CN109885664A - A kind of Intelligent dialogue method, robot conversational system, server and storage medium - Google Patents

A kind of Intelligent dialogue method, robot conversational system, server and storage medium Download PDF

Info

Publication number
CN109885664A
CN109885664A CN201910014751.1A CN201910014751A CN109885664A CN 109885664 A CN109885664 A CN 109885664A CN 201910014751 A CN201910014751 A CN 201910014751A CN 109885664 A CN109885664 A CN 109885664A
Authority
CN
China
Prior art keywords
entity
answer
replied
dialogue
data
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.)
Pending
Application number
CN201910014751.1A
Other languages
Chinese (zh)
Inventor
黄友福
肖龙源
蔡振华
李稀敏
刘晓葳
谭玉坤
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xiamen Express Business Information Consulting Co Ltd
Original Assignee
Xiamen Express Business Information Consulting Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xiamen Express Business Information Consulting Co Ltd filed Critical Xiamen Express Business Information Consulting Co Ltd
Priority to CN201910014751.1A priority Critical patent/CN109885664A/en
Publication of CN109885664A publication Critical patent/CN109885664A/en
Pending legal-status Critical Current

Links

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a kind of Intelligent dialogue method, robot conversational system, server and storage mediums, which comprises the artificial customer service dialogue data for obtaining history accumulation extracts the QA question and answer in the dialogue data to deposit QA database;Classification and Entity recognition are carried out to all problems in QA database respectively, and class label and entity vector are increased separately to each problem;According to the classification and entity of problem to be replied, the QA question and answer comprising the same category and entity are filtered out to as candidate data set from QA database;The problem of problem to be replied is matched with the problem in candidate data set, is found with problem similarity highest to be replied, and the corresponding answer of using problem similarity highest to be replied the problem of is as the answer of problem to be replied as exporting.The present invention screens problem domain using textual classification model and NER, enhances the accuracy of FAQ reply, it is ensured that the features such as " capable of asking what what answers " rather than give an irrelevant answer.

Description

A kind of Intelligent dialogue method, robot conversational system, server and storage medium
Technical field
The present invention relates to robots to talk with field, and in particular to a kind of Intelligent dialogue method, robot conversational system, service Device and storage medium.
Background technique
The artificial customer service cost for training a high quality is very high, and there are it is online slowly, the working time is limited equal to be lacked Point.The artificial above-mentioned disadvantage of customer service leads to artificial customer service recruitment, training, using difficulty, causes potential customers to be lost indirectly, gives enterprise Cause huge loss.Therefore, the robot conversational system of a set of high quality is customized just at the urgent need of many enterprises.? Robot conversational system urgent need in modern times doctor U.S.A field, that there is also this is huge for the robot conversational system in doctor U.S. field at present Vacancy.
Since major part Yi Mei enterprise is all also docked using artificial customer service with client at this stage, all add up a large amount of Artificial customer service dialogue data.By taking the beautifying medical hospital of Beijing as an example, adds up artificial customer service dialogue data and reach 2,100,000 left sides It is right.These data cover a large amount of valuable dialogues of the enterprise in each project.
Summary of the invention
It is an object of the invention to a kind of Intelligent dialogue method, machine, for data basis, are proposed with enterprise's artificial customer service dialogue Device people conversational system and storage medium can be quickly applied to medical and beauty treatment fields robot dialogue field.
The present invention provides a kind of Intelligent dialogue methods, include the following:
The artificial customer service dialogue data for obtaining history accumulation extracts the QA question and answer in the dialogue data to deposit QA data Library;
Customer issue is formulated to data to data, and according to the QA question and answer of extraction from the QA question and answer of QA database extraction section Several class scopes and several entity class ranges, and textual classification model and life is respectively trained to for training data with QA question and answer Name entity recognition model;
With trained textual classification model and Named Entity Extraction Model respectively to all problems in QA database into Row classification and Entity recognition, and class label and entity vector are increased separately to each problem, the class label of described problem is Output of the problem Jing Guo trained textual classification model, the entity vector of described problem are that problem is real by trained name The entity sets that body identification model is extracted are constituted;
Obtain problem to be replied from existing customer dialogue, and will be to be replied the problem of respectively by trained text point Class model and Named Entity Extraction Model respectively obtain the generic and entity of problem to be replied;
According to the classification and entity of problem to be replied, the QA question and answer comprising the same category and entity are filtered out from QA database To as candidate data set;
Problem to be replied is matched with the problem in candidate data set, is found highest with problem similarity to be replied Problem, and the corresponding answer of using problem similarity highest to be replied the problem of is as the answer of problem to be replied as output.
Further, the entity vector of described problem is the vector of a regular length, the dimension and entity class of vector Sum is equal, and element value in each dimension is 1 or 0, represents and includes or do not include the entity.
Further, in embodiments of the present invention, which comprises extracting the QA question and answer in dialogue data to preceding, The data of the artificial customer service dialogue data of history accumulation are cleaned or screened.The mode of the data cleansing or screening is logical The mode for crossing keyword search, regularity matching and human assistance screening, rejects meaningless dialogue.It is described meaningless Dialogue comprising it is following any one: give an irrelevant answer, length is too short, unmanned response.
Further, it is preferred that the textual classification model is xgboost textual classification model.
Further, it is preferred that the deterministic process of the similarity is to convert sentence for problem using word2vec model Vector calculates the Euclidean distance of sentence vector as index of similarity using the cosine law.
Further, it is preferred that the Entity recognition includes entity Boundary Recognition and determining entity class.
Correspondingly, the embodiment of the present invention, additionally provides a kind of robot conversational system, comprising: receiving unit, output are single Member, at least one processor and the memory being connect at least one described processor communication;Wherein, the receiving unit For receiving the conversation content of existing customer Yu the robot conversational system;The output unit is for exporting existing customer pair Talk about the answer of problem in content;The memory is stored with the instruction that can be executed by least one described processor, described instruction It is executed by least one described processor, so that at least one described processor executes above-mentioned Intelligent dialogue method.
Correspondingly, the present invention also provides a kind of servers, comprising: at least one processor;And with described at least one The memory of a processor communication connection;Wherein, the memory is stored with the finger that can be executed by least one described processor It enables, described instruction is executed by least one described processor, so that at least one described processor executes above-mentioned Intelligent dialogue side Method.
Correspondingly, the computer readable storage medium is deposited the present invention also provides a kind of computer readable storage medium Computer program is contained, the computer program realizes above-mentioned Intelligent dialogue method and step when being executed by processor.
Intelligent dialogue method, Intelligent dialogue system, server and storage medium provided by the invention, with existing artificial customer service Compare, the present invention this using text classifier (Text-Classifier), name Entity recognition (Named Entity Recognition, NER) and frequently asked questions similarity mode (FAQ Similarity) technology progress robot conversational system structure It builds, the present invention is based on the FAQ conversational systems of artificial customer service data building, are a kind of conversational systems of high-efficient simple, can satisfy The dialogue demand on basis, the method for the present invention screen problem domain using textual classification model and NER, enhance FAQ and return Multiple accuracy, it is ensured that " can ask what what answers " rather than give an irrelevant answer.
Detailed description of the invention
Attached drawing described herein is used to provide to further understand invention, constitutes a part of the invention, the present invention Illustrative embodiments and their description be used to explain the present invention, do not constitute improper limitations of the present invention, it should be apparent that, under Attached drawing in the description of face is some embodiments of the present invention, for those of ordinary skill in the art, is not paying creativeness Under the premise of labour, it is also possible to obtain other drawings based on these drawings.In the accompanying drawings:
Fig. 1 is a kind of schematic illustration of Intelligent dialogue method of the embodiment of the present invention 1;
Fig. 2 is a kind of process step schematic diagram of Intelligent dialogue method of the embodiment of the present invention 1.
Specific embodiment
In order to be clearer and more clear technical problems, technical solutions and advantages to be solved, tie below Drawings and examples are closed, the present invention will be described in further detail.It should be appreciated that specific embodiment described herein is only used To explain the present invention, it is not intended to limit the present invention.Based on the embodiments of the present invention, those of ordinary skill in the art are not having Every other embodiment obtained under the premise of creative work is made, shall fall within the protection scope of the present invention.
Embodiment 1
The present invention provides a kind of Intelligent dialogue methods, and specific implementation principle is as shown in Fig. 1, and the method includes as follows Step, as shown in Fig. 2,
Step S1 obtains the artificial customer service dialogue data of history accumulation, extracts QA question and answer in the dialogue data to depositing Enter QA database, the QA question and answer are to for problem and the corresponding answer of problem.
Step S2 formulates data to data, and according to the QA question and answer of extraction from the QA question and answer of QA database extraction section Several class scopes of customer issue and several entity class ranges, and text classification is respectively trained to for training data with QA question and answer Model and Named Entity Extraction Model;Preferably, in embodiments of the present invention, the QA question and answer from QA database extraction section It is to extract 10% QA question and answer pair of the QA question and answer to total amount of QA database to data.
Step S3, with trained textual classification model and Named Entity Extraction Model respectively to all in QA database Problem carries out classification and Entity recognition, and increases separately class label and entity vector, the classification of described problem to each problem Label is output of the problem Jing Guo trained textual classification model, and the entity vector of described problem is problem by trained The entity sets that Named Entity Extraction Model is extracted are constituted;The entity vector of described problem be regular length to Amount, the dimension of vector is equal with entity class sum, and the element value in each dimension is 1 or 0, represents and includes or do not include this Entity.As shown in table 1-2, example sentence " cut double-edged eyelid and want thousands of? " in, it can be extracted using Named Entity Extraction Model Entity is " position organ " (double-edged eyelid) and " price " (thousands of), then is 1 in " position organ " and the element value of " price ".This hair Bright embodiment carries out classification and Entity recognition to all problems in QA database, obtains entity as shown in Table 1, classification, asks Corresponding table is inscribed, entity vector is as shown in table 2.
1 entity of table, classification, problem correspond to table
2 entity vector of table
Price Time It is preferential Position organ Symptom Material instrument Expert ····
1 0 0 1 0 0 0 ····
Step S4 obtains problem to be replied from existing customer dialogue, and will be to be replied the problem of respectively by training Textual classification model and Named Entity Extraction Model respectively obtain the generic and entity of problem to be replied;The present invention is implemented Obtain that problem to be replied is " it is thousands of to cut double-edged eyelid? " in example, such as existing customer dialogue, it is " to cut double-edged eyelid by problem to be replied It is thousands of? ", obtaining generic by trained textual classification model is " consulting price ", utilizes Named Entity Extraction Model The entity that can be extracted is " position organ " (double-edged eyelid) and " price " (thousands of).
Step S5 is filtered out from QA database comprising the same category and entity according to the classification and entity of problem to be replied QA question and answer to as candidate data set, in the embodiment of the present invention, problem to be replied is " it is thousands of to cut double-edged eyelid? " the problem of finding out It is all the price that this project of double-edged eyelid is cut in inquiry;
Problem to be replied is matched with the problem in candidate data set, is found similar to problem to be replied by step S6 The problem of spending highest, and the corresponding answer of using problem similarity highest to be replied the problem of as the answer of problem to be replied as Output.
Further, in order to more export more accurately answer, this hair states method and includes:
The QA question and answer in dialogue data are being extracted to preceding, the data of the artificial customer service dialogue data of history accumulation are being carried out clear It washes or screens.The mode of data cleansing and screening is screened by keyword search, regularity matching and human assistance Mode, the meaningless dialogues such as rejecting is given an irrelevant answer, length is too short and nobody responds, retains the dialogue data of high quality.
In order to train the textual classification model of high quality, in embodiments of the present invention, it is preferred that the textual classification model For xgboost textual classification model.Xgboost (eXtreme Gradient Boosting) is by Chen Tianqi in gbdt Improved integrated learning approach on the basis of (Gradient Boosting) can be carried out using the multithreading of CPU parallel automatically It calculates, while algorithmically being improved and improving precision.Xgboost is an additivity regression model, is changed by boosting Generation one group of weak learner of construction is put to the vote, to export optimal result.Relative to LR classifier (Logistic Regression Classifier), xgboost classifier has the advantage that: not needing to do the normalization of feature, automatic progress Feature selecting, model interpretation are preferable, are adapted to a variety of loss functions such as SquareLoss, LogLoss etc., utilize Xgboost model can train the text classifier of high quality.
FAQ is the abbreviation of English Frequently Asked Questions, and the Chinese meaning is exactly that " that often asks asks Topic ", or more generically it is called " frequently asked questions and corresponding answer ".FAQ is that the main means of online help are provided in current network, is passed through Better possible often ask is organized to answer questions in advance, publication provides counseling services on webpage for user.It is proposed according to user Problem, most close problem in matching database, and the output answered using the answer for being matched to problem as customer issue.It is common Similarity calculating method be convert a vector for problem using word2vec model, utilize the cosine law to calculate sentence vector Euclidean distance is as index of similarity.In practical applications, the answer for utilizing FAQ matching problem merely, is also frequently present answer It is not in place, situations such as answering wrong field or give an irrelevant answer.
In the methods of the invention, it is preferred that the deterministic process of the similarity is to be turned problem using word2vec model A vector is turned to, calculates the Euclidean distance of sentence vector as index of similarity using the cosine law.It should be noted that wait reply The similarity of the problems in problem and candidate data set compares, and is not limited to institute's example way of the present invention, can also have other Implementation method.
Further, in the method for the present invention, the Entity recognition includes entity Boundary Recognition and determining entity class.Name Entity recognition (Named Entity Recognition, abbreviation NER), also referred to as " proper name identification " refer to and have in identification text There is the entity of certain sense, mainly includes name, place name, mechanism name, proper noun etc..Name Entity recognition be information extraction, Question answering system, syntactic analysis, machine translation, towards Semantic Web metadata mark etc. application fields important foundation work Tool occupies an important position during natural language processing technique moves towards practical.Name Entity recognition generally includes two Point: entity Boundary Recognition and determining entity class.
Embodiment 2
A kind of robot conversational system is provided in the embodiment of the present invention, the robot conversational system includes: to receive list Member, output unit, at least one processor and the memory being connect at least one described processor communication;
Wherein, the receiving unit is used to receive the conversation content of existing customer Yu the robot conversational system;
The output unit is used to export the answer of problem in existing customer conversation content;
The memory is stored with the instruction that can be executed by least one described processor, and described instruction is by described at least one A processor executes, so that at least one described processor executes Intelligent dialogue method and step.The embodiment of the present invention is intelligently right It is identical as embodiment 1 to talk about method and step, repeats no more.
Embodiment 3
A kind of server is provided in the embodiment of the present invention, the server includes at least one processor;And with institute State the memory of at least one processor communication connection;Wherein, be stored with can be by least one described processor for the memory The instruction of execution, described instruction are executed by least one described processor, so that at least one described processor execution intelligence is right Talk about method and step.The memory also stores QA database, the embodiment of the present invention Intelligent dialogue method and step and embodiment 1 It is identical, it repeats no more.
Embodiment 4
A kind of computer readable storage medium, the computer-readable recording medium storage are provided in the embodiment of the present invention There is computer program, Intelligent dialogue method and step is realized when the computer program is executed by processor.It is described computer-readable Storage medium also stores QA database, and the embodiment of the present invention Intelligent dialogue method and step is identical as embodiment 1, no longer superfluous It states.
If realized in the form of software function module the present invention provides a kind of Intelligent dialogue method and as independent Product when selling or using, also can store in a computer readable storage medium.Based on this understanding, of the invention Substantially the part that contributes to existing technology can embody the technical solution of embodiment in the form of software products in other words Out, which is stored in a storage medium, including some instructions are used so that server (can be a People's computer, Cloud Server, the network equipment or be equipment comprising processor etc.) execute described in each embodiment of the present invention The all or part of method.The computer readable storage medium includes but is not limited to read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), USB flash disk, mobile hard disk, magnetic or disk etc. The various media that can store program code.The embodiment of the present invention is not limited to any specific hardware and software and combines.
The above is only a preferred embodiment of the present invention, is merely illustrative of the technical solution of the present invention, it is to be understood that this hair Bright to be not limited to above-described embodiment, enlightenment through the invention, those skilled in the art are in conjunction with known or the prior art, knowledge The modification carried out, or equivalent substitution of some or all of the technical features also should be regarded as in protection of the invention In range.

Claims (10)

1. a kind of Intelligent dialogue method, which is characterized in that the described method includes:
The artificial customer service dialogue data for obtaining history accumulation extracts the QA question and answer in the dialogue data to deposit QA database;
From the QA question and answer of QA database extraction section to data, and it is several according to the QA question and answer of extraction to formulate customer issue to data Class scope and several entity class ranges, and textual classification model and name is respectively trained in fact to for training data with QA question and answer Body identification model;
The all problems in QA database are divided respectively with trained textual classification model and Named Entity Extraction Model Class and Entity recognition, and class label and entity vector are increased separately to each problem, the class label of described problem is problem By the output of trained textual classification model, the entity vector of described problem is that problem is known by trained name entity The entity sets that other model extracts are constituted;
Obtain problem to be replied from existing customer dialogue, and will be to be replied the problem of pass through trained text classification mould respectively Type and Named Entity Extraction Model respectively obtain the generic and entity of problem to be replied;
According to the classification and entity of problem to be replied, the QA question and answer comprising the same category and entity are filtered out to work from QA database For candidate data set;
Problem to be replied is matched with the problem in candidate data set, is found and problem similarity to be replied is highest asks Topic, and the corresponding answer of using problem similarity highest to be replied the problem of is as the answer of problem to be replied as exporting.
2. Intelligent dialogue method according to claim 1, which is characterized in that
The entity vector of described problem is the vector of a regular length, and the dimension of vector is equal with entity class sum, each Element value in dimension is 1 or 0, represents and includes or do not include the entity.
3. Intelligent dialogue method according to claim 1, which is characterized in that
The described method includes:
The QA question and answer in dialogue data are being extracted to preceding, to the data of the artificial customer service dialogue data of history accumulation carry out cleaning or Screening.
4. Intelligent dialogue method according to claim 3, which is characterized in that
The mode of the data cleansing or screening is the side by keyword search, regularity matching and human assistance screening Formula rejects meaningless dialogue.
5. Intelligent dialogue method according to claim 4, which is characterized in that
The meaningless dialogue comprising it is following any one: give an irrelevant answer, length is too short, unmanned response.
6. Intelligent dialogue method according to claim 1, which is characterized in that
The textual classification model is xgboost textual classification model.
7. Intelligent dialogue method according to claim 1, which is characterized in that
The deterministic process of the similarity is to convert a vector for problem using word2vec model, is calculated using the cosine law The Euclidean distance of sentence vector is as index of similarity.
8. a kind of robot conversational system, which is characterized in that the robot conversational system include: receiving unit, output unit, At least one processor and the memory being connect at least one described processor communication;
Wherein, the receiving unit is used to receive the conversation content of existing customer Yu the robot conversational system;
The output unit is used to export the answer of problem in existing customer conversation content;
The memory is stored with the instruction that can be executed by least one described processor, and described instruction is by described at least one It manages device to execute, so that at least one described processor perform claim requires any one of 1 to the 7 Intelligent dialogue method.
9. a kind of server, which is characterized in that the server includes:
At least one processor;And the memory being connect at least one described processor communication;Wherein, the memory is deposited The instruction that can be executed by least one described processor is contained, described instruction is executed by least one described processor, so that institute It states at least one processor perform claim and requires any one of 1 to the 7 Intelligent dialogue method.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists In,
The Intelligent dialogue method and step as described in any one of claim 1 to 7 is realized when the computer program is executed by processor.
CN201910014751.1A 2019-01-08 2019-01-08 A kind of Intelligent dialogue method, robot conversational system, server and storage medium Pending CN109885664A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910014751.1A CN109885664A (en) 2019-01-08 2019-01-08 A kind of Intelligent dialogue method, robot conversational system, server and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910014751.1A CN109885664A (en) 2019-01-08 2019-01-08 A kind of Intelligent dialogue method, robot conversational system, server and storage medium

Publications (1)

Publication Number Publication Date
CN109885664A true CN109885664A (en) 2019-06-14

Family

ID=66925735

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910014751.1A Pending CN109885664A (en) 2019-01-08 2019-01-08 A kind of Intelligent dialogue method, robot conversational system, server and storage medium

Country Status (1)

Country Link
CN (1) CN109885664A (en)

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110196931A (en) * 2019-06-28 2019-09-03 北京蓦然认知科技有限公司 A kind of dialogue generation method and device based on iamge description
CN110738511A (en) * 2019-09-06 2020-01-31 苏宁金融科技(南京)有限公司 Intelligent customer service method and device
CN110795559A (en) * 2019-10-10 2020-02-14 上海易点时空网络有限公司 Data processing method and device for customer service question answering
CN110969021A (en) * 2019-12-23 2020-04-07 竹间智能科技(上海)有限公司 Named entity recognition method, device, equipment and medium in single-round conversation
CN111062612A (en) * 2019-12-17 2020-04-24 联想(北京)有限公司 Construction method of auxiliary processing stream and electronic equipment
CN111340218A (en) * 2020-02-24 2020-06-26 支付宝(杭州)信息技术有限公司 Method and system for training problem recognition model
CN111737424A (en) * 2020-02-21 2020-10-02 北京沃东天骏信息技术有限公司 Question matching method, device, equipment and storage medium
CN112380875A (en) * 2020-11-18 2021-02-19 杭州大搜车汽车服务有限公司 Conversation label tracking method, device, electronic device and storage medium
WO2021051508A1 (en) * 2019-09-18 2021-03-25 平安科技(深圳)有限公司 Robot dialogue generating method and apparatus, readable storage medium, and robot
CN112613304A (en) * 2020-12-17 2021-04-06 合肥讯飞数码科技有限公司 Question answering method, electronic device and storage device
CN112906377A (en) * 2021-03-25 2021-06-04 平安科技(深圳)有限公司 Question answering method and device based on entity limitation, electronic equipment and storage medium
CN113010654A (en) * 2021-03-17 2021-06-22 北京十一贝科技有限公司 Question reply method and device applied to insurance industry, electronic equipment and medium
CN113918703A (en) * 2021-10-26 2022-01-11 未鲲(上海)科技服务有限公司 Intelligent customer service question and answer method, device, server and storage medium
CN115098660A (en) * 2022-06-28 2022-09-23 阿里巴巴(中国)有限公司 Question answering method, device and equipment

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030093263A1 (en) * 2001-11-13 2003-05-15 Zheng Chen Method and apparatus for adapting a class entity dictionary used with language models
CN104598445A (en) * 2013-11-01 2015-05-06 腾讯科技(深圳)有限公司 Automatic question-answering system and method
CN104965992A (en) * 2015-07-13 2015-10-07 南开大学 Text mining method based on online medical question and answer information

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030093263A1 (en) * 2001-11-13 2003-05-15 Zheng Chen Method and apparatus for adapting a class entity dictionary used with language models
CN104598445A (en) * 2013-11-01 2015-05-06 腾讯科技(深圳)有限公司 Automatic question-answering system and method
CN104965992A (en) * 2015-07-13 2015-10-07 南开大学 Text mining method based on online medical question and answer information

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110196931A (en) * 2019-06-28 2019-09-03 北京蓦然认知科技有限公司 A kind of dialogue generation method and device based on iamge description
CN110196931B (en) * 2019-06-28 2021-10-08 北京蓦然认知科技有限公司 Image description-based dialog generation method and device
CN110738511A (en) * 2019-09-06 2020-01-31 苏宁金融科技(南京)有限公司 Intelligent customer service method and device
WO2021051508A1 (en) * 2019-09-18 2021-03-25 平安科技(深圳)有限公司 Robot dialogue generating method and apparatus, readable storage medium, and robot
CN110795559A (en) * 2019-10-10 2020-02-14 上海易点时空网络有限公司 Data processing method and device for customer service question answering
CN111062612A (en) * 2019-12-17 2020-04-24 联想(北京)有限公司 Construction method of auxiliary processing stream and electronic equipment
CN111062612B (en) * 2019-12-17 2022-08-19 联想(北京)有限公司 Construction method of auxiliary processing stream and electronic equipment
CN110969021A (en) * 2019-12-23 2020-04-07 竹间智能科技(上海)有限公司 Named entity recognition method, device, equipment and medium in single-round conversation
CN111737424A (en) * 2020-02-21 2020-10-02 北京沃东天骏信息技术有限公司 Question matching method, device, equipment and storage medium
CN111340218A (en) * 2020-02-24 2020-06-26 支付宝(杭州)信息技术有限公司 Method and system for training problem recognition model
CN112380875A (en) * 2020-11-18 2021-02-19 杭州大搜车汽车服务有限公司 Conversation label tracking method, device, electronic device and storage medium
CN112613304A (en) * 2020-12-17 2021-04-06 合肥讯飞数码科技有限公司 Question answering method, electronic device and storage device
CN113010654A (en) * 2021-03-17 2021-06-22 北京十一贝科技有限公司 Question reply method and device applied to insurance industry, electronic equipment and medium
CN112906377A (en) * 2021-03-25 2021-06-04 平安科技(深圳)有限公司 Question answering method and device based on entity limitation, electronic equipment and storage medium
CN113918703A (en) * 2021-10-26 2022-01-11 未鲲(上海)科技服务有限公司 Intelligent customer service question and answer method, device, server and storage medium
CN115098660A (en) * 2022-06-28 2022-09-23 阿里巴巴(中国)有限公司 Question answering method, device and equipment

Similar Documents

Publication Publication Date Title
CN109885664A (en) A kind of Intelligent dialogue method, robot conversational system, server and storage medium
Tan et al. Improved representation learning for question answer matching
CN112667794A (en) Intelligent question-answer matching method and system based on twin network BERT model
CN109657232A (en) A kind of intension recognizing method
CN105912629A (en) Intelligent question and answer method and device
CN111368540B (en) Keyword information extraction method based on semantic role analysis
CN109325780A (en) A kind of exchange method of the intelligent customer service system in E-Governance Oriented field
CN115470338B (en) Multi-scenario intelligent question answering method and system based on multi-path recall
CN107844558A (en) The determination method and relevant apparatus of a kind of classification information
Abro et al. Natural language processing challenges and issues: A literature review
CN113570380A (en) Service complaint processing method, device and equipment based on semantic analysis and computer readable storage medium
Hayes et al. Toward improved artificial intelligence in requirements engineering: metadata for tracing datasets
CN117875292A (en) Financial knowledge intelligent question-answering method, system, terminal equipment and storage medium
WO2019183517A1 (en) Systems and methods using artificial intelligence to analyze natural language sources based on intelligent agent models
Kortum et al. Dissection of AI job advertisements: A text mining-based analysis of employee skills in the disciplines computer vision and natural language processing
EP4338089A1 (en) System and method of automatic topic detection in text
Rawat et al. Topic modelling of legal documents using NLP and bidirectional encoder representations from transformers
Ferschke et al. A lightly supervised approach to role identification in wikipedia talk page discussions
CN111680501B (en) Query information identification method and device based on deep learning and storage medium
Muangnak et al. The neural network conversation model enables the commonly asked student query agents
Lanyo et al. A Comparative Study of Supervised and Unsupervised Classifiers Utilizing Extractive Text Summarization Techniques to Support Automated Customer Query Question-Answering
Tănăsescu Sentiment Analysis for Human Resources: A Comparative Analysis of Methods That Can Be Used Along with Modern Technologies.
Zhang et al. Business chatbots with deep learning technologies: State-of-the-art, taxonomies, and future research directions
CN110162614B (en) Question information extraction method and device, electronic equipment and storage medium
Dikshit et al. Automating Questions and Answers of Good and Services Tax system using clustering and embeddings of queries

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20190614

RJ01 Rejection of invention patent application after publication