have a structured field corresponding to the risk score, which is recorded as a risk score field here.
  • the specific risk score can be recorded in the risk score field.
  • the set first risk score corresponding to the first sample may be 1 or 0. For example, when the user issues an alarm for a historical transaction request, the corresponding first risk score is set to 1; when the user does not issue an alarm for a historical transaction request, the corresponding first risk score is set to 0.
  • the interaction feature vector used to describe the transaction request category may be pre-selected, for example, the interaction feature vector is selected according to each structured field of the interaction sentence of the historical transaction request.
  • the interaction feature vector may be composed of at least one feature related to determining the type of the transaction request, and then whether the corresponding transaction request is a fraudulent transaction can be determined according to the interaction feature vector.
  • the interactive statements of historical transaction requests can be processed according to some existing processing methods to obtain features describing the interactive statements related to determining the transaction request category, and then form a category feature vector.
  • the topic model is used to process the interactive statements of historical transaction requests to extract the topic characteristics of the interactive statements of historical transaction requests.
  • embedding is used to process interactive sentences of historical transaction requests to extract stable embedding features.
  • the feature of interactive sentences of historical transaction requests is extracted by means of histgram mapping.
  • one-hot code one-hot is used to extract the characteristics of interactive sentences of historical transaction requests.
  • the selected interaction feature vector includes topic features and embedded features.
  • the mapping relationship between the interaction feature vector and the transaction request type can be obtained.
  • the transaction request type is either a fraudulent transaction type or is not a fraudulent transaction type.
  • the first risk assessment model may be a mapping function F(x), the independent variable of the mapping function F(x) is the feature vector X of the interactive statement of the historical transaction request, and the dependent variable F(x) is Is the function value determined by the vector value of the feature vector X, where the function value is true and the historical transaction request is a fraudulent transaction, the function value is false and the historical transaction request is not a fraudulent transaction, and the vector value is determined by the feature vector X The value composition of each feature of.
  • the method has high accuracy and can accurately obtain the first risk score of the interactive sentence of the target transaction request.
  • Step S210 Determine the management and control decision of the target transaction request according to the first risk score.
  • the target transaction request when the first risk score is within the preset first risk score range, the target transaction request is rejected; when the first risk score is outside the first risk score range, Agree to the target transaction request.
  • the first risk score range may be preset according to application scenarios or specific requirements. When the first risk score is within the range of the first risk score, it is determined that the target transaction request is a fraudulent transaction, and the target transaction request can be rejected; when the first risk score is outside the range of the first risk score, the target transaction request is determined to be non-fraudulent For normal transactions, you can agree to the target transaction request.
  • the range of the first risk score may be [0.7, 1], then when the first risk score is greater than or equal to 0.7 and less than or equal to 1, the target transaction request is determined to be a fraudulent transaction, and the target transaction request is rejected; When the risk score is less than 0.7 or greater than 1, it is determined that the target transaction request is a non-fraudulent normal transaction.
  • the step of determining the management and control decision of the target transaction request according to the first risk score may include: obtaining the target transaction request according to the first risk score and the second risk score of the target transaction request Comprehensive risk score: According to the comprehensive risk score, determine the control decision of the target transaction request.
  • the weights of the first risk score and the second risk score can be set separately in advance, and according to the weights, the first risk score and the second risk score of the target transaction request are weighted and averaged to obtain the target transaction The requested combined risk score.
  • the weights of the first risk score and the second risk score can be preset as ⁇ 1 and ⁇ 2 respectively .
  • the comprehensive risk score F of the transaction request can be expressed as:
  • the first risk score range may be preset according to application scenarios or specific requirements.
  • the comprehensive risk score is within the comprehensive risk score range, it is determined that the target transaction request is a fraudulent transaction, and the target transaction request can be rejected; when the comprehensive risk score is outside the comprehensive risk score range, the target transaction request is determined to be a non-fraudulent normal transaction. You can agree to the target transaction request.
  • the range of the comprehensive risk score can be [0.7,1], then when the comprehensive risk score is greater than or equal to 0.7 and less than or equal to 1, the target transaction request is determined to be a fraudulent transaction, and the target transaction request is rejected; when the comprehensive risk score is less than When 0.7 or greater than 1, it is determined that the target transaction request is a normal transaction that is not fraudulent.
  • the target transaction request may be rejected when the user closes the interactive interface. It may also be that after the user closes the interactive interface, when a specified operation is performed on the target transaction request, the target transaction request is rejected.
  • the specified operation may be an operation of confirming payment or confirming the transfer.
  • the target transaction request can be rejected, and the target transaction request can be displayed to the user In the corresponding transaction failure interface, the reason for rejecting the target transaction request can also be displayed in the transaction failure interface.
  • the reason can be preset, such as "the transaction has a risk of fraud, and the transaction will be stopped to ensure payment security", etc. . If the first risk score of the target transaction request is outside the range of the first risk score, the target transaction request may be approved, so that the user transfer or payment is successful.
  • the target transaction request can be rejected, and the target transaction request can be displayed to the user In the corresponding transaction failure interface, the reason for rejecting the target transaction request can also be displayed in the transaction failure interface.
  • the reason can be preset, such as "the transaction has a risk of fraud, and the transaction will be stopped to ensure payment security", etc. . If the comprehensive risk score of the target transaction request is outside the comprehensive risk score range, the target transaction request may be approved, so that the user transfer or payment is successful.
  • the method may further include: within a set time after the first risk score of the target transaction request is obtained, if it is detected again that a new transaction request triggered by the user occurs, according to The first risk score determines the control decision of the new transaction request.
  • the set time can be set according to the application scenario or specific requirements.
  • the set time can be 1 day. Then, it can be within 1 day after the first risk score of the target transaction request is obtained. If the set time is detected again A new transaction request triggered by the user occurs, and the management and control decision of the new transaction request is determined according to the first risk score.
  • the management and control decision of the target transaction request is determined only based on the first risk score, within a set time after the first risk score of the target transaction request is obtained, if it is detected that a new transaction request triggered by the user occurs, then The management and control decision of the target transaction request can be regarded as the management and control decision of the new transaction request.
  • the comprehensive risk score of the target transaction request is obtained; in the embodiment in which the management and control decision of the target transaction request is determined according to the comprehensive risk score, the first risk of the target transaction request is obtained Within a set time after the score, if it is detected that a new transaction request triggered by the user occurs, the new second risk score of the new transaction request can be determined according to the transaction information of the new transaction request, and the new second risk score of the new transaction request is determined according to the target transaction request.
  • the first risk score and the new second risk score of the new transaction request obtain the new comprehensive risk score of the new transaction request; according to the new comprehensive risk score, the management and control decision of the new transaction request is determined.
  • the method can provide users with automatic consultation and answering services. In one embodiment, this method can provide correct guidance in the event that the user may be deceived during the transaction. In an embodiment, the method can also determine the risk of the target transaction request based on the user's interactive statement. In one embodiment, the method can make the risk determination result of the target transaction request accurate and reliable. In one embodiment, this method can effectively reduce the occurrence of fraudulent transactions. In one embodiment, the method can provide targeted consultation and safety guidance to users, and improve the efficiency of risk transaction management and control. In some embodiments, there may be multiple of the above effects at the same time.
  • the method includes steps S502 to S516.
  • Step S502 Obtain transaction information requested by the target transaction.
  • Step S504 based on the second risk score evaluation model, obtain the second risk score of the target transaction request according to the transaction information.
  • Step S506 Determine the target risk type of the target transaction request according to the second risk score.
  • step S508 when the target risk type is the designated risk type, an entry for inputting an interactive sentence is provided for the user to input the interactive sentence.
  • Step S510 Obtain the interactive sentence input by the user in response to the target transaction request.
  • Step S512 based on the first risk score evaluation model and the interactive sentence, obtain the first risk score of the target transaction request.
  • Step S514 According to the first risk score and the second risk score of the target transaction request, the comprehensive risk score of the target transaction request is obtained.
  • an information processing device 6000 is provided. As shown in FIG. 6, the information processing device 6000 includes a transaction information acquisition module 6100, a sentence input module 6200, an interactive sentence acquisition module 6300, a first score determination module 6400, and a management control decision determination module 6500.
  • the transaction information obtaining module 6100 is used to obtain the transaction information of the target transaction request; the statement input module 6200 is used to enable the user to input an interactive statement based on the transaction information; the interactive statement obtaining module 6300 is used to obtain the user input for the target transaction request Interactive statement; the first score determination module 6400 is used to obtain the first risk score of the target transaction request based on the first risk score evaluation model and the interactive statement, where the first risk score evaluation model is used to determine the risk based on the interactive statement Score neural network model; the management and control decision determining module 6500 is used to determine the management and control decision of the target transaction request according to the first risk score.
  • the first score determining module 6400 may also be used to: obtain the vector value of the preset interaction feature vector from the interactive sentence; and input the vector value of the interaction feature vector into the first In the risk assessment model, the first risk score of the target transaction request is obtained.
  • the information processing device 6000 may further include: a module for acquiring historical interaction statements of historical transaction requests as the first sample; A module for the first risk score; a module for training the first neural network model according to the first sample and the set first risk score to obtain the first risk score evaluation model.
  • the sentence input module 6200 can also be used to: determine the target risk type of the target transaction request according to the transaction information; and provide input when the target risk category is a specified risk category The entrance of the interactive sentence for the user to input the interactive sentence.
  • the sentence input module 6200 can also be used to: determine the target risk type of the target transaction request according to the transaction information; and when the target risk category is a specified risk category, automatically request The user enters an interactive sentence.
  • determining the target risk type of the target transaction request according to the transaction information includes: obtaining the second risk score of the target transaction request based on the second risk score evaluation model and the transaction information, where the first The second risk score evaluation model is a neural network model used to determine the risk score based on transaction information; and according to the second risk score, the target risk type of the target transaction request is determined.
  • the information processing device 6000 may further include: a module for acquiring transaction information of historical transaction requests as a second sample; and a module for setting a second risk corresponding to the second sample A score module; and a module for training the second neural network model according to the second sample and the set second risk score to obtain the second risk score evaluation model.
  • the management control decision determining module 6500 may also be used to: obtain the comprehensive risk score of the target transaction request according to the first risk score and the second risk score of the target transaction request; , To determine the management and control decision of the target transaction request.
  • the management control decision determining module 6500 may also be used to reject the target transaction request when the comprehensive risk score is within the preset comprehensive risk score range.
  • the information processing device 6000 may further include: within a set time after the first risk score of the target transaction request is obtained, if a new transaction triggered by the user is detected again When the request occurs, according to the first risk score, the module for the management and control decision of the new transaction request is determined.
  • the information processing device 6000 may further include: a module for acquiring target keywords in an interactive sentence; and a module for reflecting the corresponding relationship between the keyword and the response sentence stored in advance. According to the target keyword, a module for obtaining the target response sentence corresponding to the interactive sentence; and a module for presenting the target response sentence to the user.
  • the information processing device 6000 can be implemented in various ways.
  • the information processing device 6000 can be implemented by configuring a processor with instructions.
  • the instructions can be stored in the ROM, and when the device is started, the instructions are read from the ROM into the programmable device to realize the information processing apparatus 6000.
  • the information processing device 6000 can be solidified into a dedicated device (for example, ASIC).
  • the information processing apparatus 6000 may be divided into mutually independent units, or they may be combined together for implementation.
  • the information processing device 6000 may be implemented by one of the foregoing various implementation manners, or may be implemented by a combination of two or more of the foregoing various implementation manners.
  • the information processing device 6000 can have multiple implementation forms.
  • the information processing device 6000 can be any software product or functional module running in an application program that provides information processing functions, or these software products or applications.
  • the peripheral embedded parts, plug-ins, patches, etc. of the program can also be these software products or the application itself.
  • an electronic device 7000 is also provided.
  • the electronic device 7000 may include a server 1100 as shown in FIG. 1 or a terminal device 1200 as shown in FIG. 1.
  • the electronic device 7000 may further include a processor 7100 and a memory 7200, where the memory 7200 is used to store executable instructions; the processor 7100 is used to operate the electronic device 7000 according to the control of the instructions to execute arbitrary implementations according to this specification Examples of information processing methods.
  • This manual can be an electronic device, method and/or computer program product.
  • the computer program product may include a computer-readable storage medium loaded with computer-readable program instructions for enabling a processor to implement various aspects of this specification.
  • the computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory flash memory
  • SRAM static random access memory
  • CD-ROM compact disk read-only memory
  • DVD digital versatile disk
  • memory stick floppy disk
  • mechanical encoding device such as a printer with instructions stored thereon
  • the computer-readable storage medium used here is not interpreted as the instantaneous signal itself, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or through wires Transmission of electrical signals.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing/processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network may include copper transmission cables, optical fiber transmission, wireless transmission, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives computer-readable program instructions from the network, and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device .
  • the computer program instructions used to perform the operations of this manual can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, status setting data, or in one or more programming languages.
  • Source code or object code written in any combination, the programming language includes object-oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as "C" language or similar programming languages.
  • Computer-readable program instructions can be executed entirely on the user's computer, partly on the user's computer, executed as a stand-alone software package, partly on the user's computer and partly executed on a remote computer, or entirely on the remote computer or server carried out.
  • the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (for example, using an Internet service provider to access the Internet). connection).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), can be customized by using the status information of the computer-readable program instructions.
  • the computer-readable program instructions are executed to realize various aspects of this specification.
  • These computer-readable program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine that makes these instructions when executed by the processor of the computer or other programmable data processing device , A device that implements the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams is produced. It is also possible to store these computer-readable program instructions in a computer-readable storage medium. These instructions make computers, programmable data processing apparatuses, and/or other devices work in a specific manner. Thus, the computer-readable medium storing the instructions includes An article of manufacture, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowcharts and/or block diagrams.
  • each block in the flowchart or block diagram may represent a module, program segment, or part of an instruction, and the module, program segment, or part of an instruction contains one or more components for realizing the specified logical function.
  • Executable instructions may also occur in a different order than the order marked in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, or they can sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagram and/or flowchart, and the combination of the blocks in the block diagram and/or flowchart can be implemented by a dedicated hardware-based system that performs the specified functions or actions Or it can be realized by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that realization by hardware, realization by software, and realization by a combination of software and hardware are all equivalent.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Accounting & Taxation (AREA)
  • Computer Security & Cryptography (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种信息处理方法、装置及电子设备。其中一个方法包括：获取目标交易请求的交易信息（S202）；基于交易信息，使得用户输入交互语句（S204）；获取用户针对目标交易请求输入的交互语句（S206）；基于第一风险分数评估模型，根据交互语句，得到目标交易请求的第一风险分数（S208），其中，第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；以及根据第一风险分数，确定目标交易请求的管控决策（S210）。可以提高风险交易的管控效率。

Description

信息处理方法、装置及电子设备 技术领域
本说明书涉及信息处理技术领域，更具体地，涉及信息处理方法、信息处理装置、及电子设备。
背景技术
目前，通过终端设备进行转账操作以及支付操作等交易场景的操作已经成为人们的主要交易操作渠道。随着互联网电商时代的发展，互联网欺诈的形势愈来愈严峻，欺诈者通常会编制各类的剧本对不同的用户群体进行定性欺诈。
因此，需要提供可靠的管控交易请求的技术方案。
发明内容
本说明书的实施例提供了一种管控交易请求的新技术方案。
根据本说明书的第一方面，提供了一种信息处理方法，包括：获取目标交易请求的交易信息；基于所述交易信息，使得用户输入交互语句；获取所述用户针对所述目标交易请求输入的交互语句；基于第一风险分数评估模型，根据所述交互语句，得到所述目标交易请求的第一风险分数，其中，所述第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；以及根据所述第一风险分数，确定所述目标交易请求的管控决策。
可选的，得到所述目标交易请求的第一风险分数包括：从所述交互语句中获取预设的交互特征向量的向量值；以及将所述交互特征向量的向量值输入至所述第一风险评估模型中，得到所述目标交易请求的第一风险分数。
可选的，所述方法还包括：获取历史交易请求的历史交互语句，作为第一样本；设置与所述第一样本对应的第一风险分数；根据所述第一样本和所述设置的第一风险分数，对第一神经网络模型进行训练，以得到所述第一风险分数评估模型。
可选的，基于所述交易信息，使得用户输入交互语句，包括：根据所述交易信息确定所述目标交易请求的目标风险类型；以及在所述目标风险类别为指定的风险类别的情况下，提供输入所述交互语句的入口，以供所述用户输入所述交互语句。
可选的，基于所述交易信息，使得用户输入交互语句，包括：根据所述交易信息确定所述目标交易请求的目标风险类型；以及在所述目标风险类别为指定的风险类别的情况下，自动要求所述用户输入所述交互语句。
可选的，根据所述交易信息确定所述目标交易请求的目标风险类型包括：基于第二风险分数评估模型，根据所述交易信息，得到所述目标交易请求的第二风险分数，其中，所述第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型；以及根据所述第二风险分数，确定所述目标交易请求的所述目标风险类型。
可选的，所述方法还包括：获取历史交易请求的交易信息，作为第二样本；设置与所述第二样本对应的第二风险分数；以及根据所述第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到所述第二风险分数评估模型。
可选的，确定所述目标交易请求的管控决策还包括：根据所述目标交易请求的第一风险分数和第二风险分数，得到所述目标交易请求的综合风险分数；根据所述综合风险分数，确定所述目标交易请求的管控决策。
可选的，确定所述目标交易请求的管控决策包括：在所述综合风险分数在预设的综合风险分数范围内时，拒绝所述目标交易请求。
可选的，所述方法还包括：在得到所述目标交易请求的第一风险分数之后的设定时间内，如果再次检测到所述用户触发的新的交易请求发生，根据所述第一风险分数，确定所述新的交易请求的管控决策。
可选的，所述方法还包括：获取所述交互语句中的目标关键词；基于预先存储的反映关键词和应答语句之间对应关系的对照表，根据所述目标关键词，得到与所述交互语句对应的目标应答语句；以及向用户呈现所述目标应答语句。
根据本说明书的第二方面，提供了一种信息处理装置，包括：交易信息获取模块，用于获取目标交易请求的交易信息；使得语句输入模块，用于基于所述交易信息，使得用户输入交互语句；交互语句获取模块，用于获取所述用户针对所述目标交易请求输入的交互语句；第一分数确定模块，用于基于第一风险分数评估模型，根据所述交互语句，得到所述目标交易请求的第一风险分数，其中，所述第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；以及管控决策确定模块，用于根据所述第一风险分数，确定所述目标交易请求的管控决策。
可选的，得到所述目标交易请求的第一风险分数包括：从所述交互语句中获取预设 的交互特征向量的向量值；以及将所述交互特征向量的向量值输入至所述第一风险评估模型中，得到所述目标交易请求的第一风险分数。
可选的，所述装置还包括：用于获取历史交易请求的历史交互语句，作为第一样本的模块；用于设置与所述第一样本对应的第一风险分数的模块；用于根据所述第一样本和所述设置的第一风险分数，对第一神经网络模型进行训练，以得到所述第一风险分数评估模型的模块。
可选的，所述使得语句输入模块还可以用于：根据所述交易信息确定所述目标交易请求的目标风险类型；以及在所述目标风险类别为指定的风险类别的情况下，提供输入所述交互语句的入口，以供所述用户输入所述交互语句。
可选的，所述使得语句输入模块还可以用于：根据所述交易信息确定所述目标交易请求的目标风险类型；以及在所述目标风险类别为指定的风险类别的情况下，自动要求所述用户输入所述交互语句。
可选的，根据所述交易信息确定所述目标交易请求的目标风险类型包括：基于第二风险分数评估模型，根据所述交易信息，得到所述目标交易请求的第二风险分数，其中，所述第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型；以及根据所述第二风险分数，确定所述目标交易请求的所述目标风险类型。
可选的，所述装置还包括：用于获取历史交易请求的交易信息，作为第二样本的模块；用于设置与所述第二样本对应的第二风险分数的模块；以及用于根据所述第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到所述第二风险分数评估模型的模块。
可选的，确定所述目标交易请求的管控决策还包括：根据所述目标交易请求的第一风险分数和第二风险分数，得到所述目标交易请求的综合风险分数；根据所述综合风险分数，确定所述目标交易请求的管控决策。
可选的，确定所述目标交易请求的管控决策包括：在所述综合风险分数在预设的综合风险分数范围内时，拒绝所述目标交易请求。
可选的，所述装置还包括：用于在得到所述目标交易请求的第一风险分数之后的设定时间内，如果再次检测到所述用户触发的新的交易请求发生，根据所述第一风险分数，确定所述新的交易请求的管控决策的模块。
可选的，所述装置还包括：用于获取所述交互语句中的目标关键词的模块；用于基 于预先存储的反映关键词和应答语句之间对应关系的对照表，根据所述目标关键词，得到与所述交互语句对应的目标应答语句的模块；以及用于向用户呈现所述目标应答语句的模块。
根据本说明书的第三方面，提供了一种电子设备，包括：处理器和存储器，所述存储器用于存储可执行的指令，所述指令用于控制所述处理器执行根据本说明书第一方面所述的方法。
通过以下参照附图对本说明书的示例性实施例的详细描述，本说明书的其它特征及其优点将会变得清楚。
附图说明
被结合在说明书中并构成说明书的一部分的附图示出了本说明书的实施例，并且连同其说明一起用于解释本说明书的原理。
图1是可用于实现一个实施例的信息处理系统的硬件配置的框图。
图2示出了一个实施例的信息处理方法的流程图。
图3示出了一个实施例的信息处理场景的示意图。
图4示出了另一个实施例的信息处理场景的示意图。
图5示出了信息处理方法的一个例子的流程图。
图6示出了一个实施例的信息处理装置的框图。
图7示出了一个实施例的电子设备的框图。
具体实施方式
现在将参照附图来详细描述本说明书的各种示例性实施例。
以下对至少一个示例性实施例的描述实际上仅仅是说明性的，决不作为对本说明书及其应用或使用的任何限制。
应注意到：相似的标号和字母在下面的附图中表示类似项，因此，一旦某一项在一个附图中被定义，则在随后的附图中不需要对其进行进一步讨论。
<硬件配置>
图1为可以应用根据本说明书一个实施例的信息处理方法的一种信息处理系统的组成结构示意图。
如图1所示，本实施例的信息处理系统1000包括服务器1100、终端设备1200以及网络1300。
服务器1100例如可以是刀片服务器、机架式服务器等，服务器1100也可以是部署在云端的服务器集群，在此不做限定。
如图1所示，服务器1100可以包括处理器1110、存储器1120、接口装置1130、通信装置1140、显示装置1150和输入装置1160。处理器1110例如可以是中央处理器CPU等。存储器1120例如包括ROM(只读存储器)、RAM(随机存取存储器)、诸如硬盘的非易失性存储器等。接口装置1130例如包括USB接口、串行接口等。通信装置1140例如能够进行有线或无线通信。显示装置1150例如是液晶显示屏。输入装置1160例如可以包括触摸屏、键盘等。
本实施例中，服务器1100的存储器1120用于存储指令，该指令用于控制处理器1110进行操作以执行本说明书任意实施例的信息处理方法。技术人员可以根据本说明书所公开方案设计指令。指令如何控制处理器进行操作，这是本领域公知，故在此不再详细描述。
本领域技术人员应当理解，尽管在图1中示出了服务器1100的多个装置，但是，本说明书实施例的服务器1100可以仅涉及其中的部分装置，例如，只涉及处理器1110和存储器1120。
如图1所示，终端设备1200可以包括处理器1210、存储器1220、接口装置1230、通信装置1240、显示装置1250、输入装置1260、音频输出装置1270、音频输入装置1280，等等。其中，处理器1210可以是中央处理器CPU、微处理器MCU等。存储器1220例如包括ROM(只读存储器)、RAM(随机存取存储器)、诸如硬盘的非易失性存储器等。接口装置1230例如包括USB接口、耳机接口等。通信装置1240例如能够进行有线或无线通信。显示装置1250例如是液晶显示屏、触摸显示屏等。输入装置1260例如可以包括触摸屏、键盘等。终端设备1200可以通过音频输出装置1270输出音频信息，该音频输出装置1270例如包括扬声器。终端设备1200可以通过音频拾取装置1280拾取用户输入的语音信息，该音频拾取装置1280例如包括麦克风。
终端设备1200可以是智能手机、便携式电脑、台式计算机、平板电脑等可以支持业 务系统运行的任意设备。
在本实施例中，终端设备1200的存储器1220用于存储指令，所述指令用于控制所述处理器1210进行操作以支持实现根据本说明书任意实施例的信息处理方法。技术人员可以根据本说明书所公开方案设计指令。指令如何控制处理器进行操作，这是本领域公知，故在此不再详细描述。
本领域技术人员应当理解，尽管在图1中示出了终端设备1200的多个装置，但是，本说明书实施例的终端设备1200可以仅涉及其中的部分装置，例如，只涉及处理器1210、存储器1220、显示装置1250、输入装置1260等。
通信网络1300可以是无线网络也可以是有线网络，可以是局域网也可以是广域网。终端设备1200可以通过通信网络1300与服务器1100进行通信。
图1所示的信息处理系统1000仅是解释性的，并且决不是为了要限制本说明书、其应用或用途。例如，尽管图1仅示出一个服务器1100和一个终端设备1200，但不意味着限制各自的数量，风险识别系统1000中可以包含多个服务器1100和/或多个终端设备1200。
<方法实施例>
图2为一个实施例的信息处理方法的示意性流程图。
在一个例子中，图2所示的方法可以仅由服务器或终端设备单独实施，也可以是由服务器和终端设备共同实施。在一个实施例中，终端设备可以是如图1所示的终端设备1200，服务器可以是如图1所示的服务器1100。
如图2所示，本实施例的方法包括如下步骤S202～S210。
步骤S202，获取目标交易请求的交易信息。
在本说明书的一个或多个实施例中，交易信息可以是用户针对目标交易请求所输入的信息，例如，至少可以包括交易金额和交易双方的账号。基于交易双方的账号，还可以获取交易双方自有的用户属性信息，也作为交易信息。其中，用户属性信息可以包括地理位置、年龄、存款总额、平均每月支出总额、平均每月收入总额等。
步骤S204，基于该交易信息，使得用户输入交互语句。
在本说明书的一个或多个实施例中，基于该交易信息，使得用户输入交互语句可以包括：根据交易信息确定目标交易请求的目标风险类型；在目标风险类型为指定的风险 类型的情况下，提供输入交互语句的入口，以供用户输入交互语句。
在本说明书的一个或多个实施例中，根据交易信息确定目标交易请求的目标风险类型包括：基于第二风险分数评估模型，根据交易信息，得到目标交易请求的第二风险分数；根据第二风险分数，确定目标交易请求的目标风险类型。其中，第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型。
在此基础上，该方法还可以包括如下获得第二风险分数评估模型的步骤：获取历史交易请求的交易信息，作为第二样本；设置与第二样本对应的第二风险分数；根据第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到第二风险分数评估模型。
在本实施例中，通过将目标交易请求的交易信息、和预先训练所得到的第二风险分数评估模型，可以目标交易请求的第二风险分数。根据该第二风险分数所对应的第二风险分数范围，可以得到目标交易请求的目标风险类型。
在一个实施例中，可以预先设置每个第二风险分数范围所对应的风险类型。具体的，可以根据应用场景或具体需求预先设置有多个不重叠的连续的第二风险分数范围，例如，预设的第二风险分数范围可以包括[0,a)、[a,b)和[b,1]，其中，b≥a，且a，b∈[0,1]，对应的风险类型分别为低风险、中风险和高风险。
在一个实施例中，指定的风险类型可以是一个也可以是多个。例如，指定的风险类型可以是高风险，那么，可以是在目标交易请求的目标风险类型为高风险时，认为目标风险类型为指定的风险类型。再例如，指定的风险类型可以是中风险和高风险，那么，可以是在目标交易请求的目标风险类型为中风险或者高风险时，认为目标风险类型为指定的风险类型。
在目标风险类型为指定的风险类型的情况下，可以是如图3所示，提供输入交互语句的入口，以供用户输入交互语句。
在本实施例中，如图3所示，可以是在用户点击该入口的情况下，提供用于输入交互语句的交互界面。在该交互界面中，还可以向用户提供交互语句示例，以供用户选择。
其中，交互语句示例可以是预先设置好的，在用户点击交互界面中用于输入交互语句的输入框的情况下，可以将交互语句示例展示在交互界面中，如果用户点击其中一个交互语句示例，该被点击的交互语句示例可以直接被输入到上述输入框中，供用户进行编辑和/或进行点击以将交互语句发送至执行本实施例的电子设备；或者，直接提交该被 点击的交互语句示例，以供执行本实施例的电子设备获取。
在本说明书的一个或多个实施例中，基于该交易信息，使得用户输入交互语句还可以包括：根据交易信息确定目标交易请求的目标风险类型；在目标风险类型为指定的风险类型的情况下，自动要求用户输入交互语句。
例如可以是如图4所示，在目标风险类型为指定的风险类型的情况下，无需用户操作就直接展示交互界面，以自动要求用户输入交互语句。
在本实施例中，也可以在交互界面中展示交互语句示例，以供用户选择。具体可以参照前述实施例中关于交互界面的描述，在此不再赘述。
步骤S206，获取用户针对目标交易请求输入的交互语句。
在本说明书的一个或多个实施例中，基于用户对目标交易请求所输入的交互语句，可以提供对应的应答语句。具体的，该方法还可以包括：获取交互语句中的目标关键词；基于预先存储的反映关键词和应答语句之间对应关系的对照表，根据目标关键词，得到与交互语句对应的目标应答语句；向用户呈现目标应答语句。
在本实施例中，可以预先存储有反映关键词和应答语句之间对应关系的对照表，根据该对照表，可以确定与每个关键词所对应的应答语句。根据交互语句中的目标关键词，查找该对照表，就可以得到与目标关键词所对应的应答语句，作为与交互语句对应的目标应答语句。
如果对照表中未存储有目标关键词、及与目标关键词对应的应答语句，那么，可以向用户呈现指定的应答语句，例如，该指定的应答语句可以用于提示用户重新输入其他交互语句。具体可以如图3和图4所示。
在本实施例中，基于用户对目标交易请求所输入的交互语句，来提供对应的应答语句，可以提供针对性的防骗引导服务。
在本实施例中，获取的用户针对目标交易请求输入的交互语句，可以是用户一次或多次输入的语句。
步骤S208，基于第一风险分数评估模型，根据交互语句，得到目标交易请求的第一风险分数。
其中，第一风险评估模型是用于基于交互语句确定风险分数的神经网络模型。
在本说明书的一个或多个实施例中，得到目标交易请求的第一风险分数可以包括： 从交互语句中获取预设的交互特征向量的向量值；将交互特征向量的向量值输入至第一风险评估模型中，得到目标交易请求的第一风险分数。
在本说明书的一个或多个实施例中，该方法还可以包括：获取历史交易请求的历史交互语句作为第一样本；设置与第一样本对应的第一风险分数；根据第一样本和设置的第一风险分数，对第一神经网络模型进行训练，以得到第一风险分数评估模型。
在一个实施例中，历史交易请求的交互语句中可以具有对应风险分数的结构化字段，在此记为风险分数字段，根据报警信息设置与每个第一样本对应的第一风险分数时，可以在该风险分数字段记录具体的风险分数。在一个例子中，设置的与第一样本对应的第一风险分数可以是1或者是0。例如，在用户针对历史交易请求发出报警的情况下，设置其对应的第一风险分数为1；在在用户未针对历史交易请求发出报警的情况下，设置其对应的第一风险分数为0。
在一个实施例中，可以预先选定用于描述交易请求类别的交互特征向量，例如，根据历史交易请求的交互语句具有的各个结构化字段选定该交互特征向量。
该交互特征向量可以由与确定交易请求类别有关的至少一个特征组成，进而可以根据该交互特征向量确定所对应的交易请求是否为欺诈交易。
在一个实施例中，可以根据现有的一些处理手段对历史交易请求的交互语句进行处理，以获得描述交互语句的与确定交易请求类别有关的特征，进而组成类别特征向量。例如，利用主题模型处理历史交易请求的交互语句，以提取历史交易请求的交互语句的主题特征。又例如，利用嵌入(Embedding)手段处理历史交易请求的交互语句，以提取稳步的嵌入(Embedding)特征。又例如，利用直方图(histgram)映射的手段提取历史交易请求的交互语句的特征。又例如，利用独热码(one-hot)提取历史交易请求的交互语句的特征。
在一个例子中，选定的交互特征向量包括主题特征和嵌入特征。在该例子中，可以将特征向量X表示为X＝(x 1,x 2)，其中，x 1是交互语句的主题特征，x 2是交互语句的嵌入特征。
在一个例子中，根据选定的交互特征向量，可以获得该交互特征向量与交易请求类型间的映射关系，该交易请求类型或者为欺诈交易类型，或者为不属于欺诈交易类型。
该例子中，该第一风险评估模型可以是一个映射函数F(x)，该映射函数F(x)的自变 量即为历史交易请求的交互语句的特征向量X，因变量F(x)即为由特征向量X的向量值决定的函数值，其中，函数值为真值对应历史交易请求属于欺诈交易，函数值为假值对应历史交易请求不属于欺诈交易，向量值则由特征向量X中的每一特征的取值构成。
根据本实施例提供的根据欺诈交易类型准确的第一样本训练得到第一风险评估模型的方法，具有较高的准确性，能够从准确地得到目标交易请求的交互语句的第一风险分数。
步骤S210，根据第一风险分数，确定目标交易请求的管控决策。
在本说明书的一个或多个实施例中，可以是在第一风险分数在预设的第一风险分数范围内时，拒绝目标交易请求；在第一风险分数在第一风险分数范围外时，同意目标交易请求。
在一个实施例中，可以是根据应用场景或具体需求预先设置第一风险分数范围。在第一风险分数在第一风险分数范围内时，判定目标交易请求为欺诈交易，可以拒绝目标交易请求；在第一风险分数在第一风险分数范围之外时，判定目标交易请求为非欺诈的正常交易，可以同意目标交易请求。
例如，该第一风险分数范围可以是[0.7,1]，那么，可以是在第一风险分数大于等于0.7且小于等于1时，判定目标交易请求为欺诈交易，拒绝目标交易请求；在第一风险分数小于0.7或大于1时，判定目标交易请求为非欺诈的正常交易。
在本说明书的一个或多个实施例中，根据第一风险分数，确定目标交易请求的管控决策的步骤可以包括：根据目标交易请求的第一风险分数和第二风险分数，得到目标交易请求的综合风险分数；根据综合风险分数，确定目标交易请求的管控决策。
在一个实施例中，可以预先分别设置第一风险分数和第二风险分数的权重，并根据该权重，对目标交易请求的第一风险分数和第二风险分数进行加权求平均计算，得到目标交易请求的综合风险分数。
例如，可以预先设置第一风险分数和第二风险分数的权重分别为λ 1和λ 2，在得到目标交易请求的第一风险分数为F 1，第二风险分数为F 2的情况下，目标交易请求的综合风险分数F可以表示为：
Figure PCTCN2020107889-appb-000001
在本说明书的一个或多个实施例中，可以是在综合风险分数在预设的综合风险分数 范围内时，拒绝目标交易请求；在综合风险分数在综合风险分数范围外时，同意目标交易请求。
在一个实施例中，可以是根据应用场景或具体需求预先设置第一风险分数范围。在综合风险分数在综合风险分数范围内时，判定目标交易请求为欺诈交易，可以拒绝目标交易请求；在综合风险分数在综合风险分数范围之外时，判定目标交易请求为非欺诈的正常交易，可以同意目标交易请求。
例如，该综合风险分数范围可以是[0.7,1]，那么，可以是在综合风险分数大于等于0.7且小于等于1时，判定目标交易请求为欺诈交易，拒绝目标交易请求；在综合风险分数小于0.7或大于1时，判定目标交易请求为非欺诈的正常交易。
在一个实施例中，可以是在用户关闭交互界面时，拒绝目标交易请求。还可以是在用户关闭交互界面后，针对目标交易请求执行指定操作时，拒绝目标交易请求，该指定操作可以是确认付款或确认转账的操作。
例如，可以是如图3所示，如果目标交易请求的第一风险分数在第一风险分数范围内，那么，在用户执行点击确认付款的操作时，可以拒绝目标交易请求，进而可以向用户展示相应的交易失败界面，在交易失败界面中还可以展示拒绝目标交易请求的理由，该理由可以是预先设定好的，例如可以是“该交易存在诈骗风险，为保证支付安全将停止交易”等。如果目标交易请求的第一风险分数在第一风险分数范围外，可以同意该目标交易请求，使得用户转账或付款成功。
再例如，还可以是如图4所示，如果目标交易请求的综合风险分数在综合风险分数范围内，那么，在用户执行点击确认付款的操作时，可以拒绝目标交易请求，进而可以向用户展示相应的交易失败界面，在交易失败界面中还可以展示拒绝目标交易请求的理由，该理由可以是预先设定好的，例如可以是“该交易存在诈骗风险，为保证支付安全将停止交易”等。如果目标交易请求的综合风险分数在综合风险分数范围外，可以同意该目标交易请求，使得用户转账或付款成功。
在本说明书的一个或多个实施例中，该方法还可以包括：在得到目标交易请求的第一风险分数之后的设定时间内，如果再次检测到该用户触发的新的交易请求发生，根据第一风险分数，确定新的交易请求的管控决策。
该设定时间可以是根据应用场景或具体需求设定，例如，该设定时间可以是1天，那么，可以是在得到目标交易请求的第一风险分数之后的1天内，如果再次检测到 该用户触发的新的交易请求发生，根据第一风险分数，确定新的交易请求的管控决策。
在仅根据第一风险分数确定目标交易请求的管控决策的实施例中，在得到目标交易请求的第一风险分数之后的设定时间内，如果检测到该用户触发的新的交易请求发生，则可以将目标交易请求的管控决策，作为新的交易请求的管控决策。
在根据目标交易请求的第一风险分数和第二风险分数，得到目标交易请求的综合风险分数；根据综合风险分数确定目标交易请求的管控决策的实施例中，在得到目标交易请求的第一风险分数之后的设定时间内，如果检测到该用户触发的新的交易请求发生，则可以是根据新的交易请求的交易信息确定新的交易请求的新的第二风险分数，根据目标交易请求的第一风险分数和新的交易请求的新的第二风险分数，得到新的交易请求的新的综合风险分数；根据新的综合风险分数，确定新的交易请求的管控决策。
在一个实施例中，该方法具有向用户提供自动咨询解答服务。在一个实施例中，该方法可以在用户交易时可能被骗的情况下提供正确引导。在一个实施例中，该方法还可以根据用户的交互语句判断目标交易请求的风险情况。在一个实施例中，该方法可以使得目标交易请求的风险判定结果准确可靠。在一个实施例中，该方法可以有效减少欺诈交易的发生。在一个实施例中，该方法可以对用户进行针对性咨询及安全引导，提高风险交易的管控效率。在某些实施例中，可能同时具有上述效果中的多个。
<例子1>
下面以一个具体的例子来说明信息处理方法实施的过程。如图5所示，该方法包括步骤S502至S516。
步骤S502，获取目标交易请求的交易信息。
步骤S504，基于第二风险分数评估模型，根据交易信息，得到目标交易请求的第二风险分数。
步骤S506，根据第二风险分数，确定目标交易请求的目标风险类型。
步骤S508，在目标风险类型为指定的风险类型的情况下，提供输入交互语句的入口，以供用户输入交互语句。
步骤S510，获取用户针对目标交易请求输入的交互语句。
步骤S512，基于第一风险分数评估模型，根据交互语句，得到目标交易请求的第一风险分数。
步骤S514，根据目标交易请求的第一风险分数和第二风险分数，得到目标交易请求的综合风险分数。
步骤S516，在综合风险分数在预设的综合风险分数范围内时，拒绝目标交易请求。
<装置>
在本实施例中，提供一种信息处理装置6000。如图6所示，该信息处理装置6000包括交易信息获取模块6100、使得语句输入模块6200、交互语句获取模块6300、第一分数确定模块6400和管控决策确定模块6500。该交易信息获取模块6100用于获取目标交易请求的交易信息；该使得语句输入模块6200用于基于交易信息，使得用户输入交互语句；该交互语句获取模块6300用于获取用户针对目标交易请求输入的交互语句；该第一分数确定模块6400用于基于第一风险分数评估模型，根据交互语句，得到目标交易请求的第一风险分数，其中，第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；该管控决策确定模块6500用于根据第一风险分数，确定目标交易请求的管控决策。
在本说明书的一个或多个实施例中，第一分数确定模块6400还可以用于：从交互语句中获取预设的交互特征向量的向量值；以及将交互特征向量的向量值输入至第一风险评估模型中，得到目标交易请求的第一风险分数。
在本说明书的一个或多个实施例中，该信息处理装置6000还可以包括：用于获取历史交易请求的历史交互语句，作为第一样本的模块；用于设置与第一样本对应的第一风险分数的模块；用于根据第一样本和设置的第一风险分数，对第一神经网络模型进行训练，以得到第一风险分数评估模型的模块。
在本说明书的一个或多个实施例中，使得语句输入模块6200还可以用于：根据交易信息确定目标交易请求的目标风险类型；以及在目标风险类别为指定的风险类别的情况下，提供输入交互语句的入口，以供用户输入交互语句。
在本说明书的一个或多个实施例中，使得语句输入模块6200还可以用于：根据交易信息确定目标交易请求的目标风险类型；以及在目标风险类别为指定的风险类别的情况下，自动要求用户输入交互语句。
在本说明书的一个或多个实施例中，根据交易信息确定目标交易请求的目标风险类型包括：基于第二风险分数评估模型，根据交易信息，得到目标交易请求的第二风 险分数，其中，第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型；以及根据第二风险分数，确定目标交易请求的目标风险类型。
在本说明书的一个或多个实施例中，该信息处理装置6000还可以包括：用于获取历史交易请求的交易信息，作为第二样本的模块；用于设置与第二样本对应的第二风险分数的模块；以及用于根据第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到第二风险分数评估模型的模块。
在本说明书的一个或多个实施例中，管控决策确定模块6500还可以用于：根据目标交易请求的第一风险分数和第二风险分数，得到目标交易请求的综合风险分数；根据综合风险分数，确定目标交易请求的管控决策。
在本说明书的一个或多个实施例中，管控决策确定模块6500还可以用于：在综合风险分数在预设的综合风险分数范围内时，拒绝目标交易请求。
在本说明书的一个或多个实施例中，该信息处理装置6000还可以包括：用于在得到目标交易请求的第一风险分数之后的设定时间内，如果再次检测到用户触发的新的交易请求发生，根据第一风险分数，确定新的交易请求的管控决策的模块。
在本说明书的一个或多个实施例中，该信息处理装置6000还可以包括：用于获取交互语句中的目标关键词的模块；用于基于预先存储的反映关键词和应答语句之间对应关系的对照表，根据目标关键词，得到与交互语句对应的目标应答语句的模块；以及用于向用户呈现目标应答语句的模块。
本领域技术人员应当明白，可以通过各种方式来实现信息处理装置6000。例如，可以通过指令配置处理器来实现信息处理装置6000。例如，可以将指令存储在ROM中，并且当启动设备时，将指令从ROM读取到可编程器件中来实现信息处理装置6000。例如，可以将信息处理装置6000固化到专用器件(例如ASIC)中。可以将信息处理装置6000分成相互独立的单元，或者可以将它们合并在一起实现。信息处理装置6000可以通过上述各种实现方式中的一种来实现，或者可以通过上述各种实现方式中的两种或更多种方式的组合来实现。
在本实施例中，信息处理装置6000可以具有多种实现形式，例如，信息处理装置6000可以是任何的提供信息处理功能的软件产品或者应用程序中运行的功能模块，或者是这些软件产品或者应用程序的外设嵌入件、插件、补丁件等，还可以是这些软件产品或者应用程序本身。
<电子设备>
在本实施例中，还提供一种电子设备7000。该电子设备7000可以包括如图1中所示的服务器1100，也可以是如图1中所示的终端设备1200。
如图7所示，电子设备7000还可以包括处理器7100和存储器7200，该存储器7200用于存储可执行的指令；该处理器7100用于根据指令的控制运行电子设备7000执行根据本说明书任意实施例的信息处理方法。
本说明书中的各个实施例均采用递进的方式描述，各个实施例之间相同相似的部分互相参见即可，每个实施例中的说明的都是与其他实施例的不同之处。尤其，对于装置实施例和电子设备实施例而言，由于其基本相似于方法实施例，所以描述的比较简单，相关之处参见方法实施例的部分说明即可。
本说明书可以是电子设备、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本说明书的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本说明书操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本说明书的各个方面。
这里参照根据本说明书实施例的方法、装置和计算机程序产品的流程图和/或框图描述了本说明书的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本说明书的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中 所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。对于本领域技术人员来说公知的是，通过硬件方式实现、通过软件方式实现以及通过软件和硬件结合的方式实现都是等价的。
以上已经描述了本说明书的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。本说明书的范围由所附权利要求来限定。
上述对本说明书特定实施例进行了描述。其他实施例在所附权利要求书的范围内。在一些情况下，在权利要求书中记载的动作或步骤可以按照不同于实施例中的顺序来执行并且仍然可以实现期望的效果。另外，在附图中描绘的过程不一定要求示出的特定顺序或者连接顺序才能实现期望的结果。在某些实施方式中，多任务处理和并行处理也是可以的或者可能是有利的。

Claims (23)

  1. 一种信息处理方法，包括：
    获取目标交易请求的交易信息；
    基于所述交易信息，使得用户输入交互语句；
    获取所述用户针对所述目标交易请求输入的交互语句；
    基于第一风险分数评估模型，根据所述交互语句，得到所述目标交易请求的第一风险分数，其中，所述第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；以及
    根据所述第一风险分数，确定所述目标交易请求的管控决策。
  2. 根据权利要求1所述的方法，得到所述目标交易请求的第一风险分数包括：
    从所述交互语句中获取预设的交互特征向量的向量值；以及
    将所述交互特征向量的向量值输入至所述第一风险评估模型中，得到所述目标交易请求的第一风险分数。
  3. 根据权利要求2所述的方法，还包括：
    获取历史交易请求的历史交互语句，作为第一样本；
    设置与所述第一样本对应的第一风险分数；
    根据所述第一样本和所述设置的第一风险分数，对第一神经网络模型进行训练，以得到所述第一风险分数评估模型。
  4. 根据权利要求1所述的方法，基于所述交易信息，使得用户输入交互语句，包括：
    根据所述交易信息确定所述目标交易请求的目标风险类型；以及
    在所述目标风险类型为指定的风险类型的情况下，提供输入所述交互语句的入口，以供所述用户输入所述交互语句。
  5. 根据权利要求1所述的方法，基于所述交易信息，使得用户输入交互语句，包括：
    根据所述交易信息确定所述目标交易请求的目标风险类型；以及
    在所述目标风险类型为指定的风险类型的情况下，自动要求所述用户输入所述交互语句。
  6. 根据权利要求4或5所述的方法，根据所述交易信息确定所述目标交易请求的目标风险类型包括：
    基于第二风险分数评估模型，根据所述交易信息，得到所述目标交易请求的第二风 险分数，其中，所述第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型；以及
    根据所述第二风险分数，确定所述目标交易请求的所述目标风险类型。
  7. 根据权利要求6所述的方法，还包括：
    获取历史交易请求的交易信息，作为第二样本；
    设置与所述第二样本对应的第二风险分数；以及
    根据所述第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到所述第二风险分数评估模型。
  8. 根据权利要求6所述的方法，确定所述目标交易请求的管控决策还包括：
    根据所述目标交易请求的第一风险分数和第二风险分数，得到所述目标交易请求的综合风险分数；
    根据所述综合风险分数，确定所述目标交易请求的管控决策。
  9. 根据权利要求8所述的方法，确定所述目标交易请求的管控决策包括：
    在所述综合风险分数在预设的综合风险分数范围内时，拒绝所述目标交易请求。
  10. 根据权利要求1所述的方法，还包括：
    在得到所述目标交易请求的第一风险分数之后的设定时间内，如果再次检测到所述用户触发的新的交易请求发生，根据所述第一风险分数，确定所述新的交易请求的管控决策。
  11. 根据权利要求1所述的方法，还包括：
    获取所述交互语句中的目标关键词；
    基于预先存储的反映关键词和应答语句之间对应关系的对照表，根据所述目标关键词，得到与所述交互语句对应的目标应答语句；以及
    向用户呈现所述目标应答语句。
  12. 一种信息处理装置，包括：
    交易信息获取模块，用于获取目标交易请求的交易信息；
    使得语句输入模块，用于基于所述交易信息，使得用户输入交互语句；
    交互语句获取模块，用于获取所述用户针对所述目标交易请求输入的交互语句；
    第一分数确定模块，用于基于第一风险分数评估模型，根据所述交互语句，得到所述目标交易请求的第一风险分数，其中，所述第一风险分数评估模型是用于基于交互语句确定风险分数的神经网络模型；以及
    管控决策确定模块，用于根据所述第一风险分数，确定所述目标交易请求的管控决 策。
  13. 根据权利要求12所述的装置，得到所述目标交易请求的第一风险分数包括：
    从所述交互语句中获取预设的交互特征向量的向量值；以及
    将所述交互特征向量的向量值输入至所述第一风险评估模型中，得到所述目标交易请求的第一风险分数。
  14. 根据权利要求13所述的装置，还包括：
    用于获取历史交易请求的历史交互语句，作为第一样本的模块；
    用于设置与所述第一样本对应的第一风险分数的模块；
    用于根据所述第一样本和所述设置的第一风险分数，对第一神经网络模型进行训练，以得到所述第一风险分数评估模型的模块。
  15. 根据权利要求12所述的装置，使得语句输入模块还可以用于：
    根据所述交易信息确定所述目标交易请求的目标风险类型；以及
    在所述目标风险类别为指定的风险类别的情况下，提供输入所述交互语句的入口，以供所述用户输入所述交互语句。
  16. 根据权利要求12所述的装置，使得语句输入模块还可以用于：
    根据所述交易信息确定所述目标交易请求的目标风险类型；以及
    在所述目标风险类别为指定的风险类别的情况下，自动要求所述用户输入所述交互语句。
  17. 根据权利要求15或16所述的装置，根据所述交易信息确定所述目标交易请求的目标风险类型包括：
    基于第二风险分数评估模型，根据所述交易信息，得到所述目标交易请求的第二风险分数，其中，所述第二风险分数评估模型是用于基于交易信息确定风险分数的神经网络模型；以及
    根据所述第二风险分数，确定所述目标交易请求的所述目标风险类型。
  18. 根据权利要求17所述的装置，还包括：
    用于获取历史交易请求的交易信息，作为第二样本的模块；
    用于设置与所述第二样本对应的第二风险分数的模块；以及
    用于根据所述第二样本和所设置的第二风险分数，对第二神经网络模型进行训练，以得到所述第二风险分数评估模型的模块。
  19. 根据权利要求17所述的装置，确定所述目标交易请求的管控决策还包括：
    根据所述目标交易请求的第一风险分数和第二风险分数，得到所述目标交易请求的 综合风险分数；
    根据所述综合风险分数，确定所述目标交易请求的管控决策。
  20. 根据权利要求19所述的装置，确定所述目标交易请求的管控决策包括：
    在所述综合风险分数在预设的综合风险分数范围内时，拒绝所述目标交易请求。
  21. 根据权利要求12所述的装置，还包括：
    用于在得到所述目标交易请求的第一风险分数之后的设定时间内，如果再次检测到所述用户触发的新的交易请求发生，根据所述第一风险分数，确定所述新的交易请求的管控决策的模块。
  22. 根据权利要求12所述的装置，还包括：
    用于获取所述交互语句中的目标关键词的模块；
    用于基于预先存储的反映关键词和应答语句之间对应关系的对照表，根据所述目标关键词，得到与所述交互语句对应的目标应答语句的模块；以及
    用于向用户呈现所述目标应答语句的模块。
  23. 一种电子设备，包括：处理器和存储器，所述存储器用于存储可执行的指令，所述指令用于在所述电子设备运行时控制所述处理器执行根据权利要求1至11中任一项所述的方法。
PCT/CN2020/107889 2019-10-11 2020-08-07 信息处理方法、装置及电子设备 WO2021068635A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910962388.6A CN110827033A (zh) 2019-10-11 2019-10-11 信息处理方法、装置及电子设备
CN201910962388.6 2019-10-11

Publications (1)

Publication Number Publication Date
WO2021068635A1 true WO2021068635A1 (zh) 2021-04-15

Family

ID=69549202

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/107889 WO2021068635A1 (zh) 2019-10-11 2020-08-07 信息处理方法、装置及电子设备

Country Status (2)

Country Link
CN (1) CN110827033A (zh)
WO (1) WO2021068635A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113837638A (zh) * 2021-09-29 2021-12-24 支付宝(杭州)信息技术有限公司 话术确定方法、装置及设备
CN113988483A (zh) * 2021-12-23 2022-01-28 支付宝(杭州)信息技术有限公司 风险操作行为的管控及其模型的训练方法及电子设备

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110827033A (zh) * 2019-10-11 2020-02-21 支付宝(杭州)信息技术有限公司 信息处理方法、装置及电子设备
CN115088007A (zh) * 2020-03-17 2022-09-20 深圳市欢太科技有限公司 风险评估方法、装置、电子设备及存储介质
CN111539741B (zh) * 2020-05-15 2023-02-24 支付宝(杭州)信息技术有限公司 识别操作风险的方法、系统和非暂时性存储介质
CN111539742B (zh) * 2020-05-28 2022-10-18 支付宝(杭州)信息技术有限公司 信息处理方法、装置、电子设备及存储介质
CN113256300B (zh) * 2021-05-27 2023-04-07 支付宝(杭州)信息技术有限公司 一种交易的处理方法及装置
CN114969293A (zh) * 2022-05-31 2022-08-30 支付宝(杭州)信息技术有限公司 数据处理方法、装置及设备

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014204900A2 (en) * 2013-06-19 2014-12-24 Alibaba Group Holding Limited Method and system for recommending information
CN108665366A (zh) * 2018-04-27 2018-10-16 平安科技（深圳）有限公司 确定用户风险等级的方法、终端设备及计算机可读存储介质
CN108805689A (zh) * 2017-04-26 2018-11-13 腾讯科技（深圳）有限公司 一种贷款风险评估控制方法及装置
CN109063920A (zh) * 2018-08-20 2018-12-21 阿里巴巴集团控股有限公司 一种交易风险识别方法、装置、及计算机设备
CN109064175A (zh) * 2018-06-11 2018-12-21 阿里巴巴集团控股有限公司 一种账户盗用风险防控方法及装置
CN110827033A (zh) * 2019-10-11 2020-02-21 支付宝(杭州)信息技术有限公司 信息处理方法、装置及电子设备

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8527418B2 (en) * 2011-11-22 2013-09-03 The Western Union Company Risk analysis of money transfer transactions
KR101364763B1 (ko) * 2012-02-20 2014-02-27 주식회사 한국프라임테크놀로지 금융거래패턴분석을 이용한 금융사기 경보 시스템 및 방법
CN105590158A (zh) * 2014-12-30 2016-05-18 中国银联股份有限公司 交易风险实时控制系统
CN105512938A (zh) * 2016-02-03 2016-04-20 宜人恒业科技发展（北京）有限公司 一种基于用户长期使用行为的在线信用风险评估方法
CN108133372B (zh) * 2017-12-28 2022-02-18 蚂蚁智安安全技术(上海)有限公司 评估支付风险的方法及装置
CN108121824A (zh) * 2018-01-12 2018-06-05 北京融快线科技有限公司 一种面向金融服务的聊天机器人及系统
CN108898308A (zh) * 2018-06-27 2018-11-27 中国建设银行股份有限公司 风险评估方法、装置、服务器及可读存储介质
CN109376999A (zh) * 2018-09-20 2019-02-22 阿里巴巴集团控股有限公司 一种交易的管控方法、装置及设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2014204900A2 (en) * 2013-06-19 2014-12-24 Alibaba Group Holding Limited Method and system for recommending information
CN108805689A (zh) * 2017-04-26 2018-11-13 腾讯科技（深圳）有限公司 一种贷款风险评估控制方法及装置
CN108665366A (zh) * 2018-04-27 2018-10-16 平安科技（深圳）有限公司 确定用户风险等级的方法、终端设备及计算机可读存储介质
CN109064175A (zh) * 2018-06-11 2018-12-21 阿里巴巴集团控股有限公司 一种账户盗用风险防控方法及装置
CN109063920A (zh) * 2018-08-20 2018-12-21 阿里巴巴集团控股有限公司 一种交易风险识别方法、装置、及计算机设备
CN110827033A (zh) * 2019-10-11 2020-02-21 支付宝(杭州)信息技术有限公司 信息处理方法、装置及电子设备

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113837638A (zh) * 2021-09-29 2021-12-24 支付宝(杭州)信息技术有限公司 话术确定方法、装置及设备
CN113837638B (zh) * 2021-09-29 2024-04-26 支付宝(杭州)信息技术有限公司 话术确定方法、装置及设备
CN113988483A (zh) * 2021-12-23 2022-01-28 支付宝(杭州)信息技术有限公司 风险操作行为的管控及其模型的训练方法及电子设备
CN113988483B (zh) * 2021-12-23 2022-04-29 支付宝(杭州)信息技术有限公司 风险操作行为的管控及其模型的训练方法及电子设备

Also Published As

Publication number Publication date
CN110827033A (zh) 2020-02-21

Similar Documents

Publication Publication Date Title
WO2021068635A1 (zh) 信息处理方法、装置及电子设备
US11443305B2 (en) Context augmentation for processing data from multiple sources
US11055727B1 (en) Account fraud detection
US10069891B2 (en) Channel accessible single function micro service data collection process for light analytics
US11270375B1 (en) Method and system for aggregating personal financial data to predict consumer financial health
US10891631B2 (en) Framework for generating risk evaluation models
WO2019084922A1 (zh) 信息处理方法和系统、服务器、终端、计算机存储介质
US20180365697A1 (en) Suspicious remittance detection through financial behavior analysis
US10755344B2 (en) System framework processor for channel contacts
TW201944306A (zh) 確定高風險用戶的方法及裝置
JP2020506473A (ja) リスクパラメータを調整するための方法、ならびにリスク識別のための方法およびデバイス
WO2020177478A1 (zh) 一种基于信用的资质信息审核方法、装置及设备
US10963885B2 (en) Systems and methods for using machine learning to predict events associated with transactions
US20230274282A1 (en) Transaction tracking and fraud detection using voice and/or video data
US20200234350A1 (en) Predicting delay in a process
KR20190119773A (ko) 콘텐츠 창작자와 후원자 매칭을 통한 온라인 콘텐츠 투자 시스템 및 방법
US11798006B1 (en) Automating content and information delivery
US9747175B2 (en) System for aggregation and transformation of real-time data
US20190279228A1 (en) Suspicious activity report smart validation
US11861698B2 (en) Vehicle selection platform
WO2019095569A1 (zh) 基于微博财经事件的金融分析方法、应用服务器及计算机可读存储介质
US10664457B2 (en) System for real-time data structuring and storage
US10984800B2 (en) Personal assistant device responses based on group presence
CN115795345A (zh) 信息处理方法、装置、设备及存储介质
US20210073907A1 (en) Peer-to-peer cloud-based credit lending

Legal Events

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

Ref document number: 20873392

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20873392

Country of ref document: EP

Kind code of ref document: A1