CN113486166B - Construction method, device and equipment of intelligent customer service robot and storage medium - Google Patents
Construction method, device and equipment of intelligent customer service robot and storage medium Download PDFInfo
- Publication number
- CN113486166B CN113486166B CN202110845373.9A CN202110845373A CN113486166B CN 113486166 B CN113486166 B CN 113486166B CN 202110845373 A CN202110845373 A CN 202110845373A CN 113486166 B CN113486166 B CN 113486166B
- Authority
- CN
- China
- Prior art keywords
- dialogue
- information
- customer service
- construction
- service robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0484—Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0484—Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range
- G06F3/04845—Interaction techniques based on graphical user interfaces [GUI] for the control of specific functions or operations, e.g. selecting or manipulating an object, an image or a displayed text element, setting a parameter value or selecting a range for image manipulation, e.g. dragging, rotation, expansion or change of colour
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Human Computer Interaction (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Manipulator (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a construction method of an intelligent customer service robot, which comprises the following steps: acquiring construction information of the intelligent customer service robot, wherein the construction information comprises intention information and answer information corresponding to each dialogue node in a dialogue flow set by a user; screening a target dialogue model from a preset dialogue model library according to the construction information; generating training data according to the construction information; training the target dialogue model based on training data; and constructing the intelligent customer service robot based on the trained target dialogue model. Therefore, the method can automatically complete the construction of the intelligent customer service robot based on the construction information, can automatically complete the complex processes of construction, training and the like of the dialogue model, and can package the complex processes without showing the complex processes to a user, so that the construction method of the intelligent customer service robot with simple construction process and easy operation can be provided, and the method is more beneficial to the wide popularization of the intelligent customer service robot technology. The invention also relates to the technical field of block chains.
Description
Technical Field
The invention relates to the technical field of voice semantics, in particular to a construction method, a device, computer equipment and a storage medium of an intelligent customer service robot.
Background
With the development of artificial intelligence technology, more and more intelligent robots are put into practical use to replace manual work to improve work efficiency. Intelligent customer service robots are one typical application form. Intelligent customer service robots typically require multiple rounds of automated conversational responses to provide intelligent customer service to users. In each automatic dialogue, a natural language understanding module is generally required to analyze dialogue information input by a user, so as to identify user intention in the dialogue information, and then make corresponding response according to the user intention. For example, in a round of automatic dialogue, the natural language understanding module recognizes that the user intends to buy about 5k of notebook computer, and then the corresponding answer can be about 5k of notebook computer model number of each computer brand.
However, the intelligent customer service robot is limited by the technical framework of the existing intelligent customer service robot, and the use threshold of the intelligent customer service robot is always high, so that the intelligent customer service robot technology cannot be widely popularized and used. For example, the automatic dialogue content of the intelligent customer service robots in different fields generally has large difference, so in order to ensure the practical use effect of the intelligent customer service robots, when constructing an intelligent customer service robot, users are generally required to have certain development capability to perform custom development on the intelligent customer service robots, and for example, constructing an intelligent customer service robot also generally requires enough training sample data to complete training on the intelligent customer service robots, which is obviously impractical for some small and medium-sized users (such as middle and small enterprises). Therefore, the construction process of the existing intelligent customer service robot is complex and complicated, and is not beneficial to the wide popularization of the intelligent customer service robot technology.
Disclosure of Invention
The technical problem to be solved by the invention is that the construction process of the existing intelligent customer service robot is complex and complicated, which is not beneficial to the wide popularization of the intelligent customer service robot technology.
In order to solve the technical problems, the first aspect of the invention discloses a construction method of an intelligent customer service robot, which comprises the following steps:
acquiring construction information of the intelligent customer service robot, wherein the construction information comprises intention information and response information corresponding to each dialogue node in a dialogue flow set by a user;
screening a target dialogue model from a preset dialogue model library according to the construction information;
generating training data according to the construction information;
training the target dialogue model based on the training data;
constructing an intelligent customer service robot based on the trained target dialogue model;
the intelligent customer service robot is used for realizing automation of the conversation process according to the intention information and the answer information in the construction information.
The second aspect of the invention discloses a construction device of an intelligent customer service robot, which comprises:
the system comprises an acquisition module, a response module and a control module, wherein the acquisition module is used for acquiring construction information of the intelligent customer service robot, and the construction information comprises intention information and response information corresponding to each dialogue node in a dialogue flow set by a user;
The screening module is used for screening a target dialogue model from a preset dialogue model library according to the construction information;
the generating module is used for generating training data according to the construction information;
the training module is used for training the target dialogue model based on the training data;
the building module is used for building the intelligent customer service robot based on the trained target dialogue model;
the intelligent customer service robot is used for realizing automation of the conversation process according to the intention information and the answer information in the construction information.
A third aspect of the invention discloses a computer device comprising:
a memory storing executable program code;
a processor coupled to the memory;
the processor calls the executable program codes stored in the memory to execute part or all of the steps in the construction method of the intelligent customer service robot disclosed in the first aspect of the invention.
A fourth aspect of the present invention discloses a computer storage medium storing computer instructions for executing part or all of the steps in the method for constructing an intelligent customer service robot disclosed in the first aspect of the present invention when the computer instructions are called.
In the embodiment of the invention, the construction information of the intelligent customer service robot is obtained, the target dialogue model is screened from the preset dialogue model library according to the construction information, the training data is generated according to the construction information, the target dialogue model is trained based on the training data, and finally the intelligent customer service robot is constructed based on the trained target dialogue model, so that the construction of the intelligent customer service robot can be automatically completed based on the construction information, the complex processes such as construction and training of the dialogue model can be automatically completed, and the complex processes are packaged and are not displayed to users, thereby providing the construction method of the intelligent customer service robot with simple construction process and easy operation, and being more beneficial to the wide popularization of the intelligent customer service robot technology.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a construction method of an intelligent customer service robot disclosed in the embodiment of the invention;
fig. 2 is a schematic structural diagram of a construction device of an intelligent customer service robot according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a computer device according to an embodiment of the present invention;
fig. 4 is a schematic structural view of a computer storage medium according to an embodiment of the present invention.
Detailed Description
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The terms first, second and the like in the description and in the claims and in the above-described figures are used for distinguishing between different objects and not necessarily for describing a sequential or chronological order. Furthermore, the terms "comprise" and "have," as well as any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, method, apparatus, article, or article that comprises a list of steps or elements is not limited to only those listed but may optionally include other steps or elements not listed or inherent to such process, method, article, or article.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the invention. The appearances of such phrases in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments.
The invention discloses a construction method, a device, computer equipment and a storage medium of an intelligent customer service robot, which are used for acquiring construction information of the intelligent customer service robot, screening a target dialogue model from a preset dialogue model library according to the construction information, generating training data according to the construction information, training the target dialogue model based on the training data, and finally constructing the intelligent customer service robot based on the trained target dialogue model, so that the construction of the intelligent customer service robot can be automatically completed based on the construction information, the construction, training and other complex processes of the dialogue model can be automatically completed, and the complex processes are packaged and are not displayed to users, thereby providing the construction method of the intelligent customer service robot with simple construction process and easy operation, and being more beneficial to the wide popularization of the intelligent customer service robot technology. The following will describe in detail.
Example 1
Referring to fig. 1, fig. 1 is a flow chart of a method for constructing an intelligent customer service robot according to an embodiment of the present invention. As shown in fig. 1, the construction method of the intelligent customer service robot may include the following operations:
101. and acquiring construction information of the intelligent customer service robot, wherein the construction information comprises intention information and response information corresponding to each dialogue node in a dialogue flow set by a user.
In the step 101, a story line editing interface may be provided at the front end to the user to obtain the building information of the intelligent customer service robot. Specifically, the user can draw a story line through dragging, adding, deleting, modifying, editing a box and the like on the story line editing interface. The user needs to draw dialogue nodes of the story line according to dialogue flows in the actual business scene, and edit intention information and response information corresponding to each dialogue node. After the user completes editing the story line, the building information of the intelligent customer service robot can be generated according to the story line. For example, the following dialog flow: the user: "I want to know the latest hot notebook computer", intelligent customer service robot: "recently popular notebook computers have: hua Shuo magic 14, associative air14, associative pro14, etc. "user: "I want to know the configuration of Hua Shuo magic 14", intelligent customer service robot: the CPU of the Hua Shuo phantom 14 is Ruilong R7 or Ruilong R9, and the display card is GTX2060 or GTX1660 or GTX 1650). In the storyboard drawn by the user according to the conversation process, there are two conversation nodes altogether, wherein the intention information corresponding to the first conversation node is "i want to know the latest popular notebook computer", and the answer information is "the latest popular notebook computer has: hua Shuo magic 14, associativity air14 and associativity pro14", the intention information corresponding to the second dialogue node is that I want to know the configuration of Hua Shuo magic 14, the CPU of the answer information Hua Shuo magic 14 is that of Ruilong R7 or Ruilong R9, and the display card is that of GTX2060 or GTX1660 or GTX 1650).
102. And screening a target dialogue model from a preset dialogue model library according to the construction information.
In the step 102, a common natural language recognition model may be pre-stored in a dialogue model library, for example, the dialogue model library may include a bi-directional double-layer lstm+seq2seq model, a bert model, an N-gram model, an NNLM model, and the like. In practical application, different natural language recognition models generally have certain differences in recognition effects on different dialogue contents, so that an intelligent customer service robot can be constructed by screening a proper model from a dialogue model library according to the actual dialogue contents (a specific screening process is described later), and the recognition effect of the intelligent customer service robot on the dialogue contents can be ensured.
103. And generating training data according to the construction information.
In step 103, training data of the target dialogue model is generated according to the dialogue content in the construction information, and then the training data is used for training the target dialogue model, so that the target dialogue model can have the capability of identifying the dialogue content in the construction information, and the constructed intelligent customer service robot can automatically complete the dialogue flow. Optionally, when generating the training data, the training data may be further amplified by using a data enhancement algorithm (a specific amplification process will be described later), so that it is beneficial to ensure that there is enough training data to train the target dialogue model, ensure the training effect of the target dialogue model, and solve the problem of lack of training sample data for small and medium-sized users.
104. Training the target dialogue model based on the training data.
105. Constructing an intelligent customer service robot based on the trained target dialogue model; the intelligent customer service robot is used for realizing automation of the conversation process according to the intention information and the answer information in the construction information.
After training of the target dialogue model is completed, corresponding service interfaces can be generated according to the trained target dialogue model, and then the service interfaces are connected through the front end interface, so that the construction of the intelligent customer service robot can be completed, and intelligent customer service is improved for users. Specifically, when the user needs to consult, the dialogue content to be consulted can be input in the front-end interface, then after the dialogue content is input by the user, the front-end interface inputs the dialogue content input by the user into the target dialogue model after training is completed, the target dialogue model returns the corresponding response content to the front end after analyzing the dialogue content input by the user, and finally the returned response content is displayed to the user in the front-end interface, so that automation of a dialogue process can be completed, and construction of the intelligent customer service robot is completed. Therefore, in the construction process of the intelligent customer service robot, a user only needs to draw a story line on an operation level, wherein the user only needs to relate to simple text editing, box dragging and other contents, does not need to know complex processes such as model construction, training and the like, the complex processes such as model construction, training and the like are packaged and are not displayed to the user any more, and the complex processes such as model construction, training and the like can be automatically completed, so that the construction process of the intelligent customer service robot with simple construction can be provided, and a middle-small-scale user can also freely construct a proper intelligent customer service robot according to own business requirements.
It can be seen that, by implementing the embodiment of the invention, the construction information of the intelligent customer service robot is obtained, the target dialogue model is screened from the preset dialogue model library according to the construction information, the training data is generated according to the construction information, the target dialogue model is trained based on the training data, and finally the intelligent customer service robot is constructed based on the trained target dialogue model, so that the construction of the intelligent customer service robot can be automatically completed based on the construction information, the complex processes such as construction and training of the dialogue model can be automatically completed, and the complex processes are packaged and are not displayed to users, thereby providing the construction method of the intelligent customer service robot with simple construction process and easy operation, and being more beneficial to the wide popularization of the intelligent customer service robot technology.
In an optional embodiment, the screening the target dialogue model from the preset dialogue model library according to the construction information includes:
extracting keywords in the construction information through a preset keyword extraction algorithm;
calculating a score corresponding to each dialogue model in the dialogue model library according to the keywords;
and screening the dialogue model meeting preset scoring conditions from the dialogue model library to serve as a target dialogue model according to the score of each dialogue model in the dialogue model library.
In this alternative embodiment, the keywords in the build information may be extracted by a keyword extraction algorithm such as TF-IDF, topic-model, textrank, rake, etc. For example, the intention information in the construction information is "i want to know the latest popular notebook computer", and the answer information is "the latest popular notebook computer has: hua Shuo magic 14, associative air14 and associative pro14", the keywords extracted from the intention information may be" latest popular "," notebook ", and the keywords extracted from the answer information may be" Hua Shuo magic 14"," associative air14 "and" associative pro14". In practical application, different dialogue models have a certain difference in recognition effect on different keywords, for example, the recognition effect of the bert model on the letter type keywords may be better, while the recognition effect on the Chinese character type keywords may be worse, the score of the bert model corresponding to the letter type keywords may be set to a higher value, and the score of the bert model corresponding to the Chinese character type keywords may be set to a lower value. And then summing the scores corresponding to each keyword in each dialogue model to obtain the scores corresponding to each dialogue model, wherein a specific summation process is described later. For example, the number of keywords is two, and the scores corresponding to the keywords "latest popular" and "notebook" in the bidirectional double-layer lstm+seq2seq model are 5 and 6 respectively, so that the scores corresponding to the bidirectional double-layer lstm+seq2seq model obtained by final calculation are 5+6=11. And finally, screening the target dialogue model from the dialogue model library according to the score of each dialogue model in the dialogue model library, for example, selecting the dialogue model with the highest score in the dialogue model library as the target dialogue model.
Therefore, according to the implementation of the optional embodiment, the key words in the construction information are extracted, the score corresponding to each dialogue model in the dialogue model library is calculated according to the key words, and finally the target dialogue model is screened out from the dialogue model library according to the score of each dialogue model in the dialogue model library, so that the appropriate target dialogue model can be screened out from the dialogue model library to construct the intelligent customer service robot, and the use effect of the finally constructed intelligent customer service robot is better.
In an optional embodiment, the calculating the score corresponding to each session model in the session model library according to the keyword includes:
inquiring a preset scoring table to obtain the corresponding score of each keyword in each dialogue model, wherein the scoring table records the corresponding score of each keyword in each dialogue model;
and calculating the score corresponding to each dialogue model according to the scores corresponding to all the keywords corresponding to the dialogue model.
In this alternative embodiment, the preset scoring table may be as follows:
the score table is recorded with the score corresponding to each keyword in each dialogue model, the score corresponding to each keyword in each dialogue model can be obtained by inquiring the score table, then the score corresponding to each dialogue model can be calculated according to the scores corresponding to all keywords corresponding to each dialogue model, and the specific calculation process is described later. The scores of the keywords in the scoring table in the dialogue model can be preset by a developer according to own experience or actual use effect.
Therefore, according to the implementation of the alternative embodiment, the score corresponding to each keyword in each dialogue model is obtained by querying the preset score table, and then the score corresponding to the dialogue model is calculated according to the scores corresponding to all keywords corresponding to each dialogue model, so that the score corresponding to each dialogue model in the dialogue model library can be calculated according to the keywords, and then the appropriate target dialogue model can be screened out from the dialogue model library according to the score corresponding to each dialogue model, thereby being more beneficial to the construction of the intelligent customer service robot.
In an optional embodiment, the calculating the score corresponding to each dialogue model according to the scores corresponding to all the keywords corresponding to the dialogue model includes:
inquiring a preset keyword weight table to obtain a keyword weight corresponding to each keyword corresponding to each dialogue model;
calculating the weighted sum of scores corresponding to all keywords corresponding to each dialogue model to serve as the score corresponding to the dialogue model;
the scoring weight corresponding to each keyword corresponding to each dialogue model is the keyword weight corresponding to the keyword. In this alternative embodiment, the preset keyword weight table may be as follows:
| Keyword(s) | Weight value |
| Latest hot door | 10 |
| Associative air14 | 5 |
| …… | …… |
The weight corresponding to the keyword in the keyword weight table may be preset by a developer according to his experience or actual use effect, for example, the keyword "latest trending" usually appears in the intent information, and correctly identifies the keyword "latest trending" in the intent information, which plays an important role in correctly replying to the consultation of the user for the intelligent customer service robot, so that the keyword "latest trending" may be set to a larger value, so that the identification accuracy of the target dialogue model screened according to the score on the keyword "latest trending" is higher, thereby being more beneficial to ensuring the use effect of the finally constructed intelligent customer service robot. When calculating the score corresponding to each dialogue model based on the keyword weight and the score corresponding to each keyword corresponding to each dialogue model, a weighted sum of the keyword weight and the score corresponding to each keyword corresponding to each dialogue model may be calculated as the score of the dialogue model. For example, according to the contents in the scoring table and the keyword weight table, when the keyword includes only "latest popular" and "associated air14", the score of the bi-directional double-layer lstm+seq2seq model is 8×10+8×5=120, the score of the bert model is 5×10+8×5=90, and the score of the n-gram model is 8×10+5×5=105.
Therefore, according to the implementation of the alternative embodiment, the keyword weight corresponding to each keyword corresponding to each dialogue model is obtained by querying the preset keyword weight table, and then the score corresponding to the dialogue model is calculated based on the keyword weight and the score corresponding to each keyword corresponding to each dialogue model, so that the calculated score corresponding to each dialogue model is more accurate, and the proper target dialogue model can be screened from the dialogue model library according to the score corresponding to each dialogue model, thereby being more beneficial to the construction of the intelligent customer service robot.
In an alternative embodiment, the generating training data according to the construction information includes:
judging whether the construction information comprises preset data enhancement keywords or not;
when the data enhancement keywords are included in the construction information, replacing the data enhancement keywords in the construction information with the expansion keywords corresponding to the data enhancement keywords to generate training data, wherein each data enhancement keyword is preset with at least one corresponding expansion keyword.
For small and medium-scale users, how to acquire enough training sample data when constructing an intelligent customer service robot is an important problem, so that when training data is generated according to construction information, the training data can be subjected to data enhancement to realize the augmentation of the training data, thereby being beneficial to ensuring the sufficiency of the training data. Specifically, the data enhancement keywords in the build information may be replaced with the expansion keywords corresponding to the data enhancement keywords to generate the training data. For example, the construction information is "about an associative xxxK" notebook configuration nvidiap10000", wherein" nvidiap10000 "is a data enhancement keyword, and the data enhancement keyword" nvidiap10000 "is replaced by an expansion keyword" nvidiaa6000 "and" nividiaRTX4000", so that a training data" about an associative xxxK "notebook configuration nvidiaa6000" and "about an associative xxxK" notebook configuration nividiaRTX4000 "can be obtained, and the training data can be amplified. Wherein, the data enhancement keywords are preset by the user.
Therefore, when the optional embodiment is implemented, the training data is generated by replacing the data enhancement keywords in the construction information with the expansion keywords corresponding to the data enhancement keywords when the training data is generated according to the construction information, so that the training data can be amplified, the training of the target dialogue model can be ensured to be carried out by enough training data, and the smooth construction of the intelligent customer service robot is ensured.
In an alternative embodiment, the method further comprises:
acquiring construction adjustment information of the intelligent customer service robot, wherein the construction adjustment information comprises intention information and response information corresponding to dialogue nodes to be newly added in a dialogue flow;
generating construction adjustment training data based on the construction adjustment information;
training the target dialogue model in the intelligent customer service robot again based on the construction adjustment training data so as to add the newly added dialogue node in the construction adjustment information into a dialogue flow realized by the intelligent customer service robot.
After the intelligent customer service robot is built, the user can also adjust the story line, and new dialogue nodes are continuously added into the story line. Specifically, a story line adjustment interface may be provided at the front end for the user, where the user may input intent information and answer information corresponding to the newly added dialog node, so that the obtaining of the building adjustment information may be achieved. And then, generating construction adjustment training data according to the construction adjustment information in the same training data generation mode, and finally, training the target dialogue model again by using the newly generated construction adjustment training data, so that the target dialogue model has the capability of identifying dialogue contents in the construction adjustment information, and the intelligent customer service robot can automatically complete the newly added dialogue nodes. For example, the current robot can realize only 2 dialogue nodes, if one dialogue node needs to be added again, the story line needs to be adjusted, training data is generated again, and the model is trained again, so that the model has the capability of identifying the content of the new dialogue node. The original dialogue model is trained according to the information of the first two dialogue nodes, so the dialogue model generally does not have the capability of identifying the content of the new dialogue node, and if the new node is added, the training is needed again, so that the dialogue model has the capability of identifying the content of the new dialogue node.
It can be seen that, implementing this alternative embodiment, after completing the construction of the intelligent customer service robot, the construction adjustment information of the intelligent customer service robot is obtained, then the construction adjustment training data is generated based on the construction adjustment information, and finally the target dialogue model in the intelligent customer service robot is trained again based on the construction adjustment training data, so that the target dialogue model has the capability of identifying dialogue content in the construction adjustment information, thereby realizing the new function of dialogue nodes, enabling the application of the constructed intelligent customer service robot to be more flexible, and being better adapted to more application scenarios.
In an optional embodiment, the obtaining the construction adjustment information of the intelligent customer service robot includes:
acquiring historical dialogue information of the intelligent customer service robot;
judging whether the history dialogue information contains non-response dialogue information or not;
pushing the non-response dialogue information to a preset management terminal corresponding to the intelligent customer service robot when judging that the non-response dialogue information exists in the history dialogue information;
receiving correct response information corresponding to the non-response dialogue information returned by the management terminal, wherein the correct response information is correct response information made by a user of the management terminal according to the non-response dialogue information;
And taking the non-response dialogue information and the correct response information as intention information and response information corresponding to dialogue nodes to be newly added in a dialogue flow, so as to obtain construction adjustment information of the intelligent customer service robot.
After the intelligent customer service robot is constructed, dialogue information (i.e. no-response dialogue information) which cannot be normally and automatically responded by the intelligent customer service robot can be extracted from dialogue information (i.e. history dialogue information) generated by the intelligent customer service robot in actual use. And pushing the dialogue information which cannot be normally and automatically replied by the intelligent customer service robots to a preset management terminal (such as a personal computer of a developer), manually replying the dialogue information which cannot be normally and automatically replied by the intelligent customer service robots by a user (such as the developer) of the management terminal, returning corresponding correct response information, and finally taking the non-response dialogue information and the correct response information as intention information and response information corresponding to a new dialogue node to be added in a dialogue flow so as to obtain construction adjustment information. Therefore, the conversation process of the intelligent customer service robot can be adjusted according to the non-response conversation information of the intelligent customer service robot, so that the automatic perfection of the conversation process can be realized, and the use effect of the intelligent customer service robot is improved. In the historical dialogue information of the intelligent customer service robot, normal dialogue information usually exists in a form of one-to-one answer between the robot and a user, and if two or more continuous dialogue information sent by the user appear, the dialogue information is the dialogue information which indicates that the intelligent customer service robot is very likely to be unable to automatically answer normally, at this time, the dialogue information can be extracted as non-answer dialogue information.
It can be seen that, by implementing this alternative embodiment, the non-answer dialogue information is extracted from the history dialogue information of the intelligent customer service robot, then the correct answer information corresponding to the non-answer dialogue information is formed by using the manual answer non-answer dialogue information, and finally the non-answer dialogue information and the correct answer information are used as the intention information and the answer information corresponding to the dialogue node to be newly added in the dialogue process to obtain the construction adjustment information, so that the dialogue process of the intelligent customer service robot can be adjusted according to the non-answer dialogue information of the intelligent customer service robot, the automatic improvement of the dialogue process is realized, and the use effect of the intelligent customer service robot is improved.
Optionally, it is also possible to: and uploading the construction information of the intelligent customer service robot in the construction method of the intelligent customer service robot to a blockchain.
Specifically, the construction information of the intelligent customer service robot is obtained by running the construction method of the intelligent customer service robot, and is used for recording the construction condition of the intelligent customer service robot, for example, intention information and answer information contained in the construction information, description information of a target dialogue model, training data and the like. The safety and the fairness and transparency of the intelligent customer service robot to the user can be ensured by uploading the construction information of the intelligent customer service robot to the blockchain. The user can download the construction information of the intelligent customer service robot from the blockchain so as to verify whether the construction information of the intelligent customer service robot of the construction method of the intelligent customer service robot is tampered. The blockchain referred to in this example is a novel mode of application for computer technology such as distributed data storage, point-to-point transmission, consensus mechanisms, encryption algorithms, and the like. The Blockchain (Blockchain), which is essentially a decentralised database, is a string of data blocks that are generated by cryptographic means in association, each data block containing a batch of information of network transactions for verifying the validity of the information (anti-counterfeiting) and generating the next block. The blockchain may include a blockchain underlying platform, a platform product services layer, an application services layer, and the like.
Example two
Referring to fig. 2, fig. 2 is a schematic structural diagram of a construction device of an intelligent customer service robot according to an embodiment of the present invention. As shown in fig. 2, the construction apparatus of the intelligent customer service robot may include:
an obtaining module 201, configured to obtain construction information of the intelligent customer service robot, where the construction information includes intention information and answer information corresponding to each dialogue node in a dialogue flow set by a user;
the screening module 202 is configured to screen a target dialogue model from a preset dialogue model library according to the construction information;
a generating module 203, configured to generate training data according to the construction information;
a training module 204, configured to train the target dialogue model based on the training data;
a building module 205, configured to build an intelligent customer service robot based on the trained target dialogue model;
the intelligent customer service robot is used for realizing automation of the conversation process according to the intention information and the answer information in the construction information.
In an alternative embodiment, the specific manner of screening the target dialogue model from the preset dialogue model library by the screening module 202 according to the construction information is:
Extracting keywords in the construction information through a preset keyword extraction algorithm;
calculating a score corresponding to each dialogue model in the dialogue model library according to the keywords;
and screening the dialogue model meeting preset scoring conditions from the dialogue model library to serve as a target dialogue model according to the score of each dialogue model in the dialogue model library.
In an alternative embodiment, the specific manner of calculating the score corresponding to each session model in the session model library by the screening module 202 according to the keyword is:
inquiring a preset scoring table to obtain the corresponding score of each keyword in each dialogue model, wherein the scoring table records the corresponding score of each keyword in each dialogue model;
and calculating the score corresponding to each dialogue model according to the scores corresponding to all the keywords corresponding to the dialogue model.
In an alternative embodiment, the specific manner of calculating the score corresponding to each dialogue model according to the score corresponding to all the keywords corresponding to the dialogue model by the screening module 202 is:
inquiring a preset keyword weight table to obtain a keyword weight corresponding to each keyword corresponding to each dialogue model;
Calculating the weighted sum of scores corresponding to all keywords corresponding to each dialogue model to serve as the score corresponding to the dialogue model;
the scoring weight corresponding to each keyword corresponding to each dialogue model is the keyword weight corresponding to the keyword. In an alternative embodiment, the specific manner of generating the training data by the generating module 203 according to the construction information is:
judging whether the construction information comprises preset data enhancement keywords or not;
when the data enhancement keywords are included in the construction information, replacing the data enhancement keywords in the construction information with the expansion keywords corresponding to the data enhancement keywords to generate training data, wherein each data enhancement keyword is preset with at least one corresponding expansion keyword.
In an optional embodiment, the obtaining module 201 is further configured to obtain construction adjustment information of the intelligent customer service robot, where the construction adjustment information includes intention information and answer information corresponding to a dialog node to be newly added in a dialog flow;
the generating module 203 is further configured to generate construction adjustment training data based on the construction adjustment information;
The training module 204 is further configured to train the target session model in the intelligent customer service robot again based on the building adjustment training data, so as to add the session node newly added in the building adjustment information to a session process implemented by the intelligent customer service robot.
In an alternative embodiment, the specific manner of the obtaining module 201 obtaining the building adjustment information of the intelligent customer service robot is:
acquiring historical dialogue information of the intelligent customer service robot;
judging whether the history dialogue information contains non-response dialogue information or not;
pushing the non-response dialogue information to a preset management terminal corresponding to the intelligent customer service robot when judging that the non-response dialogue information exists in the history dialogue information;
receiving correct response information corresponding to the non-response dialogue information returned by the management terminal, wherein the correct response information is correct response information made by a user of the management terminal according to the non-response dialogue information;
and taking the non-response dialogue information and the correct response information as intention information and response information corresponding to dialogue nodes to be newly added in a dialogue flow, so as to obtain construction adjustment information of the intelligent customer service robot.
For a specific description of the construction apparatus of the intelligent customer service robot, reference may be made to a specific description of the construction method of the intelligent customer service robot, and for avoiding repetition, a detailed description is omitted herein.
Example III
Referring to fig. 3, fig. 3 is a schematic structural diagram of a computer device according to an embodiment of the invention. As shown in fig. 3, the computer device may include:
a memory 301 storing executable program code;
a processor 302 connected to the memory 301;
the processor 302 invokes the executable program code stored in the memory 301 to perform the steps in the method for constructing the intelligent customer service robot according to the first embodiment of the present invention.
Example IV
Referring to fig. 4, an embodiment of the present invention discloses a computer storage medium 401, where the computer storage medium 401 stores computer instructions for executing steps in the method for constructing an intelligent customer service robot according to the embodiment of the present invention when the computer instructions are called.
The apparatus embodiments described above are merely illustrative, wherein the modules illustrated as separate components may or may not be physically separate, and the components shown as modules may or may not be physical, i.e., may be located in one place, or may be distributed over a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above detailed description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course by means of hardware. Based on such understanding, the foregoing technical solutions may be embodied essentially or in part in the form of a software product that may be stored in a computer-readable storage medium including Read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), programmable Read-Only Memory (Programmable Read-Only Memory, PROM), erasable programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), one-time programmable Read-Only Memory (OTPROM), electrically erasable programmable Read-Only Memory (EEPROM), compact disc Read-Only Memory (Compact Disc Read-Only Memory, CD-ROM) or other optical disc Memory, magnetic disc Memory, tape Memory, or any other medium that can be used for computer-readable carrying or storing data.
Finally, it should be noted that: the embodiment of the invention discloses a method, a device, a computer device and a storage medium for constructing an intelligent customer service robot, which are disclosed by the embodiment of the invention only as a preferred embodiment of the invention, and are only used for illustrating the technical scheme of the invention, but not limiting the technical scheme; although the invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art will understand that; the technical scheme recorded in the various embodiments can be modified or part of technical features in the technical scheme can be replaced equivalently; such modifications and substitutions do not depart from the spirit and scope of the corresponding technical solutions.
Claims (7)
1. The method for constructing the intelligent customer service robot is characterized by comprising the following steps of:
acquiring construction information of the intelligent customer service robot, wherein the construction information comprises intention information and response information corresponding to each dialogue node in a dialogue flow set by a user;
screening a target dialogue model from a preset dialogue model library according to the construction information;
generating training data according to the construction information;
training the target dialogue model based on the training data;
Constructing an intelligent customer service robot based on the trained target dialogue model;
the intelligent customer service robot is used for realizing the automation of the conversation process according to the intention information and the answer information in the construction information;
the screening the target dialogue model from a preset dialogue model library according to the construction information comprises the following steps:
extracting keywords in the construction information through a preset keyword extraction algorithm;
calculating a score corresponding to each dialogue model in the dialogue model library according to the keywords;
screening dialogue models meeting preset scoring conditions from the dialogue model library to serve as target dialogue models according to the score of each dialogue model in the dialogue model library;
the calculating the score corresponding to each dialogue model in the dialogue model library according to the keyword comprises the following steps:
inquiring a preset scoring table to obtain the corresponding score of each keyword in each dialogue model, wherein the scoring table records the corresponding score of each keyword in each dialogue model;
calculating the scores corresponding to the dialogue models according to the scores corresponding to all the keywords corresponding to each dialogue model;
The step of calculating the score corresponding to each dialogue model according to the scores corresponding to all the keywords corresponding to each dialogue model comprises the following steps:
inquiring a preset keyword weight table to obtain a keyword weight corresponding to each keyword corresponding to each dialogue model;
calculating the weighted sum of scores corresponding to all keywords corresponding to each dialogue model to serve as the score corresponding to the dialogue model;
the scoring weight corresponding to each keyword corresponding to each dialogue model is the keyword weight corresponding to the keyword.
2. The method for constructing an intelligent customer service robot according to claim 1, wherein generating training data according to the construction information comprises:
judging whether the construction information comprises preset data enhancement keywords or not;
when the data enhancement keywords are included in the construction information, replacing the data enhancement keywords in the construction information with the expansion keywords corresponding to the data enhancement keywords to generate training data, wherein each data enhancement keyword is preset with at least one corresponding expansion keyword.
3. The method for constructing an intelligent customer service robot according to claim 1, wherein the method further comprises:
acquiring construction adjustment information of the intelligent customer service robot, wherein the construction adjustment information comprises intention information and response information corresponding to dialogue nodes to be newly added in a dialogue flow;
generating construction adjustment training data based on the construction adjustment information;
training the target dialogue model in the intelligent customer service robot again based on the construction adjustment training data so as to add the newly added dialogue node in the construction adjustment information into a dialogue flow realized by the intelligent customer service robot.
4. A method for constructing an intelligent customer service robot according to claim 3, wherein the acquiring the construction adjustment information of the intelligent customer service robot comprises:
acquiring historical dialogue information of the intelligent customer service robot;
judging whether the history dialogue information contains non-response dialogue information or not;
pushing the non-response dialogue information to a preset management terminal corresponding to the intelligent customer service robot when judging that the non-response dialogue information exists in the history dialogue information;
Receiving correct response information corresponding to the non-response dialogue information returned by the management terminal, wherein the correct response information is correct response information made by a user of the management terminal according to the non-response dialogue information;
and taking the non-response dialogue information and the correct response information as intention information and response information corresponding to dialogue nodes to be newly added in a dialogue flow, so as to obtain construction adjustment information of the intelligent customer service robot.
5. A construction device of an intelligent customer service robot, the device comprising:
the system comprises an acquisition module, a response module and a control module, wherein the acquisition module is used for acquiring construction information of the intelligent customer service robot, and the construction information comprises intention information and response information corresponding to each dialogue node in a dialogue flow set by a user;
the screening module is used for screening a target dialogue model from a preset dialogue model library according to the construction information, and the specific mode of screening the target dialogue model from the preset dialogue model library according to the construction information is as follows: extracting keywords in the construction information through a preset keyword extraction algorithm; calculating a score corresponding to each dialogue model in the dialogue model library according to the keywords; screening dialogue models meeting preset scoring conditions from the dialogue model library to serve as target dialogue models according to the score of each dialogue model in the dialogue model library; the specific way of calculating the score corresponding to each dialogue model in the dialogue model library according to the keywords is as follows: inquiring a preset scoring table to obtain the corresponding score of each keyword in each dialogue model, wherein the scoring table records the corresponding score of each keyword in each dialogue model; calculating the scores corresponding to the dialogue models according to the scores corresponding to all the keywords corresponding to each dialogue model; calculating the score corresponding to each dialogue model according to the scores corresponding to all the keywords corresponding to the dialogue model, comprising: inquiring a preset keyword weight table to obtain a keyword weight corresponding to each keyword corresponding to each dialogue model; calculating the weighted sum of scores corresponding to all keywords corresponding to each dialogue model to serve as the score corresponding to the dialogue model; wherein, the scoring weight corresponding to each keyword corresponding to each dialogue model is the keyword weight corresponding to the keyword;
The generating module is used for generating training data according to the construction information;
the training module is used for training the target dialogue model based on the training data;
and the construction module is used for constructing the intelligent customer service robot based on the trained target dialogue model.
6. A computer device, the computer device comprising:
a memory storing executable program code;
a processor coupled to the memory;
the processor invokes the executable program code stored in the memory to perform the method of constructing an intelligent customer service robot as claimed in any one of claims 1-4.
7. A computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the method of constructing an intelligent customer service robot according to any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110845373.9A CN113486166B (en) | 2021-07-26 | 2021-07-26 | Construction method, device and equipment of intelligent customer service robot and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110845373.9A CN113486166B (en) | 2021-07-26 | 2021-07-26 | Construction method, device and equipment of intelligent customer service robot and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113486166A CN113486166A (en) | 2021-10-08 |
| CN113486166B true CN113486166B (en) | 2023-09-05 |
Family
ID=77943770
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110845373.9A Active CN113486166B (en) | 2021-07-26 | 2021-07-26 | Construction method, device and equipment of intelligent customer service robot and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113486166B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116303979A (en) * | 2023-05-18 | 2023-06-23 | 北京及时语智能科技有限公司 | Intelligent customer service dialogue model construction method and system |
| CN117370535B (en) * | 2023-12-05 | 2024-04-16 | 粤港澳大湾区数字经济研究院(福田) | Training method of medical dialogue model, medical query method, device and equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110472035A (en) * | 2019-08-26 | 2019-11-19 | 杭州城市大数据运营有限公司 | A kind of intelligent response method, apparatus, computer equipment and storage medium |
| CN111400465A (en) * | 2020-02-25 | 2020-07-10 | 支付宝(杭州)信息技术有限公司 | Generation method and device of customer service robot, electronic equipment and medium |
| WO2020148355A1 (en) * | 2019-01-17 | 2020-07-23 | Koninklijke Philips N.V. | A system for multi-perspective discourse within a dialog |
-
2021
- 2021-07-26 CN CN202110845373.9A patent/CN113486166B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020148355A1 (en) * | 2019-01-17 | 2020-07-23 | Koninklijke Philips N.V. | A system for multi-perspective discourse within a dialog |
| CN110472035A (en) * | 2019-08-26 | 2019-11-19 | 杭州城市大数据运营有限公司 | A kind of intelligent response method, apparatus, computer equipment and storage medium |
| CN111400465A (en) * | 2020-02-25 | 2020-07-10 | 支付宝(杭州)信息技术有限公司 | Generation method and device of customer service robot, electronic equipment and medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113486166A (en) | 2021-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6799574B2 (en) | Method and device for determining satisfaction with voice dialogue | |
| WO2022095380A1 (en) | Ai-based virtual interaction model generation method and apparatus, computer device and storage medium | |
| CN107992585A (en) | Universal tag method for digging, device, server and medium | |
| CN108021934B (en) | Method and device for recognizing multiple elements | |
| CN109857846B (en) | Method and device for matching user question and knowledge point | |
| CN113486166B (en) | Construction method, device and equipment of intelligent customer service robot and storage medium | |
| CN111625632A (en) | Question-answer pair recommendation method, device, equipment and storage medium | |
| CN109284502B (en) | Text similarity calculation method and device, electronic equipment and storage medium | |
| US20230121711A1 (en) | Content augmentation with machine generated content to meet content gaps during interaction with target entities | |
| US20220138770A1 (en) | Method and apparatus for analyzing sales conversation based on voice recognition | |
| CN112925911B (en) | Complaint classification method based on multi-modal data and related equipment thereof | |
| CN109087205A (en) | Prediction technique and device, the computer equipment and readable storage medium storing program for executing of public opinion index | |
| Windiatmoko et al. | Developing facebook chatbot based on deep learning using rasa framework for university enquiries | |
| CN111177307A (en) | Test scheme and system based on semantic understanding similarity threshold configuration | |
| CN107368499B (en) | Client label modeling and recommending method and device | |
| Windiatmoko et al. | Developing FB chatbot based on deep learning using RASA framework for university enquiries | |
| CN113707299A (en) | Auxiliary diagnosis method and device based on inquiry session and computer equipment | |
| CN109903122A (en) | House prosperity transaction information processing method, device, equipment and storage medium | |
| TWI674517B (en) | Information interaction method and device | |
| CN116521850A (en) | Interaction method and device based on reinforcement learning | |
| CN111931503B (en) | Information extraction method and device, equipment and computer readable storage medium | |
| CN112528638A (en) | Abnormal object identification method and device, electronic equipment and storage medium | |
| CN111723559B (en) | Real-time information extraction method and device | |
| CN113342948A (en) | Intelligent question and answer method and device | |
| CN117370512A (en) | Method, device, equipment and storage medium for replying to dialogue |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |