CN110866103B - Sentence diversity generation method and system in dialogue system - Google Patents
Sentence diversity generation method and system in dialogue system Download PDFInfo
- Publication number
- CN110866103B CN110866103B CN201911087246.6A CN201911087246A CN110866103B CN 110866103 B CN110866103 B CN 110866103B CN 201911087246 A CN201911087246 A CN 201911087246A CN 110866103 B CN110866103 B CN 110866103B
- Authority
- CN
- China
- Prior art keywords
- sentence
- answer
- answer sentence
- graph
- feature vector
- 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
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
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- 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)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a sentence diversity generation method and a sentence diversity generation system in a dialogue system, wherein the method comprises the following steps: s1, extracting a dependency tree of an answer sentence, and converting the dependency tree into an undirected graph; step S2, inputting the answer sentence and the undirected graph obtained in the step S1 into a graph structure converter to obtain a feature vector of the answer sentence; step S3, extracting feature vectors of dialogue histories of the answer sentences by using the sequence structure converter; and S4, inputting the feature vector of the answer sentence obtained in the step S2 and the feature vector of the dialogue history obtained in the step S3 into a conditional variation automatic encoder to obtain a new answer sentence of the dialogue history.
Description
Technical Field
The invention relates to the technical field of dialogue systems, in particular to a sentence diversity generation method and system for fusing sentence grammar structures in a dialogue system.
Background
The dialogue system is a research direction of natural language processing, and the research purpose of the dialogue system is to generate the next sentence of the dialogue history according to the dialogue history of a user and a dialogue robot. In the field of dialog systems, a number of related technologies have been developed, including mainly retrievable dialog systems, generative dialog systems, and dialog systems in which retrievable and generative are mixed.
In reality, for the same dialogue history, there are a plurality of different answers, which is a sentence diversity generation problem in the dialogue system. However, in the dialogue system of the prior art, the sentence generation does not use the grammar structure information of the answer sentence, so that the correlation of the generated sentences is not strong in some cases, and a good dialogue effect cannot be realized.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a sentence diversity generation method and system in a dialogue system so as to improve the sentence diversity in the dialogue system.
In order to achieve the above and other objects, the present invention provides a sentence diversity generation method in a dialogue system, comprising the steps of:
s1, extracting a dependency tree of an answer sentence, and converting the dependency tree into an undirected graph;
step S2, inputting the answer sentence and the undirected graph obtained in the step S1 into a graph structure converter to obtain a feature vector of the answer sentence;
step S3, extracting feature vectors of dialogue histories of the answer sentences by using the sequence structure converter;
and S4, inputting the feature vector of the answer sentence obtained in the step S2 and the feature vector of the dialogue history obtained in the step S3 into a conditional variation automatic encoder to obtain a new answer sentence of the dialogue history.
Preferably, step S1 further comprises:
step S100, extracting a dependency tree of the answer sentence by using an open source natural language processing tool;
step S101, representing the dependency tree by a directed graph, wherein nodes in the dependency tree are words of sentences, and directed edges in the dependency tree represent syntactic relations among the words;
and step S102, changing the directed edges in the directed graph into undirected edges to obtain an undirected graph of the answer sentence.
Preferably, in step S1, the undirected graph is represented by an adjacency matrix.
Preferably, if the answer sentence has n words, the adjacency matrix of the answer sentence is a matrix M with dimension n×n, and the value M of the ith row and jth column in the adjacency matrix M ij Is determined by the following conditions:
preferably, step S2 further comprises
Step S200, graph Attention operation is carried out on the feature V of the answer sentence and the adjacency matrix M of the undirected Graph;
step S201, adding the result of Graph attribute operation and the feature V, and performing layer normalization operation;
step S202, junction of step S201
Inputting a layer of feedforward neural network, and performing layer normalization operation to obtain the feature vector of the answer sentence.
Preferably, in step S3, m sentences of the dialogue history are acquired, the m sentences are arranged in sequence, the m sentences are spliced into a sentence C in sequence end to end, and the sentence C is input to the sequence structure converter to obtain the feature vector of the dialogue history.
Preferably, the automatic conditional variation encoder is composed of an encoder and a decoder, the feature vector E ' of the conversation history obtained in step S3 is input to the encoder of the automatic conditional variation encoder to obtain a normal distribution z ', and a plurality of samples are sampled from the normal distribution z ', and then are input to the decoder respectively to obtain a plurality of different answer sentences.
In order to achieve the above object, the present invention further provides a sentence diversity generation system in a dialogue system, including:
an answer sentence processing unit for extracting a dependency tree of an answer sentence and converting the dependency tree into an undirected graph;
an answer sentence feature vector extraction unit configured to input the answer sentence and the undirected graph of the answer sentence obtained by the answer sentence processing unit into a graph structure converter to obtain a feature vector of the answer sentence;
a dialogue history feature extraction unit for extracting feature vectors of dialogue histories of the answer sentences using a sequence structure converter;
and the diversity sentence generating unit is used for inputting the feature vectors of the answer sentences of the answer sentence feature vector extracting unit and the feature vectors of the dialogue history feature extracting unit into the conditional variation automatic encoder to obtain new answer sentences of the dialogue history.
Preferably, in the answer sentence processing unit, the conversion method of the dependency tree into an undirected graph is to change a directed edge into an undirected edge, and the undirected graph is represented by an adjacency matrix.
Preferably, in the dialogue history feature extraction unit, m sentences of the dialogue history are acquired, the m sentences are arranged in sequence, the m sentences are spliced into a sentence C in sequence end to end, and the sentence C is input to the sequence structure converter to obtain feature vectors of the dialogue history.
Compared with the prior art, the sentence diversity generation method and system in the dialogue system convert the dependency tree into the undirected graph by extracting the dependency tree of the answer sentence, then input the answer sentence and the undirected graph into the graph structure converter to obtain the feature vector of the answer sentence, extract the feature vector of the dialogue history of the answer sentence by using the sequence structure converter, and finally input the obtained feature vector of the answer sentence and the obtained feature vector of the dialogue history into the conditional variation automatic encoder to obtain the new answer sentence of the dialogue history, thereby realizing the purpose of improving the sentence generation diversity in the dialogue system.
Drawings
FIG. 1 is a flow chart showing the steps of a sentence diversity generation method in a dialogue system according to the present invention;
FIG. 2 is a schematic diagram of a dependency tree in an embodiment of the present invention;
FIG. 3 is a diagram of a directed graph representation of a dependency tree in an embodiment of the present invention;
FIG. 4 is a diagram illustrating the transformation of a dependency tree into undirected graph according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a adjacency matrix of an answer sentence in an embodiment of the invention;
FIG. 6 is a block diagram of the architecture converter (Graph Transformer) of an embodiment of the present invention;
FIG. 7 is a block diagram of a conditional variance automatic encoder in an embodiment of the invention;
fig. 8 is a system architecture diagram of a sentence diversity generation system in a dialog system according to the present invention.
Detailed Description
Other advantages and effects of the present invention will become readily apparent to those skilled in the art from the following disclosure, when considered in light of the accompanying drawings, by describing embodiments of the present invention with specific embodiments thereof. The invention may be practiced or carried out in other embodiments and details within the scope and range of equivalents of the various features and advantages of the invention.
Fig. 1 is a flowchart illustrating steps of a sentence diversity generation method in a dialogue system according to the present invention. As shown in fig. 1, the sentence diversity generating method in the dialogue system of the present invention includes the following steps:
step S1, extracting a dependency tree of the answer sentence, and converting the dependency tree into an undirected graph, wherein the undirected graph is represented by an adjacent matrix M.
Specifically, the answer sentence is an answer sentence to a question in the dialogue system, and the dependency tree of the answer sentence can be extracted by using an open-source natural language processing tool, such as Stanford CoreNLP, allenlp, and the like. The dependency tree is a directed graph, the nodes in the dependency tree are words of a sentence, and the directed edges in the dependency tree represent syntactic relationships between the words. If there is a certain syntactic relationship between words, there will be a directed edge between the nodes represented by the two words in the directed graph.
In the invention, the conversion method for converting the dependency tree into the undirected graph is to change the directed edge of the dependency tree into the undirected edge.
Specifically, assuming that the answer sentence has n words, the adjacency matrix of the answer sentence is a matrix M having a dimension n×n, and the value M of the ith row and jth column in the adjacency matrix M ij Is determined by the following conditions:
examples of extracting sentence dependency trees and computing adjacency matrices are as follows: for example, a sentence "a syntactic structure is fused in sentence feature extraction". "
First, extracting a dependency tree of the sentence by using an open-source natural language processing tool, as shown in fig. 2;
then, the above dependency tree is represented by a directed graph, the nodes in the dependency tree are words of the sentence, and the directed edges in the dependency tree represent the syntactic relationship between the words, as shown in fig. 3.
And changing the directed edge in the directed graph into an undirected edge to obtain an undirected graph of the sentence, as shown in fig. 4.
Finally, the example sentence is fused with a syntactic structure in sentence feature extraction. The undirected graph of "is converted to an adjacency matrix M as shown in fig. 5.
Step S2, inputting the answer sentence and the adjacency matrix M of the undirected graph of step S1 into a graph structure converter (Graph Transformer) to obtain the feature vector of the answer sentence;
fig. 6 is a block diagram of the graph-structure converter (Graph Transformer) according to an embodiment of the present invention, and the following description is made with reference to fig. 6 for describing a feature extraction process of the graph-structure converter (Graph Transformer) according to the present invention:
specifically, assuming that the answer sentence is composed of n words, the i-th word is composed of a k-dimensional feature vector V i Expressed, the feature of the answer sentence is expressed as v= (V 1 ，...，V n ). The feature V of the reply sentence and the adjacency matrix M of the undirected graph are input to the graph structure converter (Graph Transformer).
The characteristic extraction process of the graph structure converter is as follows:
1. and carrying out Graph Attention operation on the feature V of the answer sentence and the adjacency matrix M of the undirected Graph. Specifically, feature vector V for the ith word i The Graph Attention is calculated as follows:
wherein M is ij Is step S1The value of the ith row and jth column of the adjacency matrix M in (b).
2. Will be
And V is equal to i Adding and performing layer normalization operation, wherein the specific operation is as follows:
where LayerNorm is a layer normalization operation, the layer normalization operation is not described in detail herein since it is prior art.
3. Will be
Inputting a layer of feedforward neural network, and then performing layer normalization operation, wherein the specific operation is as follows:
wherein FFN is a layer of feedforward neural network.
So that for the i-th word the feature vector V i The transformed feature vector is obtained after Graph Transformer treatment
Thereby obtaining the characteristics of the answer sentence after Graph Transformer transformation
Finally, for the characteristics of the answer sentences
The following operations are performed to obtain the characteristics V' of the final answer sentence:
step S3, extracting the characteristics of the dialogue history of the answer sentence by using a sequence structure converter (transducer);
specifically, in a dialogue system, a dialogue sample is typically composed of dialogue history and answer sentences, an example of which is as follows:
conversation history (m sentences):
1. how does today weather?
2. Today, the weather is good and the sunshine is bright.
……
M, do you feel that the next week will not go down to storms?
The reply sentence is the next sentence of the dialogue history, for example:
i feel that the next week will experience heavy rain.
It is assumed that in the dialogue system, the dialogue history of the answer sentence is composed of m sentences, and the m sentences are arranged in order, the m sentences are sequentially spliced into one sentence C, and the sentence C is input to a sequence structure converter (converter) to obtain a feature vector of the dialogue history.
Specifically, suppose that sentence C is composed of r words, and that the i-th word in sentence C is composed of a k-dimensional feature vector E i Expressed, the feature of sentence C is expressed as e= (E 1 ，...，E r ) After E is input into the transducer, the characteristics of the converted sentence C can be obtained
Feature of sentence C->
The following operations are performed to obtain the final dialogue history feature E':
step S4, inputting the feature vector V 'of the answer sentence of step S2 and the feature vector E' of the dialogue history of step S3 into a conditional variation automatic encoder to generate the answer sentence.
The structure diagram of the condition-variable automatic encoder is shown in fig. 7, the condition-variable automatic encoder is composed of an encoder and a decoder, the characteristic E ' of the conversation history is input into the encoder of the condition-variable automatic encoder to obtain a normal distribution z ', and a plurality of different answer sentences can be obtained only by sampling a plurality of samples from the normal distribution z ' and then respectively inputting the samples into the decoder. Specifically, after the feature vector E ' of the conversation history is input to the encoder in the conditional variation automatic encoder, a normal distribution Z ' can be obtained, and then the normal distribution Z ' is sampled multiple times and input to the decoder respectively, and then the decoder generates different answer sentences respectively, thereby realizing diversity generation of the answer sentences.
Fig. 8 is a system architecture diagram of a sentence diversity generation system in a dialog system according to the present invention. As shown in fig. 8, the sentence diversity generating system in the dialogue system of the present invention includes:
an answer sentence processing unit 201 for extracting a dependency tree of an answer sentence and converting the dependency tree into an undirected graph represented by an adjacency matrix;
specifically, the answer sentence is an answer sentence to a question in the dialogue system, and the dependency tree of the answer sentence can be extracted by using an open-source natural language processing tool, such as Stanford CoreNLP, allenlp, and the like. The dependency tree is a directed graph, the nodes in the dependency tree are words of a sentence, and the directed edges in the dependency tree represent syntactic relationships between the words. If there is a certain syntactic relationship between words, there will be a directed edge between the nodes represented by the two words in the directed graph.
In the present invention, the conversion method of the reply sentence extraction unit 201 to convert the dependency tree into the undirected graph is to change the directed edge of the dependency tree into the undirected edge.
Specifically, assuming that the answer sentence has n words, the adjacency matrix of the answer sentence is a matrix M having a dimension n×n, and the value M of the ith row and jth column in the adjacency matrix M ij Is determined by the following conditionsAnd (3) determining:
an answer sentence feature vector extraction unit 202 for inputting the answer sentence and the undirected graph of the answer sentence obtained by the answer sentence processing unit 201 into a graph structure converter (Graph Transformer) to obtain a feature vector of the answer sentence.
Assuming that the answer sentence is composed of n words, the ith word is composed of a k-dimensional feature vector V i Expressed, the feature of the answer sentence is expressed as v= (V 1 ，...，V n ). The feature V of the reply sentence and the adjacency matrix M of the undirected graph are input to the graph structure converter (Graph Transformer).
The characteristic extraction process of the graph structure converter is as follows:
1. and carrying out Graph Attention operation on the feature V of the answer sentence and the adjacency matrix M of the undirected Graph. Specifically, feature vector V for the ith word i The Graph Attention is calculated as follows:
wherein M is ij Is the value of the ith row and jth column of the adjacency matrix M in step S1.
2. Will be
And V is equal to i Adding and performing layer normalization operation, wherein the specific operation is as follows:
where LayerNorm is a layer normalization operation, the layer normalization operation is not described in detail herein since it is prior art.
3. Will be
Inputting a layer of feedforward neural network, and then performing layer normalization operation, wherein the specific operation is as follows:
wherein FFN is a layer of feedforward neural network.
So that for the i-th word the feature vector V i The transformed feature vector is obtained after the processing of the graph structure converter (Graph Transformer)
Thereby obtaining the characteristic +.A reply sentence transformed by the diagram structure transformer (Graph Transformer)>
Finally, for the characteristics of the answer sentences
The following operations are performed to obtain the characteristics V' of the final answer sentence:
a dialogue history feature extraction unit 203 for acquiring an answer history of the answer sentence, extracting feature vectors of the dialogue history using a sequence structure converter (transducer);
specifically, it is assumed that in the dialogue system, the dialogue history of the answer sentence is composed of m sentences, and the m sentences are arranged in order, the m sentences are sequentially spliced into one sentence C, and the sentence C is input to a sequence structure converter (converter) to obtain a feature vector of the dialogue history.
Specifically, suppose that sentence C is composed of r words, and that the i-th word in sentence C is composed of a k-dimensional feature vector E i Expressed, the feature of sentence C is expressed as e= (E 1 ，...，E r ) After inputting E into a sequence structure converter (transducer), the characteristics of the converted sentence C can be obtained
Feature of sentence C->
The following operations are performed to obtain the final dialogue history feature E':
the diversity sentence generating unit 204 inputs the feature vector of the answer sentence feature vector extracting unit 202 and the feature vector of the dialogue history feature extracting unit 203 to the conditional variance automatic encoder, and obtains the answer sentence of the dialogue history.
The condition variation automatic encoder consists of an encoder and a decoder, the characteristic E ' of the dialogue history is input into the encoder of the condition variation automatic encoder to obtain a normal distribution z ', and a plurality of different answer sentences can be obtained only by sampling a plurality of samples from the normal distribution z ' and then respectively inputting the samples into the decoder
In summary, the sentence diversity generating method and system in the dialogue system of the present invention convert the dependency tree into the undirected graph by extracting the dependency tree of the answer sentence, then input the answer sentence and undirected graph into the graph-hook converter to obtain the feature vector of the answer sentence, extract the feature vector of the dialogue history of the answer sentence by using the sequence structure converter, and finally input the obtained feature vector of the answer sentence and the obtained feature vector of the dialogue history into the condition-variable automatic encoder to obtain the new answer sentence of the dialogue history, thereby achieving the purpose of improving the sentence diversity in the dialogue system.
The above embodiments are merely illustrative of the principles of the present invention and its effectiveness, and are not intended to limit the invention. Modifications and variations may be made to the above-described embodiments by those skilled in the art without departing from the spirit and scope of the invention. Accordingly, the scope of the invention is to be indicated by the appended claims.
Claims (8)
1. A sentence diversity generation method in a dialogue system includes the steps of:
s1, extracting a dependency tree of an answer sentence, and converting the dependency tree into an undirected graph;
step S2, inputting the answer sentence and the undirected graph obtained in the step S1 into a graph structure converter to obtain a feature vector of the answer sentence;
step S3, extracting the feature vector of the dialogue history of the answer sentence by using a sequence structure converter;
step S4, inputting the feature vector of the answer sentence obtained in the step S2 and the feature vector of the dialogue history obtained in the step S3 into a condition variation automatic encoder to obtain a new answer sentence of the dialogue history;
if the answer sentence has n words, the adjacency matrix of the answer sentence is a matrix M with dimension n x n, and the value M of the ith row and the jth column in the adjacency matrix M ij Is determined by the following conditions:
step S2 further comprises:
step S200, graph Attention operation is carried out on the feature V of the answer sentence and the adjacency matrix M of the undirected Graph;
feature vector V for the ith word i The Graph Attention is calculated as follows:
wherein M is ij Is the value of the ith row and jth column of the adjacency matrix M in step S1;
step S201, adding the result of Graph attribute operation and the feature V, and performing layer normalization operation;
step S202, junction of step S201
Inputting a layer of feedforward neural network, and performing layer normalization operation to obtain the feature vector of the answer sentence.
2. The sentence diversity generation method in a dialog system of claim 1, wherein step S1 further comprises:
step S100, extracting a dependency tree of the answer sentence by using an open source natural language processing tool;
step S101, representing the dependency tree by a directed graph, wherein nodes in the dependency tree are words of sentences, and directed edges in the dependency tree represent syntactic relations among the words;
and step S102, changing the directed edges in the directed graph into undirected edges to obtain an undirected graph of the answer sentence.
3. The sentence diversity generation method in a dialog system of claim 2, wherein: in step S1, the undirected graph is represented by an adjacency matrix.
4. The sentence diversity generation method in a dialog system of claim 1, wherein: in step S3, m sentences of the conversation history are obtained, the m sentences are arranged in sequence, the m sentences are spliced into a sentence C in sequence end to end, and the sentence C is input to the sequence structure converter to obtain the feature vector of the conversation history.
5. The sentence diversity generation method in a dialog system of claim 4, wherein: the condition variation automatic encoder is composed of an encoder and a decoder, the feature vector E ' of the conversation history obtained in the step S3 is input into the encoder of the condition variation automatic encoder to obtain a normal distribution z ', a plurality of samples are sampled from the normal distribution z ', and then the samples are respectively input into the decoder to obtain a plurality of different answer sentences.
6. A sentence diversity generation system in a dialog system for implementing the sentence diversity generation method of any of claims 1 to 5, comprising:
an answer sentence processing unit for extracting a dependency tree of an answer sentence and converting the dependency tree into an undirected graph;
an answer sentence feature vector extraction unit configured to input the answer sentence and the undirected graph of the answer sentence obtained by the answer sentence processing unit into a graph structure converter to obtain a feature vector of the answer sentence;
a dialogue history feature extraction unit for extracting feature vectors of dialogue histories of the answer sentences using a sequence structure converter;
and the diversity sentence generating unit is used for inputting the feature vectors of the answer sentences of the answer sentence feature vector extracting unit and the feature vectors of the dialogue history feature extracting unit into the conditional variation automatic encoder to obtain new answer sentences of the dialogue history.
7. The sentence diversity generation system in a dialog system of claim 6, wherein: in the answer sentence processing unit, the conversion method of the dependency tree into the undirected graph is to change the directed edge into the undirected edge, and the undirected graph is represented by an adjacency matrix.
8. The sentence diversity generation method in a dialog system of claim 6, wherein: in the dialogue history feature extraction unit, m sentences of dialogue history are obtained, the m sentences are arranged in sequence, the m sentences are spliced into a sentence C in sequence in an end-to-end mode, and the sentence C is input into the sequence structure converter to obtain feature vectors of the dialogue history.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911087246.6A CN110866103B (en) | 2019-11-08 | 2019-11-08 | Sentence diversity generation method and system in dialogue system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911087246.6A CN110866103B (en) | 2019-11-08 | 2019-11-08 | Sentence diversity generation method and system in dialogue system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110866103A CN110866103A (en) | 2020-03-06 |
| CN110866103B true CN110866103B (en) | 2023-07-07 |
Family
ID=69654516
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911087246.6A Active CN110866103B (en) | 2019-11-08 | 2019-11-08 | Sentence diversity generation method and system in dialogue system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110866103B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109543020A (en) * | 2018-11-27 | 2019-03-29 | 科大讯飞股份有限公司 | Inquiry handles method and system |
| CN109597876A (en) * | 2018-11-07 | 2019-04-09 | 中山大学 | A kind of more wheels dialogue answer preference pattern and its method based on intensified learning |
| CN109726276A (en) * | 2018-12-29 | 2019-05-07 | 中山大学 | A kind of Task conversational system based on depth e-learning |
| CN110309287A (en) * | 2019-07-08 | 2019-10-08 | 北京邮电大学 | The retrieval type of modeling dialog round information chats dialogue scoring method |
-
2019
- 2019-11-08 CN CN201911087246.6A patent/CN110866103B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109597876A (en) * | 2018-11-07 | 2019-04-09 | 中山大学 | A kind of more wheels dialogue answer preference pattern and its method based on intensified learning |
| CN109543020A (en) * | 2018-11-27 | 2019-03-29 | 科大讯飞股份有限公司 | Inquiry handles method and system |
| CN109726276A (en) * | 2018-12-29 | 2019-05-07 | 中山大学 | A kind of Task conversational system based on depth e-learning |
| CN110309287A (en) * | 2019-07-08 | 2019-10-08 | 北京邮电大学 | The retrieval type of modeling dialog round information chats dialogue scoring method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110866103A (en) | 2020-03-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112528672B (en) | Aspect-level emotion analysis method and device based on graph convolution neural network | |
| RU2722571C1 (en) | Method of recognizing named entities in network text based on elimination of probability ambiguity in neural network | |
| CN110418210B (en) | Video description generation method based on bidirectional cyclic neural network and depth output | |
| CN109964223B (en) | Session information processing method and device, storage medium | |
| CN109492113B (en) | Entity and relation combined extraction method for software defect knowledge | |
| CN109344242B (en) | Dialogue question-answering method, device, equipment and storage medium | |
| CN110704576A (en) | Text-based entity relationship extraction method and device | |
| CN111402861A (en) | Voice recognition method, device, equipment and storage medium | |
| JP6810580B2 (en) | Language model learning device and its program | |
| CN114118417A (en) | Multi-mode pre-training method, device, equipment and medium | |
| CN110795549A (en) | Short text conversation method, device, equipment and storage medium | |
| CN112528654A (en) | Natural language processing method and device and electronic equipment | |
| CN115221306B (en) | Automatic response evaluation method and device | |
| CN114925195A (en) | Standard content text abstract generation method integrating vocabulary coding and structure coding | |
| CN114254637A (en) | Summary generation method, device, equipment and storage medium | |
| CN114154518A (en) | Data enhancement model training method and device, electronic equipment and storage medium | |
| CN115994317A (en) | Incomplete multi-view multi-label classification method and system based on depth contrast learning | |
| CN112069781A (en) | Comment generation method and device, terminal device and storage medium | |
| CN113326367B (en) | Task type dialogue method and system based on end-to-end text generation | |
| CN114510576A (en) | Entity relationship extraction method based on BERT and BiGRU fusion attention mechanism | |
| WO2019171925A1 (en) | Device, method and program using language model | |
| CN114239607A (en) | Conversation reply method and device | |
| CN110866103B (en) | Sentence diversity generation method and system in dialogue system | |
| CN111813907A (en) | Question and sentence intention identification method in natural language question-answering technology | |
| CN116340507A (en) | Aspect-level emotion analysis method based on mixed weight and double-channel graph convolution |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Liang Xiaodan Inventor after: Chen Bingcheng Inventor after: Lin Jing Inventor before: Chen Bingcheng Inventor before: Liang Xiaodan Inventor before: Lin Jing |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |