CN113837638A

CN113837638A - Method, device and equipment for determining dialect

Info

Publication number: CN113837638A
Application number: CN202111153611.6A
Authority: CN
Inventors: 王昊天; 应缜哲; 王维强
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2021-09-29
Filing date: 2021-09-29
Publication date: 2021-12-24
Anticipated expiration: 2041-09-29
Also published as: CN113837638B

Abstract

The embodiment of the specification provides a conversation determining method, a conversation determining device and conversation determining equipment, wherein the method comprises the following steps: acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks; inputting the target risk aggregation into a pre-trained preset conversational recommendation model to obtain a target conversational language matched with the target risk aggregation; and interacting with the target user based on the target conversation, and determining whether to execute the target service based on an interaction result.

Description

Method, device and equipment for determining dialect

Technical Field

The embodiment of the specification relates to the technical field of data processing, in particular to a method, a device and equipment for determining a dialect.

Background

With the rapid development of the internet industry, the network risk is increased, in a wind control scene, an application service provider can interact with a user through customer service personnel before providing service for the user to determine whether the current user intention (such as account transfer, recharging, cash withdrawal and the like) has risk, and in order to reduce the cost of manual participation, risk control can be performed in a man-machine interaction mode. For example, the computer may train a preset conversational recommendation model through historical user behavior and corresponding risks, determine a corresponding conversational recommendation through the trained conversational recommendation model after receiving the user's intent, and interact with the user based on the determined conversational recommendation to control the risks.

However, when a new change occurs in a black fraud technique, a received user intention may also generate a new change, and under the conditions of a large amount of wind control data and a high updating speed, the data processing pressure of model updating is high, and the dialogue recommendation model cannot be updated in time, which may result in that a dialogue corresponding to the new user intention cannot be determined by the dialogue recommendation model, and therefore, a solution capable of determining an interactive dialogue in time in a wind control scene is required.

Disclosure of Invention

An object of the embodiments of the present specification is to provide a method, an apparatus, and a device for determining a word operation, so as to provide a solution capable of improving the efficiency of word operation determination.

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

in a first aspect, an embodiment of the present specification provides a method for determining a word technique, including: acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks; inputting the target risk aggregate into a pre-trained preset tactical recommendation model to obtain a target tactical matching with the target risk aggregate, wherein the preset tactical recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and interacting with the target user based on the target conversation, and determining whether to execute the target service based on an interaction result.

In a second aspect, an embodiment of the present specification provides a method for determining a word technique, applied to a block chain system, including: acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in the block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks; obtaining a target language matched with the target risk aggregation based on a pre-trained preset language recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and sending the target language operation to the equipment of the target user based on the first intelligent contract so that the equipment of the target user interacts with the target user based on the target language operation, and determining whether to execute the target service based on an interaction result.

In a third aspect, embodiments of the present specification provide a speech determination apparatus, including: the information acquisition module is used for acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering the target service to be executed based on the input information to obtain the user intention of triggering the target service to be executed by the target user; the aggregation obtaining module is used for obtaining a target risk aggregation corresponding to the user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks; the language operation determination module is used for inputting the target risk aggregate into a pre-trained preset language operation recommendation model to obtain a target language operation matched with the target risk aggregate, the preset language operation recommendation model is obtained by training a model built through a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and including user behaviors and corresponding risks; and the interaction module is used for interacting with the target user based on the target conversation and determining whether to execute the target service based on an interaction result.

In a fourth aspect, embodiments of the present specification provide a speech determination apparatus, which is an apparatus in a blockchain system, including: the information acquisition module is used for acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in the block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; the aggregation determining module is used for acquiring a target risk aggregation corresponding to the user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks; the language operation determination module is used for obtaining a target language operation matched with the target risk aggregation based on a pre-trained preset language operation recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language operation recommendation model is obtained by training a model built through a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and including user behaviors and corresponding risks; and the session sending module is used for sending the target session to the equipment of the target user based on the first intelligent contract so as to enable the equipment of the target user to interact with the target user based on the target session, and determining whether to execute the target service based on an interaction result.

In a fifth aspect, an embodiment of the present specification provides a speech determination apparatus, including: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to: acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks; inputting the target risk aggregate into a pre-trained preset tactical recommendation model to obtain a target tactical matching with the target risk aggregate, wherein the preset tactical recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and interacting with the target user based on the target conversation, and determining whether to execute the target service based on an interaction result.

In a sixth aspect, an embodiment of the present specification provides a tactical determination apparatus, where the tactical determination apparatus is an apparatus in a blockchain system, and includes: a processor; and a memory arranged to store computer executable instructions that, when executed, cause the processor to: acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in the block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks; obtaining a target language matched with the target risk aggregation based on a pre-trained preset language recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and sending the target language operation to the equipment of the target user based on the first intelligent contract so that the equipment of the target user interacts with the target user based on the target language operation, and determining whether to execute the target service based on an interaction result.

In a seventh aspect, embodiments of the present specification provide a storage medium for storing computer-executable instructions, where the executable instructions, when executed, implement the following processes: acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks; inputting the target risk aggregate into a pre-trained preset tactical recommendation model to obtain a target tactical matching with the target risk aggregate, wherein the preset tactical recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and interacting with the target user based on the target conversation, and determining whether to execute the target service based on an interaction result.

In an eighth aspect, embodiments of the present specification provide a storage medium for storing computer-executable instructions, which when executed by a processor implement the following process: acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in a block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user; acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks; obtaining a target language matched with the target risk aggregation based on a pre-trained preset language recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks; and sending the target language operation to the equipment of the target user based on the first intelligent contract so that the equipment of the target user interacts with the target user based on the target language operation, and determining whether to execute the target service based on an interaction result.

Drawings

In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present specification, and for those skilled in the art, other drawings can be obtained according to the drawings without any creative effort.

FIG. 1A is a flow chart of an embodiment of a method for determining speech;

FIG. 1B is a schematic diagram of a process of an embodiment of a speech determination method of the present disclosure;

FIG. 2 is a schematic diagram of information input for a target service triggering execution in the present specification;

FIG. 3 is a schematic diagram of a process of another embodiment of a speech determination method of the present disclosure;

FIG. 4A is a flow chart of yet another embodiment of a method for determining speech;

FIG. 4B is a schematic diagram of a process of another embodiment of a speech determination method;

FIG. 5 is a schematic processing diagram of another embodiment of a speech determination method;

FIG. 6 is a schematic diagram of a structure of an embodiment of a speech determination apparatus of the present disclosure;

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

fig. 8 is a schematic diagram of a configuration of a speech determination apparatus according to the present description.

Detailed Description

The embodiment of the specification provides a method, a device and equipment for determining a dialect.

In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the present specification, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present specification without any creative effort shall fall within the protection scope of the present specification.

Example one

As shown in fig. 1A and fig. 1B, an embodiment of the present specification provides a method for determining a dialect, where an execution subject of the method may be a terminal device or a server, where the terminal device may be a device such as a personal computer, or may also be a mobile terminal device such as a mobile phone and a tablet computer, and the server may be an independent server or a server cluster composed of multiple servers.

The method may specifically comprise the steps of:

in S102, information input by the target user for the target service triggered to be executed is acquired, and the intention of the target user for triggering to execute the target service is identified based on the input information, so as to obtain the user intention of the target user for triggering to execute the target service.

The target service may be any service related to user privacy, property security, and the like, for example, the target service may be a resource transfer service, a privacy information update service (such as modifying a login password, adding new user information, and the like), information input by the target user for the target service triggered to be executed may be any text information, voice information, and the like, the user intention may be an objective that the target user triggers to execute the target service, for example, the target service may be a resource transfer service, and the user intention triggered to execute the target service by the target user may be to use a transferred resource for paying a certain money, or transfer the resource to a certain object.

In implementation, with the rapid development of the internet industry, the network risk is increased, in a wind control scene, an application service provider can interact with a user through customer service personnel before providing service for the user to determine whether the current user intention (such as account transfer, recharging, cash withdrawal and the like) has risk, and in order to reduce the cost of manual participation, risk control can be performed in a man-machine interaction mode. For example, the computer may train a preset conversational recommendation model through historical user behavior and corresponding risks, determine a corresponding conversational recommendation through the trained conversational recommendation model after receiving the user's intent, and interact with the user based on the determined conversational recommendation to control the risks.

However, when a new change occurs in a black fraud technique, a received user intention may also generate a new change, and under the conditions of a large amount of wind control data and a high updating speed, the data processing pressure of model updating is high, and the dialogue recommendation model cannot be updated in time, which may result in that a dialogue corresponding to the new user intention cannot be determined by the dialogue recommendation model, and therefore, a solution capable of determining an interactive dialogue in time in a wind control scene is required. Therefore, the embodiments of the present disclosure provide a technical solution that can solve the above problems, and refer to the following specifically.

Taking the target service as the resource transfer service in the resource management application installed in the electronic device (i.e., the terminal device or the server), the target user may trigger to start the resource management application and trigger to execute the resource transfer service in the resource management application. The electronic device may receive information input by the target user when detecting that the target user triggers execution of the target service, for example, as shown in fig. 2, the electronic device may display a prompt page with preset prompt information and may receive information input by the target user on the prompt page for the preset prompt information when detecting that the target user triggers execution of the resource transfer service, where the prompt information displayed on the prompt page may be prompt information for obtaining user intention (such as a transfer object, a transfer purpose, and the like).

After receiving the information input by the target user for triggering the execution of the target service, the input information may be subjected to intention identification to obtain the user intention of the target user for triggering the execution of the target service. For example, the information input by the target user may be subjected to keyword matching through a preset keyword dictionary, and the keyword matched with the keyword dictionary is used as the user intention of the target user for triggering the execution of the target service, for example, the information input by the user is "i am in hand with reserve money", in the input information, the keyword matched with the keyword dictionary may be "reserve money", and then the user intention of the target user for triggering the execution of the target service may be "reserve money".

The method for acquiring the user intention is an optional and realizable acquisition method, and in an actual application scenario, there may be a plurality of different acquisition methods, for example, input information may be identified through a pre-trained intention identification model to acquire a corresponding user intention, and the method for acquiring the user intention may be different according to different actual application scenarios, which is not specifically limited in the embodiment of the present specification.

In S104, a target risk aggregation corresponding to the user behavior matched with the user intention in the risk aggregation is obtained.

The risk aggregation may be an aggregation including one or more risks obtained by performing risk aggregation processing on target risk data corresponding to the execution of the target service.

In implementation, one or more user behaviors matched with the user intention of the target user in the preset multiple user behaviors may be acquired, and risk aggregation corresponding to the user behaviors matched with the user intention of the target user is acquired. For example, the preset user behavior may include a spare fund and a guarantee fund, the risk aggregate corresponding to the user behavior of the spare fund may include risk aggregate 1 of risk 1 and risk 2, and the risk aggregate corresponding to the user behavior of the guarantee fund may include risk aggregate 2 of risk 2 and risk 3, and assuming that the user intention of the target user is "spare fund", the risk aggregate corresponding to the user behavior matched with the user intention is risk aggregate 1, that is, the target risk aggregate is risk aggregate 1.

In S106, the target risk aggregate is input into a pre-trained preset conversational recommendation model, and a target conversational language matched with the target risk aggregate is obtained.

The preset tactical recommendation model is obtained by training a model built through a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, the first risk data can comprise risk data which comprise user behaviors and corresponding risks and are except risk log data, for example, the first risk data can be risk data which comprise the user behaviors and the corresponding risks and are determined based on expert experience, the target risk data can comprise the first risk data, and in addition, the target risk data can also comprise risk log data except the historical risk data.

In an implementation, historical risk data may be obtained, where the historical risk data may include historical first risk data and/or historical risk log data, the historical first risk data may be expert-determined historical data that includes user behavior and corresponding risk, and the historical risk log data may be historical risk log data obtained prior to training a preset conversational recommendation model.

The user behavior and the corresponding risk contained in the historical risk data can be extracted, and risk aggregation processing is performed on the risk based on the user behavior to obtain historical risk aggregation.

And inputting the obtained historical risk aggregate into a preset machine learning algorithm for training to obtain a pre-trained preset dialect recommendation model, and inputting the target risk aggregate into the dialect recommendation model to obtain a target dialect matched with the target risk aggregate.

In addition, when risk aggregation is determined, the user behavior and the corresponding risk included in the target risk data may be extracted, and risk aggregation processing may be performed on the risk based on the user behavior to obtain risk aggregation.

The target risk data may include the first risk data, and may also include the first risk data and the historical risk data. For example, assuming that the current time is 2021 year 8 month 1 day, the historical risk data may be historical data of 2021 year 1 month 1 day to 6 month 30 days, and the first risk data may be data of 2021 year 7 month 31 day.

Then, only the first risk data of 2021 year 7, month 31 may be included in the target risk data, and at this time, the user behavior and risk in the first risk data of 2021 year 7, month 31 may be extracted. And performing aggregation processing on the risks based on the extracted user behaviors to obtain risk aggregation.

Alternatively, the target risk data may further include first risk data of 7/31/2021 and risk log data other than the historical risk data, for example, the target risk data may further include risk log data of any time from 7/1/2021 to 7/31/2021. The user behavior and the corresponding risk in the first risk data and the risk log data can be respectively extracted, and the risk is aggregated based on the user behavior to obtain risk aggregation.

Alternatively, the target risk data may further include the first risk data and the historical risk data, that is, the target risk data may include the first risk data of 7/31/2021 and the historical risk data of 1/6/30/2021.

Still alternatively, the target risk data may further include first risk data, risk log data, and historical risk data, that is, the target risk data may include first risk data of 31/7/2021, historical risk data of 30/1/2021, and risk log data of any time from 1/7/31/2021.

In addition, the risks in the historical risk data at least include the risks in the target risk data, for example, the target risk data includes risk 1, risk 2, and risk 3, and the risks in the historical risk data at least include risk 1, risk 2, and risk 3 (for example, the risks in the historical risk data may include risk 1, risk 2, risk 3, and risk 4), that is, the risk aggregation may be an aggregation obtained by performing aggregation processing on risk 1, risk 2, and risk 3, and the historical risk aggregation may be an aggregation obtained by performing aggregation processing on risk 1, risk 2, risk 3, and risk 4.

In S108, interaction is performed with the target user based on the target session, and it is determined whether to perform the target service based on the interaction result.

In implementation, assuming that the target business is a resource transfer business, the user intent of the target user is "reserve money", and the target risk aggregation is a risk aggregation containing loan risks, the determined target dialect may be a dialect related to the loan risks, and the interaction result with the target user based on the target dialect may be as follows:

an electronic device: is your transaction related to loan?

The target user: then, to.

An electronic device: mr. you may be at risk of being cheated, asking for subsequent transactions to be stopped. The cheat can help you loan for borrowing only by paying a reserve fund, and requires you to transfer money. These are all fraudulent routes, please be million not to trust.

The target user: thus? I see that.

An electronic device: the transaction is cheated, and if tens of millions of transactions are paid, the transaction is not disturbed, and you are thanked to listen and see again.

Alternatively, if the target risk aggregate is a risk aggregate containing game risks, the determined target tactics may be related to game risks, and the interaction results with the target user based on the target tactics may be as follows:

an electronic device: do you the transaction relate to games?

The target user: then, to.

An electronic device: is the payee the game platform?

The target user: he says he is an official customer service.

An electronic device: mr. please note that you are transferring to the personal account of the other party, and the regular game platform does not receive the money through the personal account. Recently, many cases of counterfeiting game platforms appear, and your transaction may be at risk of being cheated, and please stop the subsequent transaction.

The target user: and (c) adding.

An electronic device: reminding people that a cheat is wrongly filled in a bank card, unfreezing an account, paying a reserve money to draw a cash and the like, and requiring the transfer of money, wherein the money is cheated and the cheat is not believed by tens of millions.

The target user: i know.

After the interaction result is obtained, the interaction result may be analyzed, and if it can be determined that there is a risk in executing the target service according to the interaction result, the execution of the target service may be suspended, or it may be determined that the target user temporarily does not need to trigger the execution of the target service according to the interaction result, the execution of the target service may be suspended.

The embodiment of the specification provides a method for determining a word operation, which includes acquiring information input by a target user aiming at a target service triggered to be executed, identifying an intention of triggering to execute the target service based on the input information, acquiring a user intention of triggering to execute the target service by the target user, acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in a risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the executed target service and contains one or more risks, inputting the target risk aggregation into a pre-trained preset word operation recommendation model which is obtained by training a model built through a pre-set machine learning algorithm based on historical risk aggregation, and obtaining the target word operation matched with the target risk aggregation, wherein the pre-trained word operation recommendation model is obtained by training the model built through a pre-trained machine learning algorithm based on the historical risk aggregation, and the target risk data at least comprises first risk data except the historical risk data for determining the historical risk aggregation, the first risk data comprises risk data including user behaviors and corresponding risks except risk log data, interacts with the target user based on the target conversation, and determines whether to execute the target business based on an interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

Example two

As shown in fig. 3, an execution subject of the method may be a terminal device or a server, where the terminal device may be a device such as a personal computer, or may also be a mobile terminal device such as a mobile phone or a tablet computer, and the server may be an independent server, or may be a server cluster composed of multiple servers. The method may specifically comprise the steps of:

in S302, when it is detected that the target user triggers execution of the target service, preset prompt information is output.

In S304, information input by the target user for preset prompt information is received.

In implementation, as shown in fig. 2, in a case that it is detected that the target user triggers execution of the target service, preset prompt information may be output, and information input by the target user for the preset prompt information may be received.

In S306, the input information is identified based on the preset named entity identification model, so as to obtain the user intention of the target user to trigger the execution of the target service.

In implementation, the input information may be identified through a preset Named Entity Recognition (NER) model, the identified Named entities are matched through a preset keyword dictionary, and a keyword obtained through matching is used as a user intention of a target user to trigger execution of a target service.

In S308, the user behaviors and corresponding risks included in the historical risk data and the first risk data are respectively extracted through a preset extraction algorithm.

The target risk data may further include historical risk data, and the extracted user behavior may include information related to triggering the behavior, such as identity information of a user corresponding to the user behavior (e.g., the user triggering the behavior, the user corresponding to the behavior), a purpose of triggering the user behavior, and the like.

In an implementation, the historical risk data may include historical risk log data and historical first risk data (i.e., data including user behaviors and corresponding risks other than the historical risk log data), and the user behaviors and corresponding risks included in the historical risk data and the first risk data may be extracted through a preset extraction algorithm, respectively.

In S310, risk aggregation processing is performed on the risk based on the user behavior, and a risk aggregate corresponding to the user behavior is obtained.

In implementation, for example, by presetting an extraction algorithm, the extracted user behavior and corresponding risk may be as shown in table 1 below.

TABLE 1

After the user behavior and the corresponding risk are extracted, the risk may be aggregated based on the user behavior to obtain a risk aggregation corresponding to the user behavior, for example, the risk corresponding to the user behavior of "spare money" in table 1 has a part-time risk and a game risk, and then it may be determined that the risk aggregation corresponding to the user behavior of "spare money" may include the part-time risk and the game risk.

In addition, the risk corresponding to the user behavior may be screened based on the risk quantity threshold, and only the risk whose risk quantity is not less than the risk quantity threshold is selected for risk aggregation, for example, if the risk quantity threshold is 10, then, for the user behavior of "reserve fund", the quantity of game risk is less than the risk quantity threshold, so only the "part-time risk" whose quantity is not less than the risk quantity threshold may be selected for risk aggregation, that is, the risk aggregation corresponding to the user behavior of "reserve fund" may include the part-time risk.

Moreover, the risk aggregation may further include a first occurrence probability between the user behavior and the corresponding risk, for example, the risk corresponding to the user of the "spare money" includes a part-time risk and a game risk, the number of the part-time risks is 45, and the number of the game risks is 5, so that the first occurrence probability between the spare money and the part-time risk is 45/(45+5) 0.9, the first occurrence probability between the spare money and the game risk is 5/(45+5) 0.1, and the risk aggregation corresponding to the user behavior of the "spare money" includes the part-time risk-0.9 and the game risk-0.1.

In S312, the number of user behaviors contained in the historical risk data and the first risk data is acquired.

In implementation, taking the extracted user behavior as the user behavior shown in table 1 in the above-mentioned S310 as an example, the extracted user behavior of "deposit" appears 3 times in total, that is, the number of the user behaviors is 3, the number of times that the user behavior of "guarantee fee" appears is 1, that is, the number of the user behaviors is 1, and so on, the number of each user behavior can be obtained.

In S314, a user behavior library is constructed based on the first user behaviors of which the number is not less than the preset number threshold among the user behaviors.

In implementation, assuming that the preset number threshold is 2, as shown in table 1 above, the first user behaviors whose number is not less than the preset number threshold may be screened out, and if the number of the user behaviors "deposit" is 3 and is not less than the preset number threshold, the user behaviors are the first user behaviors, and a user behavior library is constructed based on the screened out first user behaviors.

Because the data volumes of the historical risk data and the first risk data are possibly large, the number of the extracted user behaviors is large, if the number of some user behaviors is small, the probability that the user behaviors reappear in a task is low, and the user behaviors can be temporarily not considered when a user behavior library is constructed, so that the processing efficiency of subsequent risk aggregation is improved.

In S316, a target user behavior matching the user intention in the user behavior library established in advance is acquired.

Wherein the user behavior library may be constructed based on user behaviors in the target risk data.

In implementation, when there is no user behavior matching the user intent in the user behavior library, a behavior similarity between each user behavior in the user behavior library and the user intent may be obtained based on a preset similarity algorithm, and a target user behavior matching the user intent may be determined based on the behavior similarity.

In S318, the risk aggregate corresponding to the target user behavior is determined as the target risk aggregate.

In S320, risk aggregation processing is performed on the risks in the historical risk data based on the user behaviors in the historical risk data, so as to obtain a historical risk aggregation corresponding to the user behaviors in the historical risk data.

Wherein the historical risk aggregate may include one or more risks and corresponding occurrence probabilities.

In the implementation, the determination method of the historical risk aggregation may refer to the related content of the determination method of the risk aggregation in S310, and is not described herein again.

In S322, the risks in the historical risk aggregation and the corresponding occurrence probabilities are input into a preset machine learning algorithm for training, so as to obtain a trained preset conversational recommendation model.

The preset machine learning algorithm may be any machine learning algorithm, such as a neural network algorithm, a decision tree algorithm, and the like, and different machine learning algorithms may be selected according to different actual application scenarios, which is not specifically limited in this embodiment of the present specification.

In S324, the risk and the first occurrence probability in the target risk aggregation are input into a pre-trained preset conversational recommendation model, so as to obtain a target conversational language matched with the target risk aggregation.

In an implementation, for example, the information input by the target user for the target business triggered to be executed may be "another party lets me download an application", the corresponding user intent may be "download an application", and the target risk aggregate corresponding to the user behavior corresponding to the user intent may include a pig killing disk risk-x, a part-time-work brushing-order risk-y, and a gambling risk-z, where x, y, and z are first occurrence probabilities of the corresponding risks, respectively, and the above risks and the first occurrence probabilities may be input into a pre-trained conversational recommendation model to obtain the corresponding target conversation.

In S326, interaction is performed with the target user based on the target utterance.

In S328, whether there is a risk in executing the target service is detected based on the preset risk recognition model and the interaction result.

In implementation, the preset risk recognition model may be a recognition model obtained by training a model constructed by a preset machine learning algorithm through a historical interaction result corresponding to the target service.

For example, assume that the interaction results are:

an electronic device: do you the transaction relate to games?

The target user: then, to.

An electronic device: is the payee the game platform?

The target user: he says he is an official customer service.

After the interaction result is obtained, keyword extraction can be performed on the interaction result, a first intention of the target user is determined, if the first intention can include 'transaction', 'game', 'official customer service', the extracted first intention can be input into a preset risk identification model, a corresponding risk value is determined, and if the risk value is higher than a preset risk threshold value, it can be determined that there is a risk in executing the target service.

Or, it may also be determined whether there is a risk in executing the target service according to the matching result of the interaction result and the preset keyword library, for example, it is assumed that the keyword matched with the keyword in the preset keyword library corresponding to the target service may have "transaction", "game" or "official customer service", and it may be determined whether there is a risk in executing the target service according to the number of keywords and the preset keyword number threshold, or it may also be determined according to a risk value of each keyword, a risk value corresponding to the interaction result, and it is determined whether there is a risk in executing the target service according to the preset risk threshold and the risk value of the interaction result.

In S330, in the case that it is detected that there is a risk in executing the target service, outputting preset risk prompt information and terminating executing the target service.

In implementation, the corresponding preset risk prompt may be matched according to the keyword in the interaction result, that is, a plurality of risk prompt messages may be preset for the target service, and the corresponding risk prompt message may be determined according to the interaction result.

The embodiment of the specification provides a method for determining a word operation, which includes the steps of acquiring information input by a target user aiming at a target business triggered to be executed, identifying the intention of triggering the target business to be executed based on the input information, acquiring the user intention of triggering the target business to be executed by the target user, acquiring a target risk aggregation corresponding to user behaviors matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by carrying out risk aggregation processing on target risk data corresponding to the target business to be executed and contains one or more risks, inputting the target risk aggregation into a pre-trained pre-set word operation recommendation model to obtain the target word operation matched with the target risk aggregation, and training the model built by a pre-set machine learning algorithm based on historical risk aggregation to obtain the target word operation matched with the target risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, the first risk data comprises risk data except the risk log data and containing user behaviors and corresponding risks, interaction is carried out with the target user based on the target language, and whether the target business is executed or not is determined based on the interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

EXAMPLE III

As shown in fig. 4A and 4B, embodiments of the present disclosure provide a method for determining a word-of-art, which may be performed by a blockchain system. The method may specifically comprise the steps of:

in S402, based on a first intelligent contract pre-deployed in the blockchain system, information input by a target user for triggering the target service to be executed is obtained, and an intention of the target user for triggering the target service to be executed is identified based on the input information, so as to obtain a user intention of the target user for triggering the target service to be executed.

In implementation, according to the functions of the blockchain, a corresponding trusted application program can be developed, and the trusted application program can be packaged into a stand-alone application program and installed in a local terminal device or a server by a target user. Alternatively, the target users may jointly agree on an application as a trusted application, and send related information (e.g., information input for a target service that triggers execution) to the blockchain system through the trusted application.

The block chain system can acquire information input by a target user aiming at the target service triggered and executed based on a first intelligent contract deployed in advance, and identify the intention of triggering and executing the target service based on the input information to obtain the user intention of triggering and executing the target service by the target user. For example, the blockchain system may check the user identity of the target user, and after the check is passed, identify the input information based on the first intelligent contract to obtain the user intention of the target user to trigger execution of the target service.

In S404, a target risk aggregation corresponding to a user behavior matched with the user intention in the risk aggregation is obtained based on the first intelligent contract.

In an implementation, the blockchain system may obtain a user behavior matching the user intent based on the first intelligent contract, and determine a risk aggregate corresponding to the user behavior as a target risk aggregate.

In S406, a target dialect matched with the target risk aggregation is obtained based on the pre-trained preset dialect recommendation model, the target risk aggregation, and the first intelligent contract.

The preset tactical recommendation model can be obtained by training a model built through a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprise first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprise risk data except risk log data and including user behaviors and corresponding risks.

In implementation, the preset tactical recommendation model may be stored in the block chain system in advance, after the target risk aggregation is determined, the preset tactical recommendation model corresponding to the target service may be acquired based on the first intelligent contract, and the target tactical matching the target risk aggregation is acquired based on the preset tactical recommendation model and the target risk aggregation.

In S408, the target session is sent to the device of the target user based on the first intelligent contract, so that the device of the target user interacts with the target user based on the target session, and whether to execute the target service is determined based on the interaction result.

In an implementation, after the target utterance is determined, the target utterance may be sent to a device of a target user based on a first smart contract.

The embodiment of the specification provides a method for determining a word operation, which includes acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in a block chain system, identifying an intention of triggering the target service to be executed based on the input information to obtain a user intention of triggering the target service to be executed by the target user, acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service to be executed and contains one or more risks, acquiring the target word operation matched with the target risk aggregation based on a pre-trained preset word operation recommendation model, the target risk aggregation and the first intelligent contract, and training a model constructed by a preset machine learning algorithm based on a historical risk aggregation by the pre-trained word operation recommendation model, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, the first risk data comprises risk data except the risk log data and containing user behaviors and corresponding risks, the target dialogue is sent to equipment of the target user based on a first intelligent contract, so that the equipment of the target user interacts with the target user based on the target dialogue, and whether the target business is executed or not is determined based on an interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

Example four

As shown in fig. 5, embodiments of the present disclosure provide a method for determining a word, which may be performed by a blockchain system. The method may specifically comprise the steps of:

in S502, based on a first intelligent contract pre-deployed in the blockchain system, information input by a target user for triggering the target service to be executed is obtained, and an intention of the target user for triggering the target service to be executed is identified based on the input information, so as to obtain a user intention of the target user for triggering the target service to be executed.

For the specific processing procedure of S502, reference may be made to relevant contents of S402 in the first embodiment, which is not described herein again.

In S504, a target risk aggregation corresponding to a user behavior matched with the user intention in the risk aggregation is obtained based on the first intelligent contract.

In an implementation, the risk aggregation may be pre-stored in the blockchain system, and the risk aggregation may be obtained directly from the blockchain system according to the first intelligent contract.

Or, the risk aggregation may also be an aggregation including one or more risks obtained by performing risk aggregation processing on target risk data corresponding to the execution target service based on a third intelligent contract pre-deployed and stored in the blockchain system.

For example, the user behavior and the corresponding risk included in the historical risk data and the first risk data may be extracted based on the third intelligent contract and the preset extraction algorithm, and risk aggregation processing may be performed on the risk based on the user behavior to obtain a risk aggregation corresponding to the user behavior.

And acquiring a target user behavior matched with the user intention in a pre-established user behavior library based on the third intelligent contract, and determining the risk aggregation corresponding to the target user behavior as the target risk aggregation. In addition, when the user behavior matched with the user intention does not exist in the user behavior library, the behavior similarity between each user behavior in the user behavior library and the user intention can be obtained based on the third intelligent contract and the preset similarity algorithm, and the target user behavior matched with the user intention is determined based on the behavior similarity.

In addition, the user behavior library may be created based on user behaviors corresponding to the third intelligent contract and the target risk data, for example, the number of the user behaviors included in the historical risk data and the first risk data may be acquired based on the third intelligent contract, and the user behavior library may be constructed based on the first user behaviors of which the number is not less than the preset number threshold among the user behaviors.

In S506, risk aggregation processing is performed on the risks in the historical risk data based on the second intelligent contract pre-deployed in the blockchain system and the user behaviors in the historical risk data, so as to obtain historical risk aggregation corresponding to the user behaviors in the historical risk data.

In implementation, aggregation processing may be performed on risks in the historical risk data based on the second intelligent contract, so as to obtain historical risk aggregation corresponding to user behaviors in the historical risk data.

In S508, a preset machine learning algorithm is trained based on the second intelligent contract, the risks in the historical risk aggregation, and the corresponding occurrence probabilities, so as to obtain a trained preset conversational recommendation model.

In S510, a target utterance matched with the target risk aggregation is obtained based on the pre-trained preset utterance recommendation model, the first intelligent contract, the risk in the target risk aggregation, and the first occurrence probability.

In S512, the target session is sent to the device of the target user based on the first intelligent contract, so that the device of the target user interacts with the target user based on the target session, and determines whether to execute the target service based on the interaction result.

For the specific processing procedures of S510 to S512, reference may be made to the relevant contents of S406 to S408 in the first embodiment, which are not described herein again.

EXAMPLE five

Based on the same idea, the method for determining a dialect provided by the embodiment of the present specification further provides a device for determining a dialect, as shown in fig. 6.

The speech determination apparatus includes: an information obtaining module 601, an aggregation obtaining module 602, a dialect determining module 603, and an interaction module 604, wherein:

an information obtaining module 601, configured to obtain information input by a target user for a target service triggered to be executed, and identify an intention of triggering to execute the target service based on the input information, so as to obtain a user intention of triggering to execute the target service by the target user;

an aggregation obtaining module 602, configured to obtain a target risk aggregation corresponding to a user behavior matched with the user intention in a risk aggregation, where the risk aggregation is an aggregation that includes one or more risks and is obtained by performing risk aggregation processing on target risk data corresponding to the target service;

a language operation determination module 603, configured to input the target risk aggregation into a pre-trained preset language operation recommendation model to obtain a target language operation matched with the target risk aggregation, where the preset language operation recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least includes first risk data other than the historical risk data used for determining the historical risk aggregation, and the first risk data includes risk data including user behaviors and corresponding risks other than risk log data;

an interaction module 604, configured to interact with the target user based on the target session, and determine whether to execute the target service based on an interaction result.

In an embodiment of this specification, the target risk data further includes the historical risk data, and the apparatus further includes:

the data extraction module is used for respectively extracting user behaviors and corresponding risks contained in the historical risk data and the first risk data through a preset extraction algorithm;

and the first aggregation module is used for carrying out risk aggregation processing on the risks based on the user behaviors to obtain the risk aggregation corresponding to the user behaviors.

In this embodiment of the present specification, the aggregation obtaining module 602 is configured to:

acquiring a target user behavior matched with the user intention in a pre-established user behavior library, wherein the user behavior library is constructed based on the user behavior in the target risk data;

determining a risk aggregate corresponding to the target user behavior as the target risk aggregate.

In an embodiment of this specification, the apparatus further includes:

the quantity acquisition module is used for acquiring the quantity of the user behaviors contained in the historical risk data and the first risk data;

and the building module is used for building the user behavior library based on the first user behaviors of which the number is not less than a preset number threshold value in the user behaviors.

under the condition that the user behavior matched with the user intention does not exist in the user behavior library, acquiring the behavior similarity between each user behavior in the user behavior library and the user intention on the basis of a preset similarity algorithm;

and determining target user behaviors matched with the user intentions based on the behavior similarity.

In an embodiment of this specification, the risk aggregation further includes a first occurrence probability between a user behavior and a corresponding risk, and the apparatus further includes:

the second aggregation module is used for carrying out risk aggregation processing on risks in the historical risk data based on user behaviors in the historical risk data to obtain historical risk aggregation corresponding to the user behaviors in the historical risk data, wherein the historical risk aggregation comprises one or more risks and corresponding occurrence probabilities;

the model training module is used for inputting risks in the historical risk aggregation and corresponding occurrence probabilities into the preset machine learning algorithm for training to obtain the trained preset conversational recommendation model;

the morphology determination module 603 is configured to:

and inputting the risk and the first occurrence probability in the target risk aggregation into the pre-trained preset conversational recommendation model to obtain the target conversational language matched with the target risk aggregation.

In this embodiment of the present specification, the information obtaining module 601 is configured to:

outputting preset prompt information under the condition that the target user is detected to trigger the execution of the target service;

and receiving information input by the target user aiming at the preset prompt information.

and identifying the input information based on a preset named entity identification model to obtain the user intention of the target user for triggering and executing the target service.

In this embodiment of the present specification, the interaction module 604 is configured to:

detecting whether the target service is executed with risks or not based on a preset risk identification model and the interaction result;

and outputting preset risk prompt information and stopping executing the target service under the condition that the risk exists in the target service execution.

The embodiment of the specification provides a dialect determining device, the embodiment of the specification provides a dialect determining method, information input by a target user aiming at a target service triggered to be executed is obtained, the intention of triggering the target service to be executed is identified based on the input information, the intention of the target user for triggering the target service to be executed is obtained, a target risk aggregation corresponding to a user behavior matched with the intention of the user in risk aggregation is obtained, the risk aggregation is an aggregation which is obtained by carrying out risk aggregation processing on target risk data corresponding to the target service to be executed and contains one or more risks, the target risk aggregation is input into a pre-trained preset dialect recommending model to obtain the target dialect matched with the target risk aggregation, the pre-trained dialect recommending model is obtained by training a model built through a pre-set machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, the first risk data comprises risk data except the risk log data and containing user behaviors and corresponding risks, interaction is carried out with the target user based on the target language, and whether the target business is executed or not is determined based on the interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

EXAMPLE six

Based on the same idea, the method for determining a dialect provided by the embodiment of the present specification further provides a device for determining a dialect, as shown in fig. 7.

The speech determination apparatus includes: an information obtaining module 701, an aggregation obtaining module 702, a dialect determining module 703 and an interaction module 704, wherein:

an information obtaining module 701, configured to obtain, based on a first intelligent contract pre-deployed in the blockchain system, information input by a target user for a target service to be triggered and executed, and identify, based on the input information, an intention of triggering and executing the target service, so as to obtain a user intention of triggering and executing the target service by the target user;

an aggregation determining module 702, configured to obtain, based on the first intelligent contract, a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, where the risk aggregation is an aggregation that includes one or more risks and is obtained by performing risk aggregation processing on target risk data corresponding to the target service;

a language-technology determining module 703, configured to obtain a target language technology matched with the target risk aggregation based on a pre-trained preset language-technology recommendation model, the target risk aggregation, and the first intelligent contract, where the preset language-technology recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least includes first risk data other than historical risk data used for determining the historical risk aggregation, and the first risk data includes risk data including user behavior and corresponding risk other than risk log data;

a session sending module 704, configured to send the target session to the device of the target user based on the first intelligent contract, so that the device of the target user interacts with the target user based on the target session, and determine whether to execute the target service based on an interaction result.

the aggregation module is used for carrying out risk aggregation processing on risks in the historical risk data based on a second intelligent contract which is deployed in the block chain system in advance and user behaviors in the historical risk data to obtain the historical risk aggregation corresponding to the user behaviors in the historical risk data, and the historical risk aggregation comprises one or more risks and corresponding occurrence probabilities;

the training module is used for training the preset machine learning algorithm based on the second intelligent contract, the risks in the historical risk aggregation and the corresponding occurrence probability to obtain the trained preset dialect recommendation model;

the tactical determination module 703 is configured to:

and obtaining the target dialect matched with the target risk aggregation based on a pre-trained preset dialect recommendation model, the first intelligent contract, the risk in the target risk aggregation and the first occurrence probability.

Acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in a block chain system, identifying the intention of triggering the target service to be executed based on the input information to obtain the user intention of triggering the target service to be executed by the target user, acquiring a target risk aggregation corresponding to user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by carrying out risk aggregation processing on target risk data corresponding to the target service to be executed and contains one or more risks, acquiring a target language matched with the target risk aggregation based on a pre-trained preset language recommendation model, a target risk aggregation and a first intelligent contract, wherein the pre-trained language recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, the first risk data comprises risk data except the risk log data and containing user behaviors and corresponding risks, the target dialogue is sent to equipment of the target user based on a first intelligent contract, so that the equipment of the target user interacts with the target user based on the target dialogue, and whether the target business is executed or not is determined based on an interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

EXAMPLE seven

Based on the same concept, the embodiments of the present specification further provide a speech technology determination device, as shown in fig. 8.

The speech determining apparatus may have a relatively large difference due to configuration or performance, and may include one or more processors 801 and a memory 802, and one or more stored applications or data may be stored in the memory 802. Wherein the memory 802 may be a transient storage or a persistent storage. The application program stored in memory 802 may include one or more modules (not shown), each of which may include a series of computer-executable instructions in a dialog determination device. Still further, the processor 801 may be configured to communicate with the memory 802, and execute a series of computer-executable instructions in the memory 802 on the speech determination device. The speech determination apparatus may also include one or more power supplies 803, one or more wired or wireless network interfaces 804, one or more input-output interfaces 805, one or more keyboards 806.

In particular, in this embodiment, the dialog determination 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 dialog determination device, and the one or more programs configured for execution by the one or more processors include computer-executable instructions for:

acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user;

acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks;

inputting the target risk aggregate into a pre-trained preset tactical recommendation model to obtain a target tactical matching with the target risk aggregate, wherein the preset tactical recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks;

and interacting with the target user based on the target conversation, and determining whether to execute the target service based on an interaction result.

Optionally, the target risk data further includes the historical risk data, and before the target risk aggregation corresponding to the user behavior matched with the user intention in the acquired risk aggregation, the method further includes:

respectively extracting user behaviors and corresponding risks contained in the historical risk data and the first risk data through a preset extraction algorithm;

and performing risk aggregation processing on the risks based on the user behaviors to obtain the risk aggregation corresponding to the user behaviors.

Optionally, the obtaining a target risk aggregation corresponding to a user behavior matched with the user intention in the risk aggregation includes:

Optionally, before the obtaining of the target user behavior matched with the user intention in the pre-established user behavior library, the method further includes:

acquiring the number of user behaviors contained in the historical risk data and the first risk data;

and constructing the user behavior library based on the first user behaviors of which the number is not less than a preset number threshold value in the user behaviors.

Optionally, the obtaining of the target user behavior matched with the user intention in the pre-established user behavior library includes:

Optionally, the risk aggregation further includes a first occurrence probability between a user behavior and a corresponding risk, and before the target risk aggregation is input into a pre-trained preset conversational recommendation model to obtain a target conversational language matched with the target risk aggregation, the method further includes:

performing risk aggregation processing on risks in the historical risk data based on user behaviors in the historical risk data to obtain historical risk aggregates corresponding to the user behaviors in the historical risk data, wherein the historical risk aggregates comprise one or more risks and corresponding occurrence probabilities;

inputting the risks in the historical risk aggregation and the corresponding occurrence probability into the preset machine learning algorithm for training to obtain the trained preset dialect recommendation model;

inputting the target risk aggregation into a pre-trained preset conversational recommendation model to obtain a target conversational language matched with the target risk aggregation, wherein the method comprises the following steps:

Optionally, the obtaining information input by the target user for the target service triggered to be executed includes:

Optionally, the identifying, based on the input information, the intention of triggering execution of the target service to obtain the user intention of the target user to trigger execution of the target service includes:

Optionally, the determining whether to execute the target service based on the interaction result includes:

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

acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in the block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user;

acquiring a target risk aggregation corresponding to a user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks;

obtaining a target language matched with the target risk aggregation based on a pre-trained preset language recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language recommendation model is obtained by training a model constructed by a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and containing user behaviors and corresponding risks;

and sending the target language operation to the equipment of the target user based on the first intelligent contract so that the equipment of the target user interacts with the target user based on the target language operation, and determining whether to execute the target service based on an interaction result.

Optionally, the risk aggregation further includes a first occurrence probability between a user behavior and a corresponding risk, and before obtaining a target utterance matched with the target risk aggregation based on a pre-trained preset utterance recommendation model, the target risk aggregation, and the first intelligent contract, the method further includes:

risk aggregation processing is carried out on risks in the historical risk data based on a second intelligent contract which is deployed in the block chain system in advance and user behaviors in the historical risk data, so that the historical risk aggregation corresponding to the user behaviors in the historical risk data is obtained, and the historical risk aggregation comprises one or more risks and corresponding occurrence probability;

training the preset machine learning algorithm based on the second intelligent contract, the risks in the historical risk aggregation and the corresponding occurrence probability to obtain the trained preset conversational recommendation model;

the obtaining of the target dialect matched with the target risk aggregation based on the pre-trained preset dialect recommendation model, the target risk aggregation and the first intelligent contract comprises:

The embodiment of the specification provides a tactical determination device, which acquires information input by a target user aiming at a target service triggered to be executed, identifies the intention of triggering to execute the target service based on the input information, acquires the user intention of triggering to execute the target service by the target user, acquires a target risk aggregation corresponding to user behaviors matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the executed target service and contains one or more risks, inputs the target risk aggregation into a pre-trained tactical recommendation model which is trained in advance, acquires the target tactical which is matched with the target risk aggregation, the pre-trained tactical recommendation model is obtained by training a model built by a pre-set machine learning algorithm based on historical risk aggregation, and the target risk data at least comprises first risk data except the historical risk data for determining the historical risk aggregation, the first risk data comprises risk data including user behaviors and corresponding risks except risk log data, interacts with the target user based on the target conversation, and determines whether to execute the target business based on an interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

Example eight

The embodiments of the present disclosure further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the computer program implements each process of the foregoing dialect determination method embodiment, and can achieve the same technical effect, and in order to avoid repetition, details are not repeated here. The computer-readable storage medium may be a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk.

The embodiment of the specification provides a computer-readable storage medium, which acquires information input by a target user for a target service triggered to be executed, identifies an intention of the target service triggered to be executed by the target user based on the input information, acquires a target risk aggregation corresponding to a user behavior matched with the user intention in a risk aggregation, the risk aggregation is an aggregation containing one or more risks obtained by performing risk aggregation processing on target risk data corresponding to the target service to be executed, inputs the target risk aggregation into a pre-trained preset tactical recommendation model, acquires a target tactical recommendation matched with the target risk aggregation, the pre-trained tactical recommendation model is obtained by training a model built through a pre-set machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data for determining the historical risk aggregation, the first risk data comprises risk data including user behaviors and corresponding risks except risk log data, interacts with the target user based on the target conversation, and determines whether to execute the target business based on an interaction result. Therefore, the dialect recommendation model is obtained by training through historical risk aggregation, so that the target dialect corresponding to the user intention can be determined through the target risk aggregation and the dialect recommendation model without training the dialect recommendation model again, and the problem that the target dialect cannot be obtained due to the fact that the dialect recommendation model is not updated timely is solved.

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

In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Hardware Description Language), traffic, pl (core universal Programming Language), HDCal (jhdware Description Language), lang, Lola, HDL, laspam, hardward Description Language (vhr Description Language), vhal (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: the ARC625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be considered a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.

The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the various elements may be implemented in the same one or more software and/or hardware implementations in implementing one or more embodiments of the present description.

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

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

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

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

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

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

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of 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 disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

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

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

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

The above description is only an example of the present specification, and is not intended to limit the present specification. Various modifications and alterations to this description will become apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification should be included in the scope of the claims of the present specification.

Claims

1. A method of determining speech, comprising:

2. The method of claim 1, the target risk data further comprising the historical risk data, further comprising, prior to a target risk aggregation corresponding to a user behavior in the obtained risk aggregation that matches the user intent:

3. The method of claim 2, wherein the obtaining of the target risk aggregation corresponding to the user behavior matched with the user intent in the risk aggregation comprises:

4. The method of claim 3, further comprising, prior to said obtaining a target user behavior in a pre-established user behavior library that matches the user intent:

5. The method of claim 4, wherein the obtaining of the target user behavior matched with the user intention from the pre-established user behavior library comprises:

6. The method of claim 5, wherein the risk aggregation further includes a first occurrence probability between a user behavior and a corresponding risk, and before the inputting the target risk aggregation into a pre-trained preset conversational recommendation model to obtain a target conversational language matching the target risk aggregation, the method further includes:

7. The method of claim 1, wherein the obtaining information input by a target user for a target service triggering execution comprises:

8. The method of claim 7, the identifying the intent to trigger execution of the target service based on the input information, resulting in the user intent of the target user to trigger execution of the target service, comprising:

9. The method of claim 8, wherein the determining whether to execute the target service based on the interaction result comprises:

10. A method for determining dialect applied to a block chain system includes:

11. The method of claim 10, wherein the risk aggregation further comprises a first occurrence probability between user behavior and corresponding risk, and before the obtaining of the target language matching the target risk aggregation based on the pre-trained pre-language recommendation model, the target risk aggregation and the first intelligent contract, further comprises:

12. A tactical determination apparatus, comprising:

the information acquisition module is used for acquiring information input by a target user aiming at a target service triggered to be executed, and identifying the intention of triggering the target service to be executed based on the input information to obtain the user intention of triggering the target service to be executed by the target user;

the aggregation obtaining module is used for obtaining a target risk aggregation corresponding to the user behavior matched with the user intention in risk aggregation, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target business and contains one or more risks;

the language operation determination module is used for inputting the target risk aggregate into a pre-trained preset language operation recommendation model to obtain a target language operation matched with the target risk aggregate, the preset language operation recommendation model is obtained by training a model built through a preset machine learning algorithm based on historical risk aggregate, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregate, and the first risk data comprises risk data except risk log data and including user behaviors and corresponding risks;

and the interaction module is used for interacting with the target user based on the target conversation and determining whether to execute the target service based on an interaction result.

13. A tactical determination apparatus, the apparatus being an apparatus in a blockchain system, comprising:

the information acquisition module is used for acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in the block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user;

the aggregation determining module is used for acquiring a target risk aggregation corresponding to the user behavior matched with the user intention in risk aggregation based on the first intelligent contract, wherein the risk aggregation is an aggregation which is obtained by performing risk aggregation processing on target risk data corresponding to the target service and contains one or more risks;

the language operation determination module is used for obtaining a target language operation matched with the target risk aggregation based on a pre-trained preset language operation recommendation model, the target risk aggregation and the first intelligent contract, wherein the preset language operation recommendation model is obtained by training a model built through a preset machine learning algorithm based on historical risk aggregation, the target risk data at least comprises first risk data except the historical risk data used for determining the historical risk aggregation, and the first risk data comprises risk data except risk log data and including user behaviors and corresponding risks;

and the session sending module is used for sending the target session to the equipment of the target user based on the first intelligent contract so as to enable the equipment of the target user to interact with the target user based on the target session, and determining whether to execute the target service based on an interaction result.

14. A tactical determination device, the tactical determination device comprising:

a processor; and

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

15. A tactical determination apparatus, the tactical determination apparatus being an apparatus in a blockchain system, comprising:

a processor; and

16. A storage medium for storing computer-executable instructions, which when executed by a processor implement the following:

17. A storage medium for storing computer-executable instructions, which when executed by a processor implement the following:

acquiring information input by a target user aiming at a target service triggered to be executed based on a first intelligent contract pre-deployed in a block chain system, and identifying the intention of triggering to execute the target service based on the input information to obtain the user intention of triggering to execute the target service by the target user;