CN109992655B

CN109992655B - Intelligent customer service method, device, equipment and storage medium

Info

Publication number: CN109992655B
Application number: CN201910256599.8A
Authority: CN
Inventors: 杨灿; 刘云峰; 文博
Original assignee: Shenzhen Zhuiyi Technology Co Ltd
Current assignee: Shenzhen Zhuiyi Technology Co Ltd
Priority date: 2019-03-29
Filing date: 2019-03-29
Publication date: 2021-04-13
Anticipated expiration: 2039-03-29
Also published as: CN109992655A

Abstract

The invention relates to an intelligent customer service method, an intelligent customer service device, intelligent customer service equipment and a storage medium, wherein the method comprises the following steps: acquiring current input information of a user; determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model; determining whether to fill the word slot or not according to the type of the target conversation behavior; if yes, obtaining elements corresponding to the current input information, and filling the elements into the word slot; if not, outputting the current response information corresponding to the current input information so as to obtain the input information fed back by the user according to the current response information again. That is to say, by adopting the technical scheme provided by the invention, even though different input information of the user may contain the same key data, the word slot is filled by combining the target dialogue behavior type to which the input information of the user belongs, so that the word slot can be accurately and effectively filled, the phenomenon that the word slot cannot be accurately filled by the machine customer service is reduced, and the practicability of the machine customer service is improved.

Description

Intelligent customer service method, device, equipment and storage medium

Technical Field

The invention relates to the technical field of artificial intelligence, in particular to an intelligent customer service method, an intelligent customer service device, intelligent customer service equipment and a storage medium.

Background

With the development of machine customer service technology and the gradual habit of users to machine customer service and the interaction form thereof, the consultation types in the machine customer service of enterprises cover various consultation scenes such as customer service consultation, product consultation and shopping guide, information inquiry, service handling and the like.

Existing machine services include two core elements: word slot filling and multiple rounds of dialogue. After the intention of user consultation triggers a task-type multi-turn scene, the machine customer service will ask reversely through multi-turn clarification and collect word slots required in the task, thereby completing certain business operations (product recommendation, information query, business transaction, etc.).

However, most of the machine services fill the word slots by hitting key data in the input information of the user, and different input information may contain the same key data, so that the machine services cannot accurately fill the word slots, thereby reducing the practicability of the machine services.

Disclosure of Invention

In view of the above, an objective of the present invention is to provide an intelligent customer service method, apparatus, device and storage medium, so as to solve the problem that the utility of the machine customer service is reduced because the machine customer service cannot accurately fill the word slot because different input information may contain the same key data.

In order to achieve the above object, the present invention provides an intelligent customer service method, comprising:

acquiring current input information of a user;

determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model;

determining whether to fill a word slot or not according to the target conversation behavior type;

if the word slot is determined to be filled, acquiring an element corresponding to the current input information, and filling the element into the word slot;

and if the word slot is determined not to be filled, outputting the current response information corresponding to the current input information so as to obtain the input information fed back by the user according to the current response information again.

Further, in the above method, the determining the target dialog behavior type in each dialog behavior type according to a preset dialog behavior model includes:

according to the dialogue behavior model, scoring the current input information according to each dialogue behavior type respectively to obtain a scoring value corresponding to each dialogue behavior type of the current input information;

and determining the target dialogue behavior type according to the numerical value of the scoring value.

Further, in the foregoing method, the determining the target dialog behavior type based on the numerical value of the scoring value includes:

and determining the dialog behavior type with the highest scoring value as the target dialog behavior type.

determining the dialogue behavior type with the highest scoring value as an initially selected dialogue behavior type;

judging whether the highest scoring value is larger than a preset threshold corresponding to the initially selected dialogue behavior type;

and if the highest scoring value is larger than the preset threshold value, determining the initially selected conversation behavior type as the target conversation behavior type.

Further, the method further includes:

if the highest score value is smaller than or equal to the preset threshold value, outputting semantic understanding information of the current input information;

obtaining a confirmation result fed back by the user according to the semantic understanding information;

and if the confirmation result shows that the conversation behavior type is correct, determining the initially selected conversation behavior type as the target conversation behavior type.

Further, the method further includes:

and if the confirmation result shows an error, prompting the user to input the explicit information of the current input information.

Further, in the above method, before outputting the current response information corresponding to the current input information, the method further includes:

and selecting current response information corresponding to the current input information from a preset corpus according to the semantic understanding information of the current input information.

Further, in the above method, the dialog behavior type includes:

providing at least one of an information behavior type, an inquiry behavior type, a question behavior type, a complaint behavior type, a positive reply behavior type, a negative reply behavior type, and a to-be-recognized reply behavior type;

the determining to fill the word slot includes:

if the conversation behavior type to which the current input information belongs is an information providing behavior type or an active reply behavior type, determining to fill the word slot;

the determining not to fill the word slot includes:

and if the conversation behavior type to which the current input information belongs is an inquiry behavior type, a question behavior type, a complaint behavior type, a negative reply behavior type or a reply behavior type to be identified, determining not to fill the word slot.

Further, in the above method, if it is determined that the word slot is filled, acquiring an element corresponding to the current input information includes:

if the conversation behavior type to which the current input information belongs is an information providing behavior type, acquiring the element from the current input information;

and if the conversation behavior type to which the current input information belongs is the positive reply behavior type, acquiring the element according to the current input information and/or the conversation content before the current input information.

The invention also provides an intelligent customer service device, which comprises an acquisition module, a determination module and an output module;

the acquisition module is used for acquiring the current input information of the user;

the determining module is used for determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model; determining whether to fill a word slot or not according to the target conversation behavior type;

the obtaining module is further configured to obtain an element corresponding to the current input information and fill the element into the word slot if the determining module determines to fill the word slot;

and the output module is used for outputting the current response information corresponding to the current input information if the determining module determines that the word slot is not filled, so that the obtaining module can obtain the input information fed back by the user according to the current response information again.

Further, in the above apparatus, the determining module is specifically configured to:

Further, in the above apparatus, the determining module is further configured to:

Further, in the above apparatus, the obtaining module is further configured to:

Further, in the above apparatus, the dialog behavior types include:

the determining module is further configured to:

the determining not to fill the word slot includes:

Further, in the above apparatus, the obtaining module is further configured to:

The invention also provides an intelligent customer service device, which comprises a processor and a memory, wherein the processor and the memory are connected through a communication bus:

the processor is used for calling and executing the program stored in the memory;

the memory for storing the program, the program at least for executing the intelligent customer service method as described above;

the intelligent customer service method comprises the following steps:

acquiring current input information of a user;

Further, the method further includes:

Further, in the above method, the dialog behavior type includes:

the determining to fill the word slot includes:

the determining not to fill the word slot includes:

The present invention also provides a storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the intelligent customer service method as described above.

The intelligent customer service method comprises the following steps:

acquiring current input information of a user;

Further, the method further includes:

Further, in the above method, the dialog behavior type includes:

the determining to fill the word slot includes:

the determining not to fill the word slot includes:

The intelligent customer service method, the intelligent customer service device, the intelligent customer service equipment and the intelligent customer service storage medium of the invention determine a target conversation behavior type to which current input information belongs in each conversation behavior type according to a preset conversation behavior model, determine whether to fill a word slot according to the target conversation behavior type, if the word slot is determined to be filled, obtain elements corresponding to the current input information, fill the elements into the word slot, if the word slot is determined not to be filled, output current response information corresponding to the current input information, so as to obtain input information fed back by a user according to the current response information again. That is to say, by adopting the technical scheme provided by the invention, even though different input information of the user may contain the same key data, the word slot is filled by combining the target dialogue behavior type to which the input information of the user belongs, so that the word slot can be accurately and effectively filled, the phenomenon that the word slot cannot be accurately filled by the machine customer service is reduced, and the practicability of the machine customer service is improved.

Drawings

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

FIG. 1 is a flow chart of a first embodiment of an intelligent customer service method of the present invention;

FIG. 2 is a flowchart of a second embodiment of the intelligent customer service method of the present invention;

FIG. 3 is a schematic structural diagram of an embodiment of an intelligent customer service device according to the present invention;

fig. 4 is a schematic structural diagram of an embodiment of an intelligent customer service device according to the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be described in detail below. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the examples given herein without any inventive step, are within the scope of the present invention.

For example, the workflow of machine customer service will be described by taking activating a credit card as an example, where U is a user, R is a machine customer service:

u: how active my card was.

R: please select your certificate type (clarify certificate type).

U: an identity card.

R: please enter your identification number (clear certificate number).

U：130xxxxxxxxxxxxxxxxxxx。

R: please enter your credit card number (clear card number).

U：620xxxxxxxxxxxxxxxxxxx。

R: please enter your card expiration date.

U: 21 years and 7 months.

R: please enter the 3-bit security code on the back of your card.

U：911。

R: please enter your passcode (clear passcode).

U：0000。

R: the information of the card is approved and whether the card is opened or not is confirmed (whether the card is opened or not is clarified). When the user input is yes, the execution function is activated, and when the user input is no, the execution function is finished.

When the machine service responds to the input information of the user, the machine service needs to accurately fill a word slot, wherein the word slot is a category of element information which needs to be collected from the user reply by the machine service in the task flow, such as 'bill time', 'consumption place', 'consumption type' in an audit scene; and "origin", "destination" and "departure time" in the air ticket booking scenario. It is different from general 'text' information, and describes a kind of 'entity' things in the real world, such as 'thank you' and 'how' do not belong to an element, and 'Shenzhen' and 'Wanyuan' belong to an element.

The word slots can be classified into the following types:

1) there is instance key data

Instances in such word slots may be enumerated and have standard names, requiring normalization processing at element extraction. For example, in the word slot "movie", there is an instance "speed and passion 8", and "speed 8" is an alias of this instance, and after "speed 8" is extracted from the user question, it needs to be normalized to "speed and passion 8".

2) Instance free Key data

The instances in such word slots are not enumeratable, and can be considered as no instances, and no normalization process is needed during element extraction. Instances may not be enumerated as in the word slot "person name" and correspondence cannot be normalized to instances.

3) Semantic type

The word slot has enumerable examples, but the key data can not be directly normalized to one example. Such as word slot "consumption type", sometimes it cannot be judged by a certain word in the user question sentence which consumption type the user inquires, but it is identified which consumption type should be mapped by combining the semantics of the whole sentence.

For multiple rounds of conversation:

a task type scene is composed of an intention inlet and a task flow; after the user asks questions, if the intention of the task type is triggered, the task flow is entered.

There are three core elements in a task flow: the node, the child node jump condition and the word slot are driven by the collection state of the word slot in the whole task. The purpose of one task is to control the task flow by setting the nodes and the jumping conditions between the nodes, thereby successfully collecting the element information and finally completing certain business operation.

One node is a behavior that the robot needs to perform in the task flow. The node types include a start node, a clarify node, and a finish node. In the task flow, different branch flows can be performed according to different collected element values, for example, if the amount of loan the user wants to loan is greater than the maximum loanable amount, the task fails, and the machine customer service needs to clarify the password again, so that the next service flow cannot be entered. If the loan amount is less than the maximum loanable amount, the user continues to be queried for a repayment account for the loan. The element value judgment rule includes: equal to, greater than, less than, containing text, not containing text, elements filled, elements unfilled, etc.

Based on the implementation method of the intelligent customer service, the following situations exist when the machine customer service performs word slot filling: for the input information of the user, the key data contained in the input information of the user is only the consultation information of the user, but not the confirmation information of the user, so that the customer service robot directly fills the word slot by taking the key data as an element, and the machine customer service can fill the word slot in an error way.

For example, U: you are good, and the bills of the last month are given by periods.

R: the service is transacting installments for 10000 yuan of the last month bill, asking how many installments you need to do, and currently supporting 6 th, 12 th and 24 th installments.

U: stage 6

R: transacting a 6-term installments for 10000 yuan of the last month of the bill, asking what way to repay the bill, 1) and other basic information; 2) equal amount of principal

U: what means the amount of the original information.

R: filling the word groove of the repayment mode into equal amount of original information (wrong filling word groove)

In order to solve the above problems in the related art, embodiments of the present invention provide the following technical solutions:

referring to fig. 1, fig. 1 is a flowchart of an embodiment of an intelligent customer service method of the present invention, as shown in fig. 1, the intelligent customer service method of the present embodiment may specifically include the following steps:

100. acquiring current input information of a user;

if the user uses the machine service to perform service handling, the machine service determines and outputs the session content of the current interaction node according to the type of the service and the session content of the previous interaction node, so that the related information input by the user according to the session content of the current interaction node is taken as the obtained current input information of the user, for example, the current input information may be in a text form, a voice form, a video form, and the like.

It should be noted that, because there may be differences between languages corresponding to different regions, in this embodiment, after the current input information of the user is recognized based on the preset language recognition model, the current input information may be converted into a default language type for machine customer service. For example, the language identification model may identify dialects, languages of different countries, and the like.

101. Determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model;

in this embodiment, classification training may be performed in advance for various dialogue behavior types to obtain a preset dialogue behavior model. Specifically, the preset dialogue behavior model is used for matching the current input information of the user, and the dialogue behavior type to which the current input information belongs, namely the target dialogue behavior type, is determined from the dialogue behavior types. Wherein the dialogue act type may include at least one of a provision information act type, a query act type, a complaint act type, a positive reply act type, a negative reply act type, and a to-be-recognized reply act type. For example, the offer information behavior type: i do 12 days; type of query behavior: the handling fees for periods 12 and 24 are slightly worse. Type of query behavior: do not know what you say; complaint behavior type: how high you a fee! Type of aggressive reply behavior: good and no problem; type of negative reply behavior: calculating the bar; the type of the reply behavior to be recognized is as follows: other dialogs.

In a specific implementation process, the current input information can be scored according to each dialogue behavior type respectively according to a preset dialogue behavior model, so that a scoring value corresponding to each dialogue behavior type of the current input information is obtained; and determining the target dialogue behavior type according to the numerical value of the scoring value, specifically, determining the dialogue behavior type with the highest scoring value as the target dialogue behavior type. For example, what the current input information means is "equivalent information. "the current input information is scored according to each dialogue behavior type to obtain the result as the behavior: the scoring value corresponding to the information behavior type is 75, the scoring value corresponding to the inquiry behavior type is 90, the scoring value corresponding to the inquiry behavior type is 80, the scoring value corresponding to the complaint behavior type is 60, the scoring value corresponding to the positive reply behavior type is 70, the scoring value corresponding to the negative reply behavior type is 50, and the scoring value corresponding to the reply behavior type to be identified is 40. At this time, the dialog behavior type to which the current input information belongs may be determined as the query behavior type.

102. Determining whether to fill the word slot or not according to the type of the target conversation behavior; if yes, go to step 103, if no, go to step 104;

if the conversation behavior type of the current input information is the information providing behavior type or the positive reply behavior type, determining to fill the word slot; and if the conversation behavior type to which the current input information belongs is a query behavior type, a complaint behavior type, a negative reply behavior type or a reply behavior type to be identified, determining not to fill the word slot.

For example, the current input information is "I do 12 th. After scoring, the word slot can be determined to be filled by determining that the word slot belongs to the type of the information providing behavior. And when the current input information is 'good and unproblematic', after scoring, determining that the input information belongs to the positive reply behavior type, and determining to fill the word slot. The current input information is "how much worse the commission for 12 th and 24 th periods is. After scoring, it can be determined that it belongs to the type of query behavior, and it is determined that the word slot is not filled. The current input information is 'not knowing what you say', and after scoring, it can be determined that the input information belongs to the type of the questioning behavior, and it is determined that the word slot is not filled. The current input information is "how high a fee is for hands! "after scoring, it can be determined that it belongs to a type of complaint behavior, and it is determined that the word slot is not filled. The current input information is 'calculated bar', and after scoring, the current input information can be determined to belong to a negative reply behavior type, and the word slot is determined not to be filled. The current business is handling staging, the current input information of the user is 'conversation irrelevant to the staging business', and after scoring, the user can determine that the current input information belongs to the reply behavior type to be identified, and determine not to fill the word slot.

In this embodiment, if it is determined that the word slot is filled, step 103 is performed, and if it is determined that the word slot is not filled, step 104 is performed.

103. Acquiring elements corresponding to current input information, and filling the acquired elements into word slots;

in this embodiment, if it is determined that the word slot is filled, the element corresponding to the current input information needs to be obtained according to the dialog behavior type to which the current input information belongs, the element is filled into the word slot, and the dialog content corresponding to the element is obtained, so that the user completes the next round of dialog according to the output dialog content. Specifically, if the dialog behavior type to which the current input information belongs is the information providing behavior type, the element corresponding to the current input information is acquired from the current input information. For example, the current input information is "I do 12 th. "the corresponding element is stage 12, and the term slot is filled with stage 12. And if the conversation behavior type to which the current input information belongs is the positive reply behavior type, acquiring the corresponding element of the current input information according to the current input information and the conversation content before the current input information. For example, if the current input information is "good and trouble-free", it is necessary to acquire the current input information corresponding element from the current input information and/or the dialog content before the current input information. The dialog content before the current input information is: "User: what the commission for period 12 is. And Bot: and dividing 10000 Yuan into 12-period repayment, each-period repayment xxx Yuan and commission charge yyy Yuan. Ask how many wordings you want to do. In this case, it is possible to determine that the current input information corresponding element is the 12 th term from the dialog contents before the current input information, and fill the word slot with the 12 th term. For another example, if the current input information is "good, the commission fee of 12 th period is acceptable", in this case, it may be determined that the element corresponding to the current input information is 12 th period according to the current input information, and the word slot is filled with 12 th period.

104. And outputting current response information corresponding to the current input information.

In this embodiment, if it is determined that the word slot is not filled, current response information corresponding to the current input information needs to be selected from a preset corpus according to semantic understanding information of the current input information. And outputting current response information corresponding to the current input information so as to obtain the input information fed back by the user according to the current response information again.

For example, for a complaint with a high installment fee, the preset response message may be "you can also select 12 or 24, and the more the repayment amount, the lower the repayment pressure, ask you whether to select other installment amounts. Therefore, if the current input information is "divide by 6, how high the fee is! The user is the pot bar, and can determine that semantic understanding information of current input information is that the client considers that the payment is high, the pot is unsatisfactory, and at the moment, the user can select' you can also select 12 th and 24 th from a preset corpus, and the more the payment is, the lower the payment pressure is, and ask whether the user needs to select other installment. "as the current response information and output for the comfort of the customer. The user may further enter "what the 12-period commission is after learning the current response message. ", so that the machine customer attends to the next round of identification.

The intelligent customer service method of the embodiment determines a target dialogue behavior type to which current input information belongs in each dialogue behavior type according to a preset dialogue behavior model by obtaining the current input information of a user, determines whether to fill a word slot according to the target dialogue behavior type, obtains elements corresponding to the current input information if the word slot is determined to be filled, fills the elements into the word slot, outputs current response information corresponding to the current input information if the word slot is determined not to be filled, so as to obtain input information fed back by the user according to the current response information again, namely, by adopting the technical scheme provided by the invention, even if different input information of the user possibly contains the same key data, the word slot is filled by combining the target dialogue behavior type to which the input information of the user belongs, so that the word slot can be filled accurately and effectively, therefore, the phenomenon that the word slot cannot be accurately filled by the machine customer service is reduced, and the practicability of the machine customer service is improved.

The technical solution of the present invention is described below with a specific example, where User is a User and Bot is a customer service system.

Example 1:

user (1): you good, give I a monthly bill an installments! (type of information provision behavior)

Bot (2): the service is transacting the service of 10000 yuan for the last and the same bill, asking how many times you need to transact. Currently, the treatment lasts for 6, 12 and 24 periods.

User (2): what the commission for period 6 is. (query behavior type)

Bot (2): and (4) paying 10000 yuan by 6 seasons, paying xxx yuan per season, and paying 250 yuan in total. Ask how many wordings you want to do.

User (3): good bar, one bar! (aggressive behavior type)

Bot (3): and 6, stage.

Example 2:

Bot (2): the user is transacting the installment for 30000 yuan of the last and last bills, and asks how many instars the user wants to do. Currently, the treatment lasts for 6, 12 and 24 periods.

User (2): what the commission for period 6 is. (query behavior type)

Bot (2): the 300 yuan is divided into 6 yuan payment, 5100 yuan payment per period and 600 yuan commission charge. Ask how many wordings you want to do.

User (3): the cost is as high as 6 times of handling, and the people are the pot-head and do not work! (type of complaint behavior or type of negative behavior)

Bot (3): you can also select 12 th and 24 th, and the more the repayment period number, the smaller the repayment pressure, and ask you whether to select other installment periods.

User (4): what the commission for period 12 is. (query behavior type)

Bot (4): the 300 yuan is paid according to the 12 th period of 30000 yuan, the payment of each period is 2525 yuan, and the commission fee is 300 yuan. Ask how many wordings you want to do.

User (4): 12 th period is cheap and a lot, and one bar is processed! (aggressive behavior type)

Bot (4): and (5) stage 12.

The above examples are only some examples of the present invention, but the present invention is not limited to the above examples, and they are not listed in the present embodiment.

In another embodiment of the present invention, in the step 101, "determining the target dialog behavior type to which the current input information belongs in each dialog behavior type according to a preset dialog behavior model", if the target dialog behavior type is determined based on the numerical value of the score, a scoring error may occur. Therefore, in order to prevent the machine customer service from identifying the current input information incorrectly due to a scoring error of the preset dialogue model, a corresponding threshold may be preset for each dialogue behavior type, where the preset thresholds corresponding to each dialogue behavior type may be the same or different, and the present embodiment is not limited specifically. Thus, after determining the dialog behavior type with the highest score, the score is further compared with a preset threshold value so as to confirm the target dialog behavior type to which the current input information belongs.

Specifically, referring to fig. 2, fig. 2 is a flowchart of a second embodiment of the intelligent customer service method of the present invention, as shown in fig. 2, the intelligent customer service method of the present embodiment describes step 101 in detail based on the embodiment shown in fig. 1, and specifically includes the following steps:

200. determining the dialogue behavior type with the highest scoring value as an initially selected dialogue behavior type;

201. judging whether the highest scoring value is larger than a preset threshold corresponding to the initially selected dialogue behavior type; if yes, go to step 202, if no, go to step 203;

202. determining the initially selected dialogue behavior type as a target dialogue behavior type;

203. outputting semantic understanding information of the current input information;

204. obtaining a confirmation result fed back by a user according to the semantic understanding information;

205. detecting whether the confirmation result fed back by the user is correct or not; if yes, go to step 206, otherwise go to step 207;

206. determining the initially selected dialogue behavior type as a target dialogue behavior type;

207. the user is prompted to enter explicit information for the currently entered information.

For example, 85 is set as a corresponding threshold for each dialog behavior type, the score value obtained by referring to the above example is 90 for the query behavior type, the score value is the highest, the query behavior type may be used as the preliminary selection dialog behavior type, and since the score value corresponding to the query behavior type is greater than the preset threshold, it may be determined that the matching result obtained by the dialog behavior model is correct, and further, it is determined that the preliminary selection dialog behavior type is the target dialog behavior type, that is, the query behavior type is the target dialog behavior type. If the score value corresponding to the inquiry behavior type is 82, the score value is the highest, the inquiry behavior type can be used as a primary selection conversation behavior type, and because the score value corresponding to the inquiry behavior type is smaller than a preset threshold value, a matching result obtained by a conversation behavior model can be determined to be wrong, and semantic understanding information of current input information can be output, so that a user can confirm whether the meaning is expressed by the user and feed back the result to a machine customer service; therefore, a confirmation result fed back by the user according to the semantic understanding information can be obtained; if the confirmation result shows that the conversation behavior type is correct, the initially selected conversation behavior type is determined to be the target conversation behavior type, so that the machine customer service can complete the subsequent flow only by answering yes or no by the user, and the customer service efficiency is improved. And if the confirmation result shows that the input information is wrong, prompting the user to input the explicit information of the current input information so as to re-determine the target conversation behavior type to which the current input information belongs.

Corresponding to the embodiment of the intelligent customer service method provided by the invention, the invention also provides an intelligent customer service device. Referring to fig. 3, fig. 3 is a schematic structural diagram of an intelligent customer service device according to an embodiment of the present invention, and as shown in fig. 3, the intelligent customer service device of the present embodiment includes an obtaining module 10, a determining module 11, and an output module 12;

an obtaining module 10, configured to obtain current input information of a user;

the determining module 11 is configured to determine, according to a preset dialogue behavior model, a target dialogue behavior type to which current input information belongs in each dialogue behavior type; determining whether to fill the word slot or not according to the type of the target conversation behavior;

the obtaining module 10 is further configured to obtain an element corresponding to the current input information and fill the element into the word slot if the determining module 11 determines to fill the word slot;

and the output module 12 is configured to output current response information corresponding to the current input information if the determining module 11 determines that the word slot is not filled, so that the obtaining module 10 obtains the input information fed back by the user according to the current response information again.

The intelligent customer service device of the embodiment determines the target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to the preset dialogue behavior model by obtaining the current input information of the user, determines whether to fill the word slot according to the target dialogue behavior type, obtains the element corresponding to the current input information if determining to fill the word slot, fills the element into the word slot, outputs the current response information corresponding to the current input information if determining not to fill the word slot, so as to obtain the input information fed back by the user according to the current response information again, that is, by adopting the technical scheme provided by the invention, even if different input information of the user possibly contains the same key data, the word slot is filled by combining the target dialogue behavior type to which the input information of the user belongs, so that the word slot can be filled accurately and effectively, therefore, the phenomenon that the word slot cannot be accurately filled by the machine customer service is reduced, and the practicability of the machine customer service is improved.

and if the highest score value is larger than a preset threshold value, determining the initially selected conversation behavior type as the target conversation behavior type.

if the highest score value is smaller than or equal to a preset threshold value, outputting semantic understanding information of the current input information;

obtaining a confirmation result fed back by a user according to the semantic understanding information;

and if the confirmation result shows that the input information is wrong, prompting the user to input the clear information of the current input information.

Further, in the above apparatus, the obtaining module is further configured to:

Further, in the above apparatus, the dialog behavior types include:

a determination module further configured to:

if the conversation behavior type of the current input information is the information providing behavior type or the positive reply behavior type, determining to fill the word slot;

determining not to fill the word slot, comprising:

Further, in the above apparatus, the obtaining module is further configured to:

if the conversation behavior type to which the current input information belongs is the information providing behavior type, acquiring elements from the current input information;

It should be noted that, because the present invention further provides an intelligent customer service device corresponding to the embodiment of the intelligent customer service method provided by the present invention, the embodiment of the intelligent customer service device is not described in detail herein, and reference may be made to the method part of the embodiments for corresponding content.

Fig. 4 is a schematic structural diagram of an embodiment of the intelligent customer service device of the present invention, and as shown in fig. 4, the intelligent customer service device of the present embodiment includes a processor 20 and a memory 21, the processor 20 and the memory 21 are connected through a communication bus:

wherein, the processor 20 is used for calling and executing the program stored in the memory 21;

a memory 21 for storing a program for executing at least the intelligent customer service method of the above embodiment.

Specifically, the intelligent customer service method of the embodiment includes:

acquiring current input information of a user;

determining whether to fill the word slot or not according to the type of the target conversation behavior;

if the word slot is determined to be filled, acquiring elements corresponding to the current input information, and filling the elements into the word slot;

and if the word slot is determined not to be filled, outputting current response information corresponding to the current input information so as to obtain the input information fed back by the user according to the current response information again.

Further, in the above method, determining a target dialogue behavior type in each dialogue behavior type according to a preset dialogue behavior model includes:

Further, in the above method, determining the target dialog behavior type based on the numerical value of the score includes:

Further, the method further includes:

Further, in the above method, before outputting current response information corresponding to the current input information, the method further includes:

Further, in the above method, the dialog behavior types include:

determining to fill a word slot, comprising:

determining not to fill the word slot, comprising:

Further, in the above method, if it is determined to fill the word slot, acquiring an element corresponding to the current input information includes:

The present invention also provides a storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the intelligent customer service method as set forth above.

acquiring current input information of a user;

Further, the method further includes:

Further, in the above method, the dialog behavior types include:

determining to fill a word slot, comprising:

determining not to fill the word slot, comprising:

It is understood that the same or similar parts in the above embodiments may be mutually referred to, and the same or similar parts in other embodiments may be referred to for the content which is not described in detail in some embodiments.

It should be noted that the terms "first," "second," and the like in the description of the present invention are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Further, in the description of the present invention, the meaning of "a plurality" means at least two unless otherwise specified.

Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and alternate implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.

It should be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.

It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.

In addition, functional units in the embodiments of the present invention may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.

The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.

In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.

Claims

1. An intelligent customer service method, comprising:

acquiring current input information of a user;

determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model; wherein the dialog behavior types include: providing at least one of an information behavior type, an inquiry behavior type, a question behavior type, a complaint behavior type, a positive reply behavior type, a negative reply behavior type, and a to-be-recognized reply behavior type;

if the word slot is determined to be filled, acquiring an element corresponding to the current input information, and filling the element into the word slot; if the conversation behavior type to which the current input information belongs is an information providing behavior type or an active reply behavior type, determining to fill the word slot;

if the word slot is determined not to be filled, outputting current response information corresponding to the current input information so as to obtain input information fed back by the user according to the current response information again; and if the conversation behavior type to which the current input information belongs is an inquiry behavior type, a question behavior type, a complaint behavior type, a negative reply behavior type or a reply behavior type to be identified, determining not to fill the word slot.

2. The method according to claim 1, wherein the determining the target dialogue action type in each dialogue action type according to a preset dialogue action model comprises:

3. The method according to claim 2, wherein the determining the target dialog behavior type based on the numerical magnitude of the score value comprises:

4. The method according to claim 2, wherein the determining the target dialog behavior type based on the numerical magnitude of the score value comprises:

5. The method of claim 4, further comprising:

6. The method of claim 5, further comprising:

7. The method of claim 1, wherein before outputting the current response information corresponding to the current input information, the method further comprises:

8. The method according to claim 1, wherein the obtaining the element corresponding to the current input information if the word slot is determined to be filled comprises:

9. An intelligent customer service device is characterized by comprising an acquisition module, a determination module and an output module;

the determining module is used for determining a target dialogue behavior type to which the current input information belongs in each dialogue behavior type according to a preset dialogue behavior model; wherein the dialog behavior types include: providing at least one of an information behavior type, an inquiry behavior type, a question behavior type, a complaint behavior type, a positive reply behavior type, a negative reply behavior type, and a to-be-recognized reply behavior type; determining whether to fill a word slot or not according to the target conversation behavior type;

the obtaining module is further configured to obtain an element corresponding to the current input information and fill the element into the word slot if the determining module determines to fill the word slot; if the conversation behavior type to which the current input information belongs is an information providing behavior type or an active reply behavior type, determining to fill the word slot;

the output module is configured to output current response information corresponding to the current input information if the determining module determines that the word slot is not filled, so that the obtaining module obtains input information fed back by a user according to the current response information again; and if the conversation behavior type to which the current input information belongs is an inquiry behavior type, a question behavior type, a complaint behavior type, a negative reply behavior type or a reply behavior type to be identified, determining not to fill the word slot.

10. An intelligent customer service device, comprising a processor and a memory, wherein the processor and the memory are connected by a communication bus:

the memory for storing the program for performing at least the intelligent customer service method of any of claims 1-8.

11. A storage medium having stored thereon a computer program which, when executed by a processor, carries out the steps of the intelligent customer service method according to any one of claims 1-8.