CN111078846A - Multi-turn dialog system construction method and system based on business scene - Google Patents
Multi-turn dialog system construction method and system based on business scene Download PDFInfo
- Publication number
- CN111078846A CN111078846A CN201911166714.9A CN201911166714A CN111078846A CN 111078846 A CN111078846 A CN 111078846A CN 201911166714 A CN201911166714 A CN 201911166714A CN 111078846 A CN111078846 A CN 111078846A
- Authority
- CN
- China
- Prior art keywords
- user
- model
- intention
- round
- rule matching
- 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
Links
- 238000010276 construction Methods 0.000 title claims abstract description 31
- 238000000034 method Methods 0.000 claims abstract description 74
- 238000004458 analytical method Methods 0.000 claims abstract description 30
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000005516 engineering process Methods 0.000 claims abstract description 11
- 238000011156 evaluation Methods 0.000 claims description 21
- 238000012360 testing method Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000013145 classification model Methods 0.000 claims description 4
- 238000013135 deep learning Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000013461 design Methods 0.000 claims description 3
- 230000000007 visual effect Effects 0.000 claims description 3
- 238000012706 support-vector machine Methods 0.000 description 6
- 238000007477 logistic regression Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000007635 classification algorithm Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
Images
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
- 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
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- 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
- G06F16/35—Clustering; Classification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Human Computer Interaction (AREA)
- Machine Translation (AREA)
Abstract
The invention relates to a multi-round dialogue system construction method based on a business scene, which comprises the following steps: designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node; respectively matching the linguistic data and the keywords of each user node in the multi-turn conversation process according to a text classification technology and a rule matching training intention judgment model and a rule matching model; after the user node receives the user statement, analyzing and predicting the user statement through the intention judgment model and the rule matching model respectively, and outputting an intention analysis result; the nodes in the multi-round conversation process can be adjusted according to the intention analysis result, the conversation process is convenient, flexible and fast to establish, meanwhile, the multi-round conversation process can be optimized fast and accurately, and conversation intelligence is improved.
Description
Technical Field
The invention relates to the technical field of intelligent conversation, in particular to a method and a system for constructing a multi-round conversation system based on a service scene.
Background
The intelligent dialogue system is widely applied to the fields of intelligent customer service, robots, automobiles, navigation and the like, most of the currently adopted multi-turn dialogue models are more suitable for the situation of living chatting and do not consider the situation based on the business scene, and the multi-turn dialogue system is not flexible enough and the process is difficult to construct.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a method for building a multi-turn dialog system based on a service scene and a system for building a multi-turn dialog system based on a service scene, aiming at the above defects in the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows:
a multi-round dialogue system construction method based on a business scene is constructed, and the realization method comprises the following steps:
the first step is as follows: designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node;
the second step is that: respectively matching the linguistic data and the keywords of each user node in the multi-turn conversation process according to a text classification technology and a rule matching training intention judgment model and a rule matching model;
the third step: and after the user node receives the user statement, analyzing and predicting the user statement respectively through the intention judging model and the rule matching model, and outputting an intention analysis result.
The invention relates to a multi-round dialogue system construction method based on a business scene, wherein the second step implementation method comprises the following steps:
cleaning the corpus, and removing special symbols and punctuation marks in the corpus;
using a Chinese word segmentation tool to segment the cleaned corpus;
removing stop words, removing words without actual meanings in the corpus, and obtaining corpus words;
vectorizing and expressing the corpus words by using a word2vec model to be used as a training set of an intention judgment model;
training and storing an intention judgment model;
the operator finds out the user statement expression under the specific service scene, and organizes the rule matching strategy according to the rules to construct a rule matching model.
The invention relates to a multi-round dialogue system construction method based on a business scene, wherein the third step implementation method comprises the following steps:
setting a high threshold and a low threshold for the intent decision model;
obtaining classification categories and corresponding confidence degrees of user sentences through an intention judgment model;
if the confidence degree is greater than a high threshold value, namely the classification result has very high reliability, directly outputting an intention analysis result;
if the confidence coefficient is lower than the high threshold and higher than the low threshold, the rule matching model is used, and if the rule matching model is successfully matched, the result is directly output; if the matching fails, comparing the confidence coefficient with a low threshold, and if the confidence coefficient is greater than the low threshold, outputting an intention analysis result;
and if the confidence coefficient is smaller than a low threshold value, namely the confidence coefficient of the classification result is very low, the decision is rejected, and the analysis result of the mixed model is not given.
The invention relates to a multi-round dialogue system construction method based on a business scene, which comprises the following steps: and performing performance evaluation on the multi-turn conversation process according to the intention analysis result, and adjusting the intention judgment model and the rule matching model according to the evaluation result.
The invention relates to a multi-round dialogue system construction method based on a business scene, wherein the fourth step implementation method comprises the following steps:
acquiring a batch of test data, wherein the test data comprises user statements and manually calibrated user intentions;
executing the third step by the user sentences in batch to obtain intention recognition results, and finally calculating the intention recognition accuracy, error rate and unrecognized rate of the user sentences;
and marking error recognition and unrecognized user sentences, adding the error recognition and unrecognized user sentences into the corpus and the keywords of the corresponding category, if the error recognition and unrecognized user sentences do not belong to any category, creating new user nodes, and redesigning the conversation process.
The invention relates to a business scene-based multi-turn conversation system construction method, wherein in the first step, a visual interface is adopted to design the multi-turn conversation process.
The multi-round dialogue system construction method based on the business scene is characterized in that the classification technology in the second step is an SVM, LR, NB or deep learning classification model.
The invention relates to a multi-round dialogue system construction method based on a business scene, wherein the fourth step further comprises the following steps:
and processing a certain number of error recognition user statements and unrecognized user statements by using a clustering algorithm, and adding the error recognition user statements and the unrecognized user statements into the corpora and the keywords of the corresponding categories in batches.
A multi-round dialogue system construction system based on a business scene is disclosed, which comprises a flow configuration module, a training module and an identification module;
the process configuration module is used for designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node;
the training module is used for respectively matching the linguistic data and the keywords of each user node in the multi-round conversation process with a training intention judgment model and a rule matching model according to a text classification technology and a rule;
and the identification module is used for analyzing and predicting the user statement through the intention judgment model and the rule matching model respectively after the user node receives the user statement, and outputting an intention analysis result.
The invention relates to a multi-round dialogue system construction system based on a business scene, which further comprises an evaluation module;
and the evaluation module is used for carrying out performance evaluation on the multi-turn conversation process according to the intention analysis result and adjusting the intention judgment model and the rule matching model according to the evaluation result.
The invention has the beneficial effects that: designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node; the linguistic data and the keywords of each user node in the multi-turn conversation process are respectively matched with the training intention judgment model and the rule matching model according to the text classification technology and the rules, after the user nodes receive user sentences, the user sentences are analyzed and predicted through the intention judgment model and the rule matching model respectively, intention analysis results are output, the nodes in the multi-turn conversation process can be adjusted according to the intention analysis results, the conversation process is convenient and flexible to establish and fast, meanwhile, the multi-turn conversation process can be optimized fast and accurately, and conversation intelligence is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the present invention will be further described with reference to the accompanying drawings and embodiments, wherein the drawings in the following description are only part of the embodiments of the present invention, and for those skilled in the art, other drawings can be obtained without inventive efforts according to the accompanying drawings:
FIG. 1 is a flow chart of a method for building a multi-turn dialog system based on a service scenario according to a preferred embodiment of the present invention;
FIG. 2 is a flowchart of model training of a multi-turn dialog system construction method based on business scenarios according to a preferred embodiment of the present invention;
FIG. 3 is a flowchart illustrating intent determination of a multi-turn dialog architecture construction method based on business scenarios according to a preferred embodiment of the present invention;
FIG. 4 is a flowchart illustrating the evaluation of the multi-turn dialog architecture construction method based on the service scenario according to the preferred embodiment of the present invention;
FIG. 5 is a schematic diagram of a hierarchical mixed-intent decision model of a multi-turn dialog system construction method based on a business scenario according to a preferred embodiment of the present invention;
FIG. 6 is a schematic block diagram of a mixed intent decision model training process of a multi-turn dialog system construction method based on business scenarios according to a preferred embodiment of the present invention;
FIG. 7 is a schematic block diagram of a working process of a mixed-intent decision model of a multi-turn dialog system construction method based on a business scenario according to a preferred embodiment of the present invention;
FIG. 8 is a flow test data table of a multi-turn dialog architecture construction method based on service scenarios according to a preferred embodiment of the present invention;
FIG. 9 is a schematic block diagram of a multi-turn dialog architecture construction system based on business scenarios in accordance with a preferred embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention, and it is obvious that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without inventive step, are within the scope of the present invention.
The method for constructing a multi-turn dialog system based on a service scenario in the preferred embodiment of the present invention is shown in fig. 1, and is implemented by referring to fig. 2 to 8 as follows:
s01: designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node;
s02: respectively matching the linguistic data and the keywords of each user node in the multi-turn conversation process according to a text classification technology and a rule matching training intention judgment model and a rule matching model;
s03: after the user node receives the user statement, analyzing and predicting the user statement through an intention judgment model and a rule matching model respectively, and outputting an intention analysis result;
the nodes in the multi-round conversation process can be adjusted according to the intention analysis result, the conversation process is convenient, flexible and fast to establish, meanwhile, the multi-round conversation process can be optimized fast and accurately, and conversation intelligence is improved.
Preferably, the second step is realized by the following steps:
s021: cleaning the corpus, and removing special symbols and punctuation marks in the corpus;
s022: using a Chinese word segmentation tool to segment the cleaned corpus;
s023: removing stop words, removing words without actual meanings in the corpus, and obtaining corpus words;
s024: vectorizing and expressing the corpus words by using a word2vec model to be used as a training set of an intention judgment model;
s025: training and storing an intention judgment model;
s026: finding out user statement expression under a specific service scene by operators, organizing rule matching strategies according to the rules, and constructing a rule matching model;
at the core of the multi-turn dialog system, a hierarchical Mixed Intention Decision Model (MIDM) is trained, and as shown in fig. 5 and 6, a plurality of mixed intention decision models are included in the multi-turn dialog system. The structure of the hierarchical model corresponds to the designed conversation process one by one, and different models are obtained by training different linguistic data and key words and are used for identifying different intentions;
a Mixed Intention Decision Model (MIDM) is established according to the text classification technique and the rule matching as shown in fig. 7, thereby outputting an intention analysis result.
The Chinese word segmentation tool can adopt jieba and the like.
Preferably, the third step implementation method is as follows:
s031: setting a high threshold and a low threshold for the intent decision model;
s032: obtaining classification categories and corresponding confidence degrees of user sentences through an intention judgment model;
s033: if the confidence degree is greater than a high threshold value, namely the classification result has very high reliability, directly outputting an intention analysis result;
if the confidence coefficient is lower than the high threshold and higher than the low threshold, the rule matching model is used, and if the rule matching model is successfully matched, the result is directly output; if the matching fails, comparing the confidence coefficient with a low threshold, and if the confidence coefficient is greater than the low threshold, outputting an intention analysis result;
if the confidence coefficient is smaller than the low threshold value, namely the confidence coefficient of the classification result is very low, the decision is refused, and the analysis result of the mixed model is not given;
and identifying the user intention according to the dialogue sentences of the user, giving corresponding replies, and realizing natural and smooth dialogue between the user and the voice robot. After being processed, user sentences enter a rule matching model and an intention judging model respectively, the rule matching model and the intention judging model can analyze and predict the user sentences, and finally, a final decision result is given through an intention decision module, and an intention analysis result of the user is output.
Preferably, the method further comprises the step of S04: and performing performance evaluation on the multi-round conversation process according to the intention analysis result, adjusting an intention judgment model and a rule matching model according to the evaluation result, and optimizing the model.
Preferably, the fourth step is realized by:
s041: acquiring a batch of test data, wherein the test data comprises user statements and manually calibrated user intentions;
s042: executing the third step by the user sentences in batch to obtain intention recognition results, and finally calculating the intention recognition accuracy, error rate and unrecognized rate of the user sentences;
s043: and marking error recognition and unrecognized user sentences, adding the error recognition and unrecognized user sentences into the corpus and the keywords of the corresponding category, if the error recognition and unrecognized user sentences do not belong to any category, creating new user nodes, and redesigning the conversation process.
Preferably, in the first step, a visual interface is adopted to design a plurality of turns of conversation processes; convenient operation and high flexibility.
Preferably, the classification technology in the second step is SVM, LR, NB or deep learning classification model; the deep learning classification model may be selected from TextCNN, TextRNN, etc.
Preferably, the fourth step further comprises the method of:
and processing a certain number of error recognition user statements and unrecognized user statements by using a clustering algorithm, and adding the error recognition user statements and the unrecognized user statements into the corpora and the keywords of the corresponding categories in batches.
A multi-round dialogue system construction system based on a business scene comprises a flow configuration module 1, a training module 2 and an identification module 3;
the flow configuration module 1 is used for designing a multi-round conversation flow according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation flow, and configuring a title and a conversation in each machine node;
the training module 2 is used for matching the linguistic data and the keywords of each user node in the multi-round conversation process with a training intention judgment model and a rule matching model respectively according to a text classification technology and a rule;
the recognition module 3 is used for analyzing and predicting the user statements through the intention judgment model and the rule matching model respectively after the user nodes receive the user statements, and outputting intention analysis results;
designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node; the language material and the key words of each user node in the multi-turn conversation process are respectively matched with the training intention judgment model and the rule matching model according to the text classification technology and the rules, after the user nodes receive user sentences, the user sentences are respectively analyzed and predicted through the intention judgment model and the rule matching model, intention analysis results are output, the nodes in the multi-turn conversation process can be adjusted according to the intention analysis results, the conversation process is convenient and flexible to establish and fast, meanwhile, the multi-turn conversation process can be optimized fast and accurately, and conversation intelligence is improved.
Preferably, the device further comprises an evaluation module 4;
the evaluation module 4 is used for carrying out performance evaluation on the multi-turn conversation process according to the intention analysis result and adjusting an intention judgment model and the rule matching model according to the evaluation result;
the module mainly realizes two functions, one of which evaluates the performance of a multi-turn dialogue system, and the other two of which automatically adjust the model according to the evaluation result and optimize the model. The flow evaluation method comprises the steps of firstly obtaining a batch of test data, obtaining intention recognition results of user sentences through a mixed intention judgment model in batch, and finally calculating the intention recognition accuracy, error rate and unrecognized rate of the user sentences. And the model adjusting scheme is to mark error recognition and unrecognized user sentences, add the error recognition and unrecognized user sentences to the corpus and the keywords of the corresponding category, create new user nodes if the error recognition and unrecognized user sentences do not belong to any category, and redesign the conversation process. The model adjusting scheme is used for expanding the model identification range and improving the model identification accuracy.
Description of related terms:
jieba: the Chinese word segmentation tool is used for segmenting a Chinese character sequence into a single word by word segmentation;
SVM: a Support Vector Machine (SVM) is a classification algorithm, and both linear classification and nonlinear classification are supported. The SVM is effective in solving the classification problem and the regression problem of the high-dimensional features, and still has a good effect when the feature dimension is greater than the sample number; when the sample size is not mass data, the classification accuracy is high, and the generalization capability is strong;
LR: logistic Regression (LR) is a classification algorithm that can handle binary classification and multivariate classification problems;
NB: naive Bayes (NB, Naive Bayes) classifiers are a series of simple probability classifiers based on the Bayes theorem under the strong independence among assumed features, and are often applied to text classification tasks;
word2 vec: word2Vec is a model for unsupervised learning of semantic knowledge from a large corpus of text, which is heavily used in natural language processing. Word2Vec is actually to represent semantic information of words in a Word vector mode through learning a text, namely, words similar in semanteme are close to each other in the space through an embedding space;
TextCNN: TextCNN is an algorithm for classifying texts using a convolutional neural network, which includes a convolutional layer and a pooling layer;
TextRNN: the TextRNN is a recurrent neural network used for text processing, and is capable of expressing context information better than the TextCNN.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.
Claims (10)
1. A multi-round dialogue system construction method based on a business scene is characterized in that the realization method is as follows:
the first step is as follows: designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node;
the second step is that: respectively matching the linguistic data and the keywords of each user node in the multi-turn conversation process according to a text classification technology and a rule matching training intention judgment model and a rule matching model;
the third step: and after the user node receives the user statement, analyzing and predicting the user statement respectively through the intention judging model and the rule matching model, and outputting an intention analysis result.
2. The multi-round dialog system construction method based on the service scenario as claimed in claim 1, wherein the second step implementation method is:
cleaning the corpus, and removing special symbols and punctuation marks in the corpus;
using a Chinese word segmentation tool to segment the cleaned corpus;
removing stop words, removing words without actual meanings in the corpus, and obtaining corpus words;
vectorizing and expressing the corpus words by using a word2vec model to be used as a training set of an intention judgment model;
training and storing an intention judgment model;
the operator finds out the user statement expression under the specific service scene, and organizes the rule matching strategy according to the rules to construct a rule matching model.
3. The multi-round dialog system construction method based on the service scenario as claimed in claim 1, wherein the third step implementation method is:
setting a high threshold and a low threshold for the intent decision model;
obtaining classification categories and corresponding confidence degrees of user sentences through an intention judgment model;
if the confidence degree is greater than a high threshold value, namely the classification result has very high reliability, directly outputting an intention analysis result;
if the confidence coefficient is lower than the high threshold and higher than the low threshold, the rule matching model is used, and if the rule matching model is successfully matched, the result is directly output; if the matching fails, comparing the confidence coefficient with a low threshold, and if the confidence coefficient is greater than the low threshold, outputting an intention analysis result;
and if the confidence coefficient is smaller than a low threshold value, namely the confidence coefficient of the classification result is very low, the decision is rejected, and the analysis result of the mixed model is not given.
4. A method for building a multi-turn dialog system based on service scenarios according to any of claims 1 to 3, characterized in that it comprises a fourth step of: and performing performance evaluation on the multi-turn conversation process according to the intention analysis result, and adjusting the intention judgment model and the rule matching model according to the evaluation result.
5. The multi-turn dialog system construction method based on the service scenario as claimed in claim 4, wherein the fourth step is implemented as follows:
acquiring a batch of test data, wherein the test data comprises user statements and manually calibrated user intentions;
executing the third step by the user sentences in batch to obtain intention recognition results, and finally calculating the intention recognition accuracy, error rate and unrecognized rate of the user sentences;
and marking error recognition and unrecognized user sentences, adding the error recognition and unrecognized user sentences into the corpus and the keywords of the corresponding category, if the error recognition and unrecognized user sentences do not belong to any category, creating new user nodes, and redesigning the conversation process.
6. A multi-turn dialog architecture construction method based on business scenarios according to any of claims 1-3, characterized in that in the first step, a visual interface is used to design the multi-turn dialog flow.
7. The multi-round dialog system construction method based on business scenarios according to any of claims 1-3, characterized in that the classification technique in the second step is SVM, LR, NB or deep learning classification model.
8. The multi-turn dialog architecture construction method based on business scenario of claim 5, characterized in that the fourth step further comprises the method of:
and processing a certain number of error recognition user statements and unrecognized user statements by using a clustering algorithm, and adding the error recognition user statements and the unrecognized user statements into the corpora and the keywords of the corresponding categories in batches.
9. A multi-round dialogue system construction system based on business scenes, the multi-round dialogue system construction method based on business scenes according to any one of claims 1 to 8, characterized by comprising a process configuration module, a training module and an identification module;
the process configuration module is used for designing a multi-round conversation process according to different service scenes, configuring linguistic data and keywords for each user node in the multi-round conversation process, and configuring a title and a conversation in each machine node;
the training module is used for respectively matching the linguistic data and the keywords of each user node in the multi-round conversation process with a training intention judgment model and a rule matching model according to a text classification technology and a rule;
and the identification module is used for analyzing and predicting the user statement through the intention judgment model and the rule matching model respectively after the user node receives the user statement, and outputting an intention analysis result.
10. The business scenario based multi-turn dialog architecture system of claim 9, further comprising an evaluation module;
and the evaluation module is used for carrying out performance evaluation on the multi-turn conversation process according to the intention analysis result and adjusting the intention judgment model and the rule matching model according to the evaluation result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911166714.9A CN111078846A (en) | 2019-11-25 | 2019-11-25 | Multi-turn dialog system construction method and system based on business scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911166714.9A CN111078846A (en) | 2019-11-25 | 2019-11-25 | Multi-turn dialog system construction method and system based on business scene |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111078846A true CN111078846A (en) | 2020-04-28 |
Family
ID=70311558
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911166714.9A Pending CN111078846A (en) | 2019-11-25 | 2019-11-25 | Multi-turn dialog system construction method and system based on business scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111078846A (en) |
Cited By (26)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111611368A (en) * | 2020-05-22 | 2020-09-01 | 北京百度网讯科技有限公司 | Method and device for backtracking common scene conversation in multi-turn conversation |
| CN111625637A (en) * | 2020-05-27 | 2020-09-04 | 苏州思必驰信息科技有限公司 | Intention training and analyzing method and device |
| CN111683182A (en) * | 2020-06-11 | 2020-09-18 | 中国银行股份有限公司 | Service node processing method and system |
| CN111767219A (en) * | 2020-06-24 | 2020-10-13 | 中国工商银行股份有限公司 | Method and device for checking intelligent customer service conversation process |
| CN111858856A (en) * | 2020-07-23 | 2020-10-30 | 海信电子科技(武汉)有限公司 | Multi-round search type chatting method and display equipment |
| CN111883111A (en) * | 2020-07-30 | 2020-11-03 | 平安国际智慧城市科技股份有限公司 | Dialect training processing method and device, computer equipment and readable storage medium |
| CN111916111A (en) * | 2020-07-20 | 2020-11-10 | 中国建设银行股份有限公司 | Intelligent voice outbound method and device with emotion, server and storage medium |
| CN111933118A (en) * | 2020-08-17 | 2020-11-13 | 苏州思必驰信息科技有限公司 | Method and device for optimizing voice recognition and intelligent voice dialogue system applying same |
| CN111985206A (en) * | 2020-07-17 | 2020-11-24 | 联想(北京)有限公司 | Corpus understanding method and equipment |
| CN112035632A (en) * | 2020-08-21 | 2020-12-04 | 惠州市德赛西威汽车电子股份有限公司 | Preferred distribution method and system suitable for multi-conversation robot collaboration task |
| CN112131357A (en) * | 2020-08-21 | 2020-12-25 | 国网浙江省电力有限公司杭州供电公司 | User intention identification method and device based on intelligent dialogue model |
| CN112150694A (en) * | 2020-08-12 | 2020-12-29 | 国网浙江省电力有限公司营销服务中心 | Intelligent voice electric charge payment urging system and method |
| CN112199486A (en) * | 2020-10-21 | 2021-01-08 | 中国电子科技集团公司第十五研究所 | Task type multi-turn conversation method and system for office scene |
| CN112487144A (en) * | 2020-11-30 | 2021-03-12 | 北京天润融通科技股份有限公司 | Multi-round conversation robot intention hit optimization method and device |
| CN112489639A (en) * | 2020-11-26 | 2021-03-12 | 北京百度网讯科技有限公司 | Audio signal processing method, device, system, electronic equipment and readable medium |
| CN112487170A (en) * | 2020-12-14 | 2021-03-12 | 南京三眼精灵信息技术有限公司 | Scene configuration-oriented human-computer interaction dialogue robot system |
| CN112612462A (en) * | 2020-12-29 | 2021-04-06 | 平安科技（深圳）有限公司 | Method and device for adjusting phone configuration, electronic equipment and storage medium |
| CN112686674A (en) * | 2020-12-25 | 2021-04-20 | 科讯嘉联信息技术有限公司 | Customer service conversation work order summarizing method |
| CN112765331A (en) * | 2020-12-31 | 2021-05-07 | 杭州摸象大数据科技有限公司 | Dialogue knowledge template construction method and device, computer equipment and storage medium |
| CN112784024A (en) * | 2021-01-11 | 2021-05-11 | 软通动力信息技术(集团)股份有限公司 | Man-machine conversation method, device, equipment and storage medium |
| CN112905749A (en) * | 2021-03-12 | 2021-06-04 | 电子科技大学 | Task-type multi-turn dialogue method based on intention-slot value rule tree |
| CN112988963A (en) * | 2021-02-19 | 2021-06-18 | 平安科技（深圳）有限公司 | User intention prediction method, device, equipment and medium based on multi-process node |
| CN112989003A (en) * | 2021-04-01 | 2021-06-18 | 网易（杭州）网络有限公司 | Intention recognition method, device, processing equipment and medium |
| CN113139059A (en) * | 2021-05-13 | 2021-07-20 | 八维(杭州)科技有限公司 | Intention grading method based on man-machine conversation |
| CN113761183A (en) * | 2020-07-30 | 2021-12-07 | 北京汇钧科技有限公司 | Intention recognition method and intention recognition device |
| CN117389541A (en) * | 2023-12-13 | 2024-01-12 | 中国人民解放军国防科技大学 | Configuration system and device for generating template based on dialogue retrieval |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107316643A (en) * | 2017-07-04 | 2017-11-03 | 科大讯飞股份有限公司 | Voice interactive method and device |
| CN108763568A (en) * | 2018-06-05 | 2018-11-06 | 北京玄科技有限公司 | The management method of intelligent robot interaction flow, more wheel dialogue methods and device |
| CN109508376A (en) * | 2018-11-23 | 2019-03-22 | 四川长虹电器股份有限公司 | It can online the error correction intension recognizing method and device that update |
| CN109871446A (en) * | 2019-01-31 | 2019-06-11 | 平安科技（深圳）有限公司 | Rejection method for identifying, electronic device and storage medium in intention assessment |
| CN110147445A (en) * | 2019-04-09 | 2019-08-20 | 平安科技（深圳）有限公司 | Intension recognizing method, device, equipment and storage medium based on text classification |
| CN110209791A (en) * | 2019-06-12 | 2019-09-06 | 百融云创科技股份有限公司 | It is a kind of to take turns dialogue intelligent speech interactive system and device more |
-
2019
- 2019-11-25 CN CN201911166714.9A patent/CN111078846A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107316643A (en) * | 2017-07-04 | 2017-11-03 | 科大讯飞股份有限公司 | Voice interactive method and device |
| CN108763568A (en) * | 2018-06-05 | 2018-11-06 | 北京玄科技有限公司 | The management method of intelligent robot interaction flow, more wheel dialogue methods and device |
| CN109508376A (en) * | 2018-11-23 | 2019-03-22 | 四川长虹电器股份有限公司 | It can online the error correction intension recognizing method and device that update |
| CN109871446A (en) * | 2019-01-31 | 2019-06-11 | 平安科技（深圳）有限公司 | Rejection method for identifying, electronic device and storage medium in intention assessment |
| CN110147445A (en) * | 2019-04-09 | 2019-08-20 | 平安科技（深圳）有限公司 | Intension recognizing method, device, equipment and storage medium based on text classification |
| CN110209791A (en) * | 2019-06-12 | 2019-09-06 | 百融云创科技股份有限公司 | It is a kind of to take turns dialogue intelligent speech interactive system and device more |
Cited By (37)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111611368B (en) * | 2020-05-22 | 2023-07-04 | 北京百度网讯科技有限公司 | Method and device for backtracking public scene dialogue in multiple rounds of dialogue |
| CN111611368A (en) * | 2020-05-22 | 2020-09-01 | 北京百度网讯科技有限公司 | Method and device for backtracking common scene conversation in multi-turn conversation |
| CN111625637A (en) * | 2020-05-27 | 2020-09-04 | 苏州思必驰信息科技有限公司 | Intention training and analyzing method and device |
| CN111625637B (en) * | 2020-05-27 | 2022-07-08 | 思必驰科技股份有限公司 | Intention training and analyzing method and device |
| CN111683182A (en) * | 2020-06-11 | 2020-09-18 | 中国银行股份有限公司 | Service node processing method and system |
| CN111767219A (en) * | 2020-06-24 | 2020-10-13 | 中国工商银行股份有限公司 | Method and device for checking intelligent customer service conversation process |
| CN111767219B (en) * | 2020-06-24 | 2023-08-15 | 中国工商银行股份有限公司 | Method and device for checking intelligent customer service dialogue flow |
| CN111985206A (en) * | 2020-07-17 | 2020-11-24 | 联想(北京)有限公司 | Corpus understanding method and equipment |
| CN111916111A (en) * | 2020-07-20 | 2020-11-10 | 中国建设银行股份有限公司 | Intelligent voice outbound method and device with emotion, server and storage medium |
| CN111858856A (en) * | 2020-07-23 | 2020-10-30 | 海信电子科技(武汉)有限公司 | Multi-round search type chatting method and display equipment |
| CN111883111A (en) * | 2020-07-30 | 2020-11-03 | 平安国际智慧城市科技股份有限公司 | Dialect training processing method and device, computer equipment and readable storage medium |
| CN113761183A (en) * | 2020-07-30 | 2021-12-07 | 北京汇钧科技有限公司 | Intention recognition method and intention recognition device |
| CN111883111B (en) * | 2020-07-30 | 2024-05-31 | 平安国际智慧城市科技股份有限公司 | Method, device, computer equipment and readable storage medium for processing speech training |
| CN112150694A (en) * | 2020-08-12 | 2020-12-29 | 国网浙江省电力有限公司营销服务中心 | Intelligent voice electric charge payment urging system and method |
| CN111933118A (en) * | 2020-08-17 | 2020-11-13 | 苏州思必驰信息科技有限公司 | Method and device for optimizing voice recognition and intelligent voice dialogue system applying same |
| CN112131357A (en) * | 2020-08-21 | 2020-12-25 | 国网浙江省电力有限公司杭州供电公司 | User intention identification method and device based on intelligent dialogue model |
| CN112035632A (en) * | 2020-08-21 | 2020-12-04 | 惠州市德赛西威汽车电子股份有限公司 | Preferred distribution method and system suitable for multi-conversation robot collaboration task |
| CN112199486A (en) * | 2020-10-21 | 2021-01-08 | 中国电子科技集团公司第十五研究所 | Task type multi-turn conversation method and system for office scene |
| CN112489639A (en) * | 2020-11-26 | 2021-03-12 | 北京百度网讯科技有限公司 | Audio signal processing method, device, system, electronic equipment and readable medium |
| CN112487144A (en) * | 2020-11-30 | 2021-03-12 | 北京天润融通科技股份有限公司 | Multi-round conversation robot intention hit optimization method and device |
| CN112487170B (en) * | 2020-12-14 | 2023-12-15 | 南京三眼精灵信息技术有限公司 | Man-machine interaction dialogue robot system facing scene configuration |
| CN112487170A (en) * | 2020-12-14 | 2021-03-12 | 南京三眼精灵信息技术有限公司 | Scene configuration-oriented human-computer interaction dialogue robot system |
| CN112686674A (en) * | 2020-12-25 | 2021-04-20 | 科讯嘉联信息技术有限公司 | Customer service conversation work order summarizing method |
| CN112612462A (en) * | 2020-12-29 | 2021-04-06 | 平安科技（深圳）有限公司 | Method and device for adjusting phone configuration, electronic equipment and storage medium |
| CN112765331A (en) * | 2020-12-31 | 2021-05-07 | 杭州摸象大数据科技有限公司 | Dialogue knowledge template construction method and device, computer equipment and storage medium |
| CN112765331B (en) * | 2020-12-31 | 2022-11-18 | 杭州摸象大数据科技有限公司 | Dialogue knowledge template construction method and device, computer equipment and storage medium |
| CN112784024B (en) * | 2021-01-11 | 2023-10-31 | 软通动力信息技术(集团)股份有限公司 | Man-machine conversation method, device, equipment and storage medium |
| CN112784024A (en) * | 2021-01-11 | 2021-05-11 | 软通动力信息技术(集团)股份有限公司 | Man-machine conversation method, device, equipment and storage medium |
| CN112988963A (en) * | 2021-02-19 | 2021-06-18 | 平安科技（深圳）有限公司 | User intention prediction method, device, equipment and medium based on multi-process node |
| CN112988963B (en) * | 2021-02-19 | 2024-05-10 | 平安科技（深圳）有限公司 | User intention prediction method, device, equipment and medium based on multi-flow nodes |
| CN112905749B (en) * | 2021-03-12 | 2022-07-29 | 电子科技大学 | Task-based multi-turn dialogue method based on intention-slot value rule tree |
| CN112905749A (en) * | 2021-03-12 | 2021-06-04 | 电子科技大学 | Task-type multi-turn dialogue method based on intention-slot value rule tree |
| CN112989003B (en) * | 2021-04-01 | 2023-04-18 | 网易（杭州）网络有限公司 | Intention recognition method, device, processing equipment and medium |
| CN112989003A (en) * | 2021-04-01 | 2021-06-18 | 网易（杭州）网络有限公司 | Intention recognition method, device, processing equipment and medium |
| CN113139059A (en) * | 2021-05-13 | 2021-07-20 | 八维(杭州)科技有限公司 | Intention grading method based on man-machine conversation |
| CN117389541A (en) * | 2023-12-13 | 2024-01-12 | 中国人民解放军国防科技大学 | Configuration system and device for generating template based on dialogue retrieval |
| CN117389541B (en) * | 2023-12-13 | 2024-02-23 | 中国人民解放军国防科技大学 | Configuration system and device for generating template based on dialogue retrieval |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111078846A (en) | Multi-turn dialog system construction method and system based on business scene | |
| CN111414479B (en) | Label extraction method based on short text clustering technology | |
| CN109460459B (en) | Log learning-based dialogue system automatic optimization method | |
| CN111191442A (en) | Similar problem generation method, device, equipment and medium | |
| CN112036168B (en) | Event main body recognition model optimization method, device, equipment and readable storage medium | |
| CN111639186B (en) | Multi-category multi-label text classification model and device with dynamic embedded projection gating | |
| CN112347760A (en) | Method and device for training intention recognition model and method and device for recognizing intention | |
| CN113505200A (en) | Sentence-level Chinese event detection method combining document key information | |
| WO2023065642A1 (en) | Corpus screening method, intention recognition model optimization method, device, and storage medium | |
| TWI734085B (en) | Dialogue system using intention detection ensemble learning and method thereof | |
| CN113449084A (en) | Relationship extraction method based on graph convolution | |
| CN110633475A (en) | Natural language understanding method, device and system based on computer scene and storage medium | |
| CN108681532B (en) | Sentiment analysis method for Chinese microblog | |
| CN112183106A (en) | Semantic understanding method and device based on phoneme association and deep learning | |
| CN116049387A (en) | Short text classification method, device and medium based on graph convolution | |
| CN111581346A (en) | Event extraction method and device | |
| CN112579730A (en) | High-expansibility multi-label text classification method and device | |
| CN112100368B (en) | Method and device for identifying dialogue interaction intention | |
| CN110633468B (en) | Information processing method and device for object feature extraction | |
| CN112528657A (en) | Text intention recognition method and device based on bidirectional LSTM, server and medium | |
| CN117056506A (en) | Public opinion emotion classification method based on long-sequence text data | |
| CN116384379A (en) | Chinese clinical term standardization method based on deep learning | |
| CN116070642A (en) | Text emotion analysis method and related device based on expression embedding | |
| CN115357718A (en) | Method, system, device and storage medium for discovering repeated materials of theme integration service | |
| CN114254622A (en) | Intention identification method and device |
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: 20200428 |
|
| RJ01 | Rejection of invention patent application after publication |