CN114297450A - Deep learning-based dialogue system and dialogue method thereof - Google Patents
Deep learning-based dialogue system and dialogue method thereof Download PDFInfo
- Publication number
- CN114297450A CN114297450A CN202111665499.4A CN202111665499A CN114297450A CN 114297450 A CN114297450 A CN 114297450A CN 202111665499 A CN202111665499 A CN 202111665499A CN 114297450 A CN114297450 A CN 114297450A
- Authority
- CN
- China
- Prior art keywords
- module
- user
- dialogue
- question
- 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
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a deep learning-based dialogue system and a dialogue method thereof, wherein the system comprises a man-machine interaction module, a dialogue switching module, a dialogue environment analysis module, a data providing module, a dialogue storage module and a dialogue processing module, and the dialogue method comprises user questioning, scene switching, data searching and calculating and outputting responses; according to the invention, through the conversation environment switching of languages and moods, the conversation process is richer and more personalized, the intelligent conversation experience of the user is improved, the data support is provided through the data providing module, the user question is subjected to convolution calculation through the deep learning module, and the optimal response is output, so that the system can perform detailed response aiming at the questions in different knowledge fields of the user, the system can reply by integrating the whole conversation process during response through the analysis and extraction of the historical conversation record through the conversation storage module, so that the client does not need to perform question-asking and conversation operations for many times, and the intelligent degree is higher.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a deep learning-based dialogue system and a deep learning-based dialogue method.
Background
The establishment of a dialog system capable of naturally communicating with human beings is a long vision in the field of artificial intelligence, at present, an automatic human-computer dialog system has gained considerable attention in both academia and industry, the dialog system is a human-computer interaction system capable of selecting reply sentences returned to a user from candidate reply sentences according to question sentences input by the user, with the establishment of a large number of artificial intelligence companies, the technology behind the dialog system is continuously accumulated and decrypted, with the increasingly deep exploration of researchers, the dialog system is gradually going from science fiction movies to real life, in recent years, the research of the artificial intelligence field for the dialog system has rapidly developed, on one hand, the artificial intelligence field benefits from the progress of deep learning technology and the increase of internet dialog data volume, on the other hand, the application prospect and commercial value brought by the landing of products such as intelligent assistants and chatty robots and the like, therefore, the research of the dialogue system has important significance for the development of technology and industry;
for the current human-computer conversation field, the conversation system based on deep learning is the development direction of the current conversation system, the conversation system in the prior art is mainly divided into the types of online customer service, entertainment, education, personal assistant, intelligent question answering and the like, the current conversation system has single function, the conversation system can not answer questions in different knowledge fields in detail and can not reply by integrating the whole conversation process in the human-computer interaction conversation process, so that a client needs to ask questions and perform conversation operation for many times, the intelligent degree is low, the tone replied by the current conversation system in the human-computer interaction conversation process is always inconvenient to fix and can not switch corresponding conversation scenes according to the questions asked by the user in the conversation process, the whole conversation process is monotonous and not personalized, and the artificial intelligence experience brought to the user is poor, therefore, the invention provides a deep learning-based dialogue system and a dialogue method thereof to solve the problems in the prior art.
Disclosure of Invention
In view of the above problems, the present invention aims to provide a deep learning-based dialog system and a dialog method thereof, wherein the system provides data support through a data providing module, performs convolution calculation on user questions through a deep learning module, and outputs an optimal response, so that the system can perform detailed responses for the questions in different knowledge fields of the user, and can perform responses by integrating the whole dialog process when performing responses through analysis and extraction of historical dialog records by a dialog storage module, thereby enabling a client not to perform multiple question-asking and dialog operations.
In order to achieve the purpose of the invention, the invention is realized by the following technical scheme: a dialogue system based on deep learning comprises a man-machine interaction module, a dialogue switching module, a dialogue environment analysis module, a data providing module, a dialogue storage module and a dialogue processing module, and is characterized in that:
the man-machine interaction module comprises a question input module for inputting a user question and a reply output module for giving an answer to the user question;
the dialect switching module comprises an identity authentication unit for authenticating the identity information of the user and a dialect storage unit for storing different types of dialects;
the dialogue environment analysis module comprises a language analysis unit for analyzing the user question language and a language analysis unit for analyzing the user question language;
the data providing module provides big data support for system output reply;
the dialogue storage module is used for storing and analyzing the dialogue history records of the user and the system, extracting key information and providing assistance for the dialogue processing module;
the dialogue processing module comprises a question receiving unit for receiving user questions, a deep learning module for calculating the questions and obtaining the best answer, and an answer sending unit for transmitting the best answer to the answer output module.
The further improvement lies in that: the identity authentication module comprises a face identification unit, a fingerprint identification unit and a voice identification unit, wherein the face identification unit, the fingerprint identification unit and the voice identification unit are used for identifying the identity of the user, and the dialogues switching module authenticates the identity information of the user through the identity authentication module and enables the dialogues storage unit to provide corresponding dialogues according to the analysis result.
The further improvement lies in that: the question input module comprises a question input unit and an expression mode modification unit, the question input unit inputs a question provided by a user into the expression mode modification unit, and the expression mode modification unit modifies a semantic format of the question provided by the user into a general format convenient for system understanding.
The further improvement lies in that: the language analysis unit analyzes the user question language and switches the corresponding reply language according to the analysis result, and the tone analysis unit analyzes the user question tone and switches the corresponding reply language according to the analysis result.
The further improvement lies in that: the data providing module comprises a knowledge domain judging unit used for judging the domain in which the user raises a problem, a knowledge domain searching unit used for searching the corresponding knowledge domain data according to the judging result and a knowledge data reading unit used for reading the corresponding knowledge domain data.
The further improvement lies in that: the dialogue storage module comprises a history storage unit used for storing the dialogue history of the user and the system, a history analysis unit used for analyzing the dialogue history and a key information extraction unit used for extracting key information in the dialogue history.
The further improvement lies in that: the deep learning module comprises a data sorting module and a deep learning model, the data sorting module comprises a data set collecting unit for collecting knowledge data of corresponding fields, a data set labeling unit for labeling the collected data and a data set classifying unit for classifying the labeled data into a training set and a testing set, and the deep learning model is trained and learned by sorting the classified data training set and data testing set through the data sorting module and obtains the best answer through convolution calculation.
The further improvement lies in that: the positioning system is selected from one of a Beidou positioning system and a GPS positioning system, and preferably is the Beidou positioning system.
A dialogue method of a deep learning based dialogue system, comprising the steps of:
the method comprises the following steps: user questions
The user firstly verifies the identity information through the identity verification module, then the speech switching module switches the speech type provided by the corresponding speech storage unit according to the verification result, then the user asks questions, and the question input module inputs the user questions and modifies the semantic format of the questions into a general format convenient for system understanding;
step two: scenario switching
According to the first step, after the question input module inputs a question, the system analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result, and simultaneously, the system also analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result to realize the adaptive switching of the conversation scene;
step three: data search
According to the first step, the question input module is received by the question receiving unit after entering the question, the data providing module judges and searches according to the user question at the moment, reads the knowledge field data corresponding to the question and provides the data to the deep learning module, then the deep learning module collects, marks and classifies the read knowledge data by using the data sorting module, and then the deep learning module sorts the classified data training set and data testing set by the data sorting module to train and learn;
step four: computing and outputting replies
According to the third step, the deep learning model calculates the user questions and outputs the best answers of the user questions after the training and learning are finished, then the best answers are sent to the answer output module by the answer sending unit, and finally the questions are output to the user through the answer output module.
The further improvement lies in that: in the fourth step, the system stores, analyzes and extracts the conversation history record through the conversation storage module, the positioning system positions the user in real time, and the deep learning model performs optimal answer output in combination with the key information of the history conversation record extracted by the conversation storage module and the geographical position of the user in the process of settling the optimal answer.
The invention has the beneficial effects that: the invention modifies the semantic format of the question proposed by the user into a general format which is convenient for system understanding through the expression mode modification unit, so that the system can adaptively modify and understand and analyze the question which is not expressed by the user, the problem that the traditional dialogue system can not understand the user and ask the question in an abnormal way is avoided, the system can provide different dialogue techniques according to the identities of different users through the dialogue switching, the dialogue comfortable feeling of the users is improved, the dialogue process is richer and more individual through the dialogue environment switching of languages and languages, the intelligent dialogue experience of the users is improved, the data support is provided through the data providing module, the convolution calculation is carried out on the question of the user through the deep learning module, and the best answer is output, so that the system can carry out detailed answer aiming at the question in different knowledge fields of the users, the analysis and extraction of the history dialogue records are carried out through the dialogue storage module, the system can synthesize the whole dialogue process to reply when replying, so that the client does not need to ask questions and carry out dialogue operation for many times, the dialogue process is more simplified, and the intelligent degree is higher.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic diagram of the system architecture of the present invention;
FIG. 2 is a flow chart of a dialog method of the present invention;
FIG. 3 is a schematic diagram of the deep learning model application of the present invention;
fig. 4 is a schematic structural diagram of a system according to a third embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention. Furthermore, the terms "first," "second," "third," "fourth," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Example one
Referring to fig. 1, the embodiment provides a deep learning-based dialog system, which includes a human-computer interaction module, a dialog switching module, a dialog environment analysis module, a data providing module, a dialog storage module, and a dialog processing module, the human-computer interaction module comprises a question input module for inputting a user question and a reply output module for giving a reply to the user question, the question input module comprises a question input unit and an expression mode modification unit, the question input unit inputs the question provided by the user into the expression mode modification unit, the expression mode modification unit modifies the semantic format of the question presented by the user into a general format convenient for system understanding, so that the system can adaptively modify and understand and analyze the question which is not expressed by the user, and the problem that the traditional dialogue system cannot understand the user's abnormal question is avoided;
the dialect switching module comprises an identity authentication unit for authenticating user identity information and a dialect storage unit for storing different types of dialects, the identity authentication module comprises a face identification unit for identifying the user identity, a fingerprint identification unit and a voice identification unit, the dialect switching module authenticates the user identity information through the identity authentication module and enables the dialect storage unit to provide corresponding dialects according to an analysis result, so that the system can provide different dialects according to the identity of different users (for example, the dialect style of natural rot is switched when the user is authenticated as child identity), and the dialogue comfort of the user is improved;
the dialogue environment analysis module comprises a language analysis unit for analyzing the user question language and a language analysis unit for analyzing the user question language, the language analysis unit analyzes the user question language, and switches the corresponding reply language according to the analysis result, the language atmosphere analyzing unit analyzes the user query language atmosphere and switches the corresponding reply language atmosphere according to the analysis result, the system can serve users with different languages through the language switching function, can provide different dialogue languages for users with different emotions through the language switching function (for example, when the user inquires a question with low emotion, the dialogue is switched to the language with positive comfort to output a response to the user), therefore, the system is convenient for switching conversation environments, the conversation process is richer and more personalized, and the intelligent conversation experience of the user is improved;
the data providing module provides big data support for system output response, and the data providing module comprises a knowledge field judging unit for judging the field of question making by a user, a knowledge field searching unit for searching corresponding knowledge field data according to the judging result and a knowledge data reading unit for reading the corresponding knowledge field data, so that the system can perform detailed response aiming at questions in different knowledge fields of the user;
the dialogue storage module is used for storing, analyzing and extracting key information of dialogue history records of a user and a system and providing assistance for the dialogue processing module, the dialogue storage module comprises a history record storage unit used for storing the dialogue history of the user and the system, a history record analysis unit used for analyzing the dialogue history and a key information extraction unit used for extracting the key information in the dialogue history, and the system can reply by integrating the whole dialogue process when answering through analyzing and extracting the history dialogue records, so that a client does not need to ask questions and perform dialogue operations for many times, the dialogue process is simplified, and the intelligent degree is high;
the dialogue processing module comprises a question receiving unit for receiving user questions, a deep learning module for calculating the questions and obtaining the best answer, and an answer sending unit for transmitting the best answer to the answer output module, the deep learning module comprises a data sorting module and a deep learning model, the data sorting module comprises a data set collecting unit for collecting knowledge data of corresponding fields, a data set labeling unit for labeling the collected data, and a data set classifying unit for classifying the labeled data into a training set and a test set, the deep learning model is trained and learned by sorting the classified data training set and data test set through the data sorting module and obtaining the best answer through convolution calculation, as shown in fig. 3, the deep learning model comprises an input layer, a calculation layer and an output layer, the input layer is used for inputting word vectors, the calculation layer is used for performing convolution calculation on the word vectors, and the output layer is used for outputting optimal answers in a pooling mode;
the Beidou positioning system is used for positioning the position of the user in real time, so that the system can conveniently perform optimal reply output in combination with the position of the user in the calculation process.
Referring to fig. 2, the present embodiment further provides a dialog method of a dialog system based on deep learning, including the following steps:
the method comprises the following steps: user questions
The user firstly verifies the identity information through the identity verification module, then the speech switching module switches the speech type provided by the corresponding speech storage unit according to the verification result, then the user asks questions, and the question input module inputs the user questions and modifies the semantic format of the questions into a general format convenient for system understanding;
step two: scenario switching
According to the first step, after the question input module inputs a question, the system analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result, and simultaneously, the system also analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result to realize the adaptive switching of the conversation scene;
step three: data search
According to the first step, the question input module is received by the question receiving unit after entering the question, the data providing module judges and searches according to the user question at the moment, reads the knowledge field data corresponding to the question and provides the data to the deep learning module, then the deep learning module collects, marks and classifies the read knowledge data by using the data sorting module, and then the deep learning module sorts the classified data training set and data testing set by the data sorting module to train and learn;
step four: computing and outputting replies
According to the third step, the deep learning model calculates the user questions and outputs the best answers of the user questions after training and learning are completed, then the best answers are sent to the answer output module by the answer sending unit, finally the questions are output to the user through the answer output module, the conversation history record is stored, analyzed and extracted through the conversation storage module in the calculation process, the user is positioned in real time through the positioning system, and the deep learning model performs the best answer output in combination with the key information of the history conversation record extracted by the conversation storage module and the geographical position of the user in the process of clearing the best answers.
Example two
Referring to fig. 1, the embodiment provides a deep learning-based dialog system, which includes a human-computer interaction module, a dialog switching module, a dialog environment analysis module, a data providing module, a dialog storage module, and a dialog processing module, the human-computer interaction module comprises a question input module for inputting a user question and a reply output module for giving a reply to the user question, the question input module comprises a question input unit and an expression mode modification unit, the question input unit inputs the question provided by the user into the expression mode modification unit, the expression mode modification unit modifies the semantic format of the question presented by the user into a general format convenient for system understanding, so that the system can adaptively modify and understand and analyze the question which is not expressed by the user, and the problem that the traditional dialogue system cannot understand the user's abnormal question is avoided;
the dialect switching module comprises an identity authentication unit for authenticating user identity information and a dialect storage unit for storing different types of dialects, the identity authentication module comprises a face identification unit for identifying the user identity, a fingerprint identification unit and a voice identification unit, the dialect switching module authenticates the user identity information through the identity authentication module and enables the dialect storage unit to provide corresponding dialects according to an analysis result, so that the system can provide different dialects according to the identity of different users (for example, the dialect style of natural rot is switched when the user is authenticated as child identity), and the dialogue comfort of the user is improved;
the dialogue environment analysis module comprises a language analysis unit for analyzing the user question language and a language analysis unit for analyzing the user question language, the language analysis unit analyzes the user question language, and switches the corresponding reply language according to the analysis result, the language atmosphere analyzing unit analyzes the user query language atmosphere and switches the corresponding reply language atmosphere according to the analysis result, the system can serve users with different languages through the language switching function, can provide different dialogue languages for users with different emotions through the language switching function (for example, when the user inquires a question with low emotion, the dialogue is switched to the language with positive comfort to output a response to the user), therefore, the system is convenient for switching conversation environments, the conversation process is richer and more personalized, and the intelligent conversation experience of the user is improved;
the data providing module provides big data support for system output response, and the data providing module comprises a knowledge field judging unit for judging the field of question making by a user, a knowledge field searching unit for searching corresponding knowledge field data according to the judging result and a knowledge data reading unit for reading the corresponding knowledge field data, so that the system can perform detailed response aiming at questions in different knowledge fields of the user;
the dialogue storage module is used for storing, analyzing and extracting key information of dialogue history records of a user and a system and providing assistance for the dialogue processing module, the dialogue storage module comprises a history record storage unit used for storing the dialogue history of the user and the system, a history record analysis unit used for analyzing the dialogue history and a key information extraction unit used for extracting the key information in the dialogue history, and the system can reply by integrating the whole dialogue process when answering through analyzing and extracting the history dialogue records, so that a client does not need to ask questions and perform dialogue operations for many times, the dialogue process is simplified, and the intelligent degree is high;
the dialogue processing module comprises a question receiving unit for receiving user questions, a deep learning module for calculating the questions and obtaining the best answer, and an answer sending unit for transmitting the best answer to the answer output module, the deep learning module comprises a data sorting module and a deep learning model, the data sorting module comprises a data set collecting unit for collecting knowledge data of corresponding fields, a data set labeling unit for labeling the collected data, and a data set classifying unit for classifying the labeled data into a training set and a test set, the deep learning model is trained and learned by sorting the classified data training set and data test set through the data sorting module and obtaining the best answer through convolution calculation, as shown in fig. 3, the deep learning model comprises an input layer, a calculation layer and an output layer, the input layer is used for inputting word vectors, the calculation layer is used for performing convolution calculation on the word vectors, and the output layer is used for outputting optimal answers in a pooling mode;
the system also comprises a positioning module, wherein the positioning module is used for positioning the position of the user in real time through a GPS positioning system, so that the system can conveniently perform optimal reply output by combining the position of the user in the calculation process.
Referring to fig. 2, the present embodiment further provides a dialog method of a dialog system based on deep learning, including the following steps:
the method comprises the following steps: user questions
The user firstly verifies the identity information through the identity verification module, then the speech switching module switches the speech type provided by the corresponding speech storage unit according to the verification result, then the user asks questions, and the question input module inputs the user questions and modifies the semantic format of the questions into a general format convenient for system understanding;
step two: scenario switching
According to the first step, after the question input module inputs a question, the system analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result, and simultaneously, the system also analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result to realize the adaptive switching of the conversation scene;
step three: data search
According to the first step, the question input module is received by the question receiving unit after entering the question, the data providing module judges and searches according to the user question at the moment, reads the knowledge field data corresponding to the question and provides the data to the deep learning module, then the deep learning module collects, marks and classifies the read knowledge data by using the data sorting module, and then the deep learning module sorts the classified data training set and data testing set by the data sorting module to train and learn;
step four: computing and outputting replies
According to the third step, the deep learning model calculates the user questions and outputs the best answers of the user questions after training and learning are completed, then the best answers are sent to the answer output module by the answer sending unit, finally the questions are output to the user through the answer output module, the conversation history record is stored, analyzed and extracted through the conversation storage module in the calculation process, the user is positioned in real time through the positioning system, and the deep learning model performs the best answer output in combination with the key information of the history conversation record extracted by the conversation storage module and the geographical position of the user in the process of clearing the best answers.
EXAMPLE III
The deep learning-based dialogue system further comprises a voice adjusting module, a volume adapting module and a personalized dialogue module, wherein the voice adjusting module comprises a voice packet database and a voice packet switching unit, the voice packet database is connected with the Internet network and updates the types of voice packets in real time, and the voice packet switching unit switches different voices for the dialogue system according to the selection of a user;
the volume adaptation module comprises a volume analysis unit and a volume adjustment unit, the volume analysis unit senses the volume of a question asked by a user through a volume sensor and analyzes the environment where the user is located and the hearing ability of the user in the environment, and the volume adjustment unit adjusts the response volume of the response output module according to the analysis result to bring the best hearing experience to the user;
the personalized dialogue module searches the Internet question-answer community aiming at the personalized question of the user, screens out the best answer and outputs the best answer to the user, and the screening condition is the answer praise number of the Internet question-answer community, so that the personalized question-answer requirement of the user can be met, and the user experience is improved.
The dialogue system in the example was compared with the dialogue system of the prior art to test question and answer dialogues, as shown in table 1 below:
TABLE 1
As can be seen from the above table, the present invention can not only reply to questions in different knowledge fields, but also comprehensively analyze the subject and meaning of the current question of the user according to the historical dialogue records, and the traditional dialogue system can only reply according to the current question of the user, which results in that the system cannot output a reply to the question asked by the user under the condition of lack of partial subject and ambiguous meaning, and thus the user needs to ask questions many times, and the last question in the above table is visible.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A dialogue system based on deep learning comprises a man-machine interaction module, a dialogue switching module, a dialogue environment analysis module, a data providing module, a dialogue storage module and a dialogue processing module, and is characterized in that:
the man-machine interaction module comprises a question input module for inputting a user question and a reply output module for giving an answer to the user question;
the dialect switching module comprises an identity authentication unit for authenticating the identity information of the user and a dialect storage unit for storing different types of dialects;
the dialogue environment analysis module comprises a language analysis unit for analyzing the user question language and a language analysis unit for analyzing the user question language;
the data providing module provides big data support for system output reply;
the dialogue storage module is used for storing and analyzing the dialogue history records of the user and the system, extracting key information and providing assistance for the dialogue processing module;
the dialogue processing module comprises a question receiving unit for receiving user questions, a deep learning module for calculating the questions and obtaining the best answer, and an answer sending unit for transmitting the best answer to the answer output module.
2. The deep learning based dialog system of claim 1, wherein: the identity authentication module comprises a face identification unit, a fingerprint identification unit and a voice identification unit, wherein the face identification unit, the fingerprint identification unit and the voice identification unit are used for identifying the identity of the user, and the dialogues switching module authenticates the identity information of the user through the identity authentication module and enables the dialogues storage unit to provide corresponding dialogues according to the analysis result.
3. The deep learning based dialog system of claim 1, wherein: the question input module comprises a question input unit and an expression mode modification unit, the question input unit inputs a question provided by a user into the expression mode modification unit, and the expression mode modification unit modifies a semantic format of the question provided by the user into a general format convenient for system understanding.
4. The deep learning based dialog system of claim 1, wherein: the language analysis unit analyzes the user question language and switches the corresponding reply language according to the analysis result, and the tone analysis unit analyzes the user question tone and switches the corresponding reply language according to the analysis result.
5. The deep learning based dialog system of claim 1, wherein: the data providing module comprises a knowledge domain judging unit used for judging the domain in which the user raises a problem, a knowledge domain searching unit used for searching the corresponding knowledge domain data according to the judging result and a knowledge data reading unit used for reading the corresponding knowledge domain data.
6. The deep learning based dialog system of claim 1, wherein: the dialogue storage module comprises a history storage unit used for storing the dialogue history of the user and the system, a history analysis unit used for analyzing the dialogue history and a key information extraction unit used for extracting key information in the dialogue history.
7. The deep learning based dialog system of claim 1, wherein: the deep learning module comprises a data sorting module and a deep learning model, the data sorting module comprises a data set collecting unit for collecting knowledge data of corresponding fields, a data set labeling unit for labeling the collected data and a data set classifying unit for classifying the labeled data into a training set and a testing set, and the deep learning model is trained and learned by sorting the classified data training set and data testing set through the data sorting module and obtains the best answer through convolution calculation.
8. The deep learning based dialog system of claim 1, wherein: the positioning system is selected from one of a Beidou positioning system and a GPS positioning system, and preferably is the Beidou positioning system.
9. A dialogue method of a dialogue system based on deep learning is characterized by comprising the following steps:
the method comprises the following steps: user questions
The user firstly verifies the identity information through the identity verification module, then the speech switching module switches the speech type provided by the corresponding speech storage unit according to the verification result, then the user asks questions, and the question input module inputs the user questions and modifies the semantic format of the questions into a general format convenient for system understanding;
step two: scenario switching
According to the first step, after the question input module inputs a question, the system analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result, and simultaneously, the system also analyzes the question language of the user through the language analysis unit and switches the corresponding reply language according to the analysis result to realize the adaptive switching of the conversation scene;
step three: data search
According to the first step, the question input module is received by the question receiving unit after entering the question, the data providing module judges and searches according to the user question at the moment, reads the knowledge field data corresponding to the question and provides the data to the deep learning module, then the deep learning module collects, marks and classifies the read knowledge data by using the data sorting module, and then the deep learning module sorts the classified data training set and data testing set by the data sorting module to train and learn;
step four: computing and outputting replies
According to the third step, the deep learning model calculates the user questions and outputs the best answers of the user questions after the training and learning are finished, then the best answers are sent to the answer output module by the answer sending unit, and finally the questions are output to the user through the answer output module.
10. The dialogue method of the deep learning based dialogue system according to claim 9, wherein: in the fourth step, the system stores, analyzes and extracts the conversation history record through the conversation storage module, the positioning system positions the user in real time, and the deep learning model performs optimal answer output in combination with the key information of the history conversation record extracted by the conversation storage module and the geographical position of the user in the process of settling the optimal answer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111665499.4A CN114297450A (en) | 2021-12-31 | 2021-12-31 | Deep learning-based dialogue system and dialogue method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111665499.4A CN114297450A (en) | 2021-12-31 | 2021-12-31 | Deep learning-based dialogue system and dialogue method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114297450A true CN114297450A (en) | 2022-04-08 |
Family
ID=80973137
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111665499.4A Pending CN114297450A (en) | 2021-12-31 | 2021-12-31 | Deep learning-based dialogue system and dialogue method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114297450A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116303962A (en) * | 2023-03-21 | 2023-06-23 | 北京百度网讯科技有限公司 | Dialogue generation method, training method, device and equipment for deep learning model |
-
2021
- 2021-12-31 CN CN202111665499.4A patent/CN114297450A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116303962A (en) * | 2023-03-21 | 2023-06-23 | 北京百度网讯科技有限公司 | Dialogue generation method, training method, device and equipment for deep learning model |
| CN116303962B (en) * | 2023-03-21 | 2024-05-28 | 北京百度网讯科技有限公司 | Dialogue generation method, training method, device and equipment for deep learning model |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108959627B (en) | Question-answer interaction method and system based on intelligent robot | |
| CN103458056B (en) | Speech intention judging system based on automatic classification technology for automatic outbound system | |
| CN111026842A (en) | Natural language processing method, natural language processing device and intelligent question-answering system | |
| CN108806667A (en) | The method for synchronously recognizing of voice and mood based on neural network | |
| CN107464568A (en) | Based on the unrelated method for distinguishing speek person of Three dimensional convolution neutral net text and system | |
| CN106682090A (en) | Active interaction implementing device, active interaction implementing method and intelligent voice interaction equipment | |
| CN111177310B (en) | Intelligent scene conversation method and device for power service robot | |
| US20110231190A1 (en) | Method of and system for providing adaptive respondent training in a speech recognition application | |
| Prasad et al. | Intelligent chatbot for lab security and automation | |
| CN110610705A (en) | Voice interaction prompter based on artificial intelligence | |
| CN114841841A (en) | Intelligent education platform interaction system and interaction method for teaching interaction | |
| CN115640530A (en) | Combined analysis method for dialogue sarcasm and emotion based on multi-task learning | |
| CN114297450A (en) | Deep learning-based dialogue system and dialogue method thereof | |
| Shan et al. | [Retracted] Research on Classroom Online Teaching Model of “Learning” Wisdom Music on Wireless Network under the Background of Artificial Intelligence | |
| CN114048335A (en) | Knowledge base-based user interaction method and device | |
| CN109256128A (en) | A kind of method and system determining user role automatically according to user's corpus | |
| CN116340476A (en) | Virtual man dialogue method, system and storage medium | |
| CN115527404A (en) | Artificial intelligence self-adaptation interactive teaching system | |
| CN111382230B (en) | Fuzzy recognition method for legal consultation options | |
| JP4451037B2 (en) | Information search system and information search method | |
| CN114333770A (en) | Automatic pronunciation assessment method, device, equipment and storage medium | |
| CN113763934A (en) | Training method and device of audio recognition model, storage medium and electronic equipment | |
| CN112307186A (en) | Question-answering service method, system, terminal device and medium based on emotion recognition | |
| CN113688685B (en) | Sign language identification method based on interaction scene | |
| CN111475634B (en) | Representative speaking segment extraction device and method based on seat voice segmentation |
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 |