Intelligent question-answering system and method based on deep learning

Abstract

The invention discloses an intelligent question answering system and method based on deep learning, and belongs to the field of intelligent questions.The intelligent question answering system analyzes the questioning habit of a questioner, and obtains real-time question keywords according to the analyzed questioning habit of the questioner and real-time question data of the questioner; searching question answers according to the obtained real-time question keywords, answering the corresponding question answers, transmitting the corresponding question answers to the questioners, analyzing the questioning habits of the questioners, replacing the habitual spoken words of the users with recognizable language vocabularies, and obtaining the real-time questions according to the recognized language vocabularies. Meanwhile, the keyword of the question is obtained according to the questioning habit of the questioner, then content analysis is conducted on the question according to the importance degree of the keyword, the accuracy of question content recognition is further improved, and then the accuracy of question answering is improved.

Description

Technical Field

[0001] This application belongs to the field of intelligent Q&A, and specifically relates to an intelligent Q&A system and method based on deep learning. Background Art

[0002] The machine intelligent Q&A system is a question answering system based on artificial intelligence technology, which can understand the questions raised by users and retrieve, generate or infer accurate answers from a knowledge base or data source. The specific steps are as follows: parsing the user's question, extracting key information, retrieving relevant information from the knowledge base or data source, generating a natural language answer based on the retrieval result, and optimizing the system performance according to the user's feedback. The machine intelligent Q&A systems in the prior art can answer the questions raised by customers with a relatively high accuracy rate (such as GPT);

[0003] However, the machine intelligent Q&A systems in the prior art all have a drawback that they cannot analyze the questioning habits of the questioners, replace the habitual spoken words of the users with recognizable language words, and at the same time cannot obtain the keywords of the questions according to the questioning habits of the questioners, and thus cannot perform content analysis on the questions according to the importance degree of the keywords, which affects the accuracy of content recognition of the questions. Most of the prior art has the above problems;

[0004] In order to improve the accuracy of question answering, this application designs an intelligent Q&A system and method based on deep learning. Summary of the Invention

[0005] To solve the deficiencies in the prior art mentioned in the background art, this application proposes an intelligent Q&A system and method based on deep learning. This application analyzes the questioning habits of the questioners, obtains real-time question keywords according to the analyzed questioning habits of the questioners and the real-time question data of the questioners, searches for question answers according to the obtained real-time question keywords, and answers the corresponding question answers, transmits the corresponding question answering answers to the questioners, analyzes the questioning habits of the questioners, replaces the habitual spoken words of the users with recognizable language words, and at the same time obtains the keywords of the questions according to the questioning habits of the questioners, and thus performs content analysis on the questions according to the importance degree of the keywords, further improving the accuracy of content recognition of the questions, and thus improving the accuracy of question answering.

[0006] To achieve the above object, this application provides the following technical solutions: In the first aspect, this application provides an intelligent Q&A method based on deep learning, which includes the following specific steps:

[0007] Step 1, obtain the question data of the questioner, and at the same time obtain the historical question data and answer acquisition data of the questioner;

[0008] Step 2: Based on the obtained historical question data of the questioner and the answer acquisition data, construct a deep learning neural network model to analyze the questioner's question habits;

[0009] Step 3: Obtain real-time question keywords according to the analyzed questioner's question habits and the real-time question data of the questioner;

[0010] Step 4: Search for question answers based on the obtained real-time question keywords and answer the corresponding question answers;

[0011] Step 5: Transmit the corresponding question answer to the questioner.

[0012] Preferably, the question data of the questioner in Step 1 includes the constituent character data of the real-time question, the historical question data of the questioner in Step 1 includes the constituent character data of the historical question, and at the same time, the answer acquisition data in Step 1 is the question data corresponding to the final answer data obtained after the questioner's question. The various obtained data are stored in the storage component correspondingly for easy retrieval.

[0013] Preferably, the analysis of the questioner's question habits includes the following specific steps:

[0014] Step 201: Obtain the constituent character data of the historical question and the question data corresponding to the final answer data obtained after the questioner's question, and the constituent character data of the corresponding historical question and the constituent character data of the question corresponding to the corresponding final answer data;

[0015] Step 202: Obtain the constituent character data of the corresponding historical question and the constituent character data of the question corresponding to the corresponding final answer data, and construct a deep learning neural network model with the input being the constituent character data of the corresponding historical question and the output being the constituent character data of the question corresponding to the corresponding final answer data, that is, the questioner's question model;

[0016] The specific construction method is: divide the obtained constituent character data of the corresponding historical question and the constituent character data of the question corresponding to the corresponding final answer data into an 85% weight and bias training set and a 15% weight and bias test set; input the 85% weight and bias training set into the deep learning neural network model for training to obtain an initial deep learning neural network model; use the 15% weight and bias test set to test the initial deep learning neural network model, and output the optimal initial deep learning neural network model output that meets the accuracy of the constituent character data of the question corresponding to the preset corresponding final answer data as the neural network model. Among them, the formula in the neural network model is: , where is the output k of the s-th neuron in the m+1 layer, is the connection weight G between the j-th neuron in the m-th layer and the s-th neuron in the (m + 1)-th layer, represents the input R of the j-th neuron in the m-th layer, is the bias z representing the linear relationship between the j-th neuron in the m-th layer and the s-th neuron in the (m + 1)-th layer, represents the Sigmoid activation function, and w is the number of input neurons in the neural network model of the m-th layer.

[0017] Step 203: Furthermore, obtain the question habit model corresponding to the questioner. In this way, replace the habitual spoken words of the user with recognizable language words, for example, replace habitual spoken words such as dialects with recognizable language words.

[0018] Preferably, the acquisition of the real-time question keywords includes the following specific steps:

[0019] Import the obtained question habit model corresponding to the questioner and the real-time question data of the corresponding questioner into the constructed question habit model. Through the continuous operation of the question habit model, obtain the character data composition of the corresponding real-time question data, and obtain the character data composition of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions;

[0020] Based on the occurrence times and occurrence times of the character data composition of the real-time question data corresponding to the character data composition of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, conduct a character habit importance evaluation. Among them, the calculation formula for the importance of the i-th character habit is: , where mi is the number of corresponding questions in which the i-th character appears in the character data composition of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, sj is the number of times the i-th character appears in the j-th question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, tm is the set time standard value used to standardize the time interval, and tj is the time interval from the j-th question data corresponding to the final answer data obtained after the corresponding questioner's historical questions to the current time;

[0021] Arrange the character data composition of the corresponding real-time question data in descending order of the importance of the character habit, and select a relatively large number of constituent characters as the keywords of the corresponding real-time question data.

[0022] Preferably, the searching for question answers based on the obtained real-time question keywords includes the following specific steps: obtaining the keywords corresponding to the real-time question data, sorting the keywords according to the importance of character habits, and querying questions in the following way: sorting by the importance of character habits of the keywords, and querying questions in the order of the importance of character habits, that is, first querying the keyword with the greatest importance of character habits, and then querying the keyword with the second greatest importance of character habits on the basis of the query results, and so on, until several remaining questions are obtained;

[0023] Obtaining the answers to the questions based on the obtained several remaining questions.

[0024] In a second aspect, the present application provides an intelligent Q&A system based on deep learning, which is implemented based on the above-mentioned intelligent Q&A method based on deep learning, and specifically includes a data acquisition module, a questioning habit analysis module, a keyword acquisition module, a question answer module, and an answer transmission module; wherein, the data acquisition module is used to acquire the question data of the questioner, and at the same time acquire the historical question data and answer acquisition data of the questioner; the questioning habit analysis module constructs a deep learning neural network model based on the acquired historical question data and answer acquisition data of the questioner to analyze the questioning habits of the questioner; the keyword acquisition module acquires real-time question keywords according to the analyzed questioning habits of the questioner and the real-time question data of the questioner; the question answer module is used to search for question answers based on the obtained real-time question keywords and answer the corresponding question answers, and the answer transmission module is used to transmit the corresponding question answer to the questioner.

[0025] In a third aspect, the present application provides an electronic device, including: a processor and a memory, wherein, a computer program that can be called by the processor is stored in the memory;

[0026] The processor executes the above-mentioned intelligent Q&A method based on deep learning by calling the computer program stored in the memory.

[0027] In a fourth aspect, the present application provides a computer-readable storage medium storing instructions, which when run on a computer, cause the computer to execute the above-mentioned intelligent Q&A method based on deep learning.

[0028] Meanwhile, compared with the prior art, the technical effects and advantages of this application are as follows: A deep learning neural network model is constructed based on the obtained historical question data and answer acquisition data of the questioner, the questioner's question habits are analyzed, real-time question keywords are obtained according to the analyzed questioner's question habits and the real-time question data of the questioner, question answers are searched according to the obtained real-time question keywords, and corresponding question answers are provided. The corresponding question answer is transmitted to the questioner, the questioner's question habits are analyzed, the user's habitual colloquial words are replaced with recognizable language vocabulary, and at the same time, keywords of the question are obtained according to the questioner's question habits. Furthermore, content analysis of the question is carried out according to the importance degree of the keywords, which further improves the accuracy of question content recognition and thus the accuracy of question answering. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following described drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0030] Figure 1 It is a schematic diagram of the overall process of a wisdom question-answering method based on deep learning in this application;

[0031] Figure 2 It is a schematic diagram of the specific process of step 3 of a wisdom question-answering method based on deep learning in this application;

[0032] Figure 3 It is a schematic diagram of the overall framework of a wisdom question-answering system based on deep learning in this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The following will clearly and completely describe the technical solutions in the embodiments of this application with reference to the drawings in the embodiments of this application. Obviously, the described embodiments are only some embodiments of this application, rather than all embodiments. The following description of at least one exemplary embodiment is actually only illustrative and in no way constitutes a limitation to this application and its application or use.

[0034] In addition, the attached drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and thus repeated descriptions thereof will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0035] It should be understood that although terms such as "first" and "second" may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0036] Embodiment 1

[0037] To solve the technical problems raised in the background art, the present application provides a preferred embodiment: As Figure 1 - Figure 2 shown, a wisdom question-answering method based on deep learning, which includes the following specific steps:

[0038] Step 1, obtain the question data of the questioner, and at the same time obtain the historical question data of the questioner and the answer acquisition data;

[0039] In this embodiment, the question data of the questioner in Step 1 includes the constituent character data of the question asked in real time, the historical question data of the questioner in Step 1 includes the constituent character data of the questions asked historically, and at the same time the answer acquisition data in Step 1 is the question data corresponding to the final answer data obtained after the questioner's question, and the various data obtained are stored in the storage component correspondingly for easy retrieval;

[0040] At the same time, it should be pointed out in this embodiment that if the technical solution of the present disclosure involves personal information, before the product applying the technical solution of the present disclosure processes personal information, it has clearly informed the personal information processing rules and obtained the individual's independent consent. If the technical solution of the present disclosure involves sensitive personal information, before the product applying the technical solution of the present disclosure processes sensitive personal information, it has obtained the individual's separate consent and at the same time meets the requirement of "express consent". The personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed;

[0041] Step 2: Based on the obtained historical question data of the questioner and the answer acquisition data, construct a deep learning neural network model to analyze the questioner's questioning habits;

[0042] In this embodiment, analyzing the questioner's questioning habits includes the following specific steps:

[0043] Step 201: Obtain the constituent character data of the historical questions and the question data corresponding to the final answer data obtained after the questioner's question. Combine the constituent character data of the corresponding historical questions and the constituent character data of the questions corresponding to the corresponding final answer data;

[0044] Step 202: Obtain the constituent character data of the corresponding historical questions and the constituent character data of the questions corresponding to the corresponding final answer data, and construct a deep learning neural network model with the input being the constituent character data of the corresponding historical questions and the output being the constituent character data of the questions corresponding to the corresponding final answer data, that is, the questioner's question model;

[0045] The specific construction method is as follows: Divide the obtained constituent character data of the corresponding historical questions and the constituent character data of the questions corresponding to the corresponding final answer data into an 85% weight and bias training set and a 15% weight and bias test set; Input the 85% weight and bias training set into the deep learning neural network model for training to obtain an initial deep learning neural network model; Use the 15% weight and bias test set to test the initial deep learning neural network model, and output the optimal initial deep learning neural network model output that meets the accuracy of the constituent character data of the questions corresponding to the preset corresponding final answer data as the neural network model. Among them, the formula in the neural network model is: , where is the output k of the s-th neuron in the m + 1 layer, is the connection weight G between the j-th neuron in the m-th layer and the s-th neuron in the m + 1 layer, represents the input R of the j-th neuron in the m-th layer, represents the bias z of the linear relationship between the j-th neuron in the m-th layer and the s-th neuron in the m + 1 layer, represents the Sigmoid activation function, and w is the number of input neurons in the m-th layer of the neural network model;

[0046] Step 203: Further obtain the questioning habit model of the corresponding questioner, and replace the user's habitual spoken words with recognizable language vocabulary, such as replacing habitual spoken words like dialects with recognizable language vocabulary;

[0047] Step 3: Obtain real-time question keywords based on the analyzed questioning habits of the questioner and the real-time question data of the questioner;

[0048] In this embodiment, the acquisition of real-time problem keywords includes the following specific steps:

[0049] The obtained question habit model corresponding to the questioner and the real-time question data corresponding to the questioner are imported into the constructed question habit model. Through the continuous operation of the question habit model, the character data components of the corresponding real-time question data are obtained, and the character data components of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions are obtained;

[0050] Based on the occurrence times and occurrence times of the character data components of the real-time question data corresponding to the character data components of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, a character habit importance evaluation is performed. Among them, the calculation formula for the importance of the i-th character habit is: , where mi is the number of corresponding questions in which the i-th character appears in the character data components of the question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, sj is the number of times the i-th character appears in the j-th question data corresponding to the final answer data obtained after the corresponding questioner's historical questions, tm is the set time standard value for normalizing the time interval, and tj is the time interval from the j-th question data corresponding to the final answer data obtained after the corresponding questioner's historical questions to the current time;

[0051] Arrange the character data components of the corresponding real-time question data in descending order of character habit importance, and select a relatively large number of constituent characters as the keywords of the corresponding real-time question data;

[0052] Step 4: Search for the question answer according to the obtained real-time question keywords and answer the corresponding question answer;

[0053] In this embodiment, the search for question answers based on the obtained real-time question keywords includes the following specific contents: obtaining the keywords corresponding to the real-time question data, sorting them according to the importance of the characters of the keywords, and querying the questions in the following way: sorting by the importance of the characters of the keywords, and querying the questions in the order of the importance of the characters. That is, first query the keyword with the greatest importance of character habits, and then query the keyword with the second greatest importance of character habits based on the query results, and so on, until several remaining questions are finally obtained. For example, in the question of the meaning of choosing Matlab to draw graphics, the order of the importance of character habits is: Matlab, draw graphics, meaning, choose. First, query Matlab, and questions about Matlab are obtained. Then query draw graphics, and questions about draw graphics in the questions about Matlab are obtained. Then query meaning, and questions about meaning in the questions about Matlab drawing graphics are obtained. This can ensure the extraction effect of keywords and prevent key keywords from being lost during question answering, which may cause errors in the questions;

[0054] Obtain the answers to the questions based on the obtained several remaining questions;

[0055] Step 5: Transmit the corresponding question answer to the questioner.

[0056] Finally, the advantages of this embodiment are described here. This embodiment constructs a deep learning neural network model based on the obtained historical question data and answer acquisition data of the questioner, analyzes the questioner's question habits, obtains real-time question keywords according to the analyzed questioner's question habits and the questioner's real-time question data, searches for question answers based on the obtained real-time question keywords, and answers the corresponding question answers. Transmit the corresponding question answer to the questioner, analyze the questioner's question habits, replace the user's habitual oral words with recognizable language words, and at the same time obtain the keywords of the question according to the questioner's question habits, and then analyze the content of the question according to the importance of the keywords, further improving the accuracy of question content recognition and thus the accuracy of question answering.

[0057] Embodiment 2

[0058] Such as Figure 3As shown in the figure, this embodiment provides an intelligent Q&A system based on deep learning, which is implemented based on the above-mentioned intelligent Q&A method based on deep learning. Specifically, it includes a data acquisition module, a questioning habit analysis module, a keyword acquisition module, a question answer module, and an answer transmission module. Among them, the data acquisition module is used to acquire the question data of the questioner, and at the same time acquire the historical question data and answer acquisition data of the questioner. The questioning habit analysis module constructs a deep learning neural network model based on the acquired historical question data and answer acquisition data of the questioner to analyze the questioning habits of the questioner. The keyword acquisition module acquires real-time question keywords according to the analyzed questioning habits of the questioner and the real-time question data of the questioner. The question answer module is used to search for question answers according to the obtained real-time question keywords and answer the corresponding question answers. The answer transmission module is used to transmit the corresponding question answer to the questioner. In this embodiment, the specific steps of each module have been elaborated in detail in the method embodiment of the above-mentioned embodiment 1 and will not be repeated here. For the arrow directions in the appendix Figure 3 The arrow directions in the appendix represent the data transmission directions.

[0059] Embodiment 3

[0060] This embodiment provides an electronic device, including: a processor and a memory. Among them, a computer program that can be called by the processor is stored in the memory;

[0061] The processor executes the above-mentioned intelligent Q&A method based on deep learning by calling the computer program stored in the memory.

[0062] This electronic device may have relatively large differences due to different configurations or performances, and can include one or more processors and one or more memories. Among them, at least one computer program is stored in the memory, and this computer program is loaded and executed by the processor to implement the intelligent Q&A method based on deep learning provided by the above-mentioned method embodiment. This electronic device can also include other components for implementing the functions of the device. For example, this electronic device can also have components such as wired or wireless network interfaces and input / output interfaces for data input and output. This embodiment will not be elaborated here.

[0063] Embodiment 4

[0064] This embodiment proposes a computer-readable storage medium, on which a rewritable computer program is stored;

[0065] When the computer program runs on a computer device, it enables the computer device to execute the above-mentioned intelligent Q&A method based on deep learning.

[0066] For example, a computer-readable storage medium can be a read-only memory, a random access memory, a compact disc read-only memory, magnetic tape, a floppy disk, and an optical data storage device, etc.

[0067] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions according to the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired network or / and a wireless network. The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or a data center that contains one or more collections of available media. The available media can be a magnetic medium (e.g., a floppy disk, a hard disk, magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0068] The term "comprising", "including", or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, method, article, or apparatus.

[0069] In addition, parts of the above technical solutions provided in the embodiments of the present application that are consistent with the corresponding technical solutions in the prior art in terms of implementation principles are not described in detail to avoid excessive elaboration.

[0070] As described above in the specific embodiments, the objectives, technical solutions, and beneficial effects of the present invention have been further described in detail. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A smart question-answering method based on deep learning, characterized in that: It includes the following specific steps: Get the question data of the questioner, as well as the historical question data and answer acquisition data of the questioner; A deep learning neural network model is built based on the historical question data and answer acquisition data of the questioner to analyze the questioning habits of the questioner; Acquire real-time question keywords based on the questioner's questioning habits and real-time question data obtained through analysis; Search for answers to questions based on the real-time question keywords obtained, and provide answers to the corresponding questions; The answer to the corresponding question is transmitted to the questioner.

2. The deep learning-based intelligent question-answering method according to claim 1, characterized in that: The analysis of the questioner's questioning habits comprises the following specific steps: Obtain the question data corresponding to the character data of the historical questions and the final answer data obtained after the questioner asks the question, and convert the character data of the historical questions and the character data of the final answer data into the character data of the question; Obtain the constituent character data of historical questions and the constituent character data of questions corresponding to final answer data to construct a deep learning neural network model whose input is the constituent character data of historical questions and output is the constituent character data of questions corresponding to final answer data, i.e., the questioner questioning model.

3. The deep learning-based intelligent question-answering method according to claim 2, characterized in that: The deep learning neural network model is constructed as follows: the constituent character data of the historical questions and the constituent character data of the questions corresponding to the final answer data are obtained and divided into an 85% weighted and biased training set and a 15% weighted and biased test set; the 85% weighted and biased training set is input into the deep learning neural network model for training to obtain an initial deep learning neural network model; the initial deep learning neural network model is tested using the 15% weighted and biased test set, and the optimal initial deep learning neural network model output that meets the preset accuracy of the constituent character data of the questions corresponding to the final answer data is output as the neural network model.

4. The deep learning-based intelligent question-answering method according to claim 3, characterized in that: The acquisition of the real-time question keywords includes the following specific steps: The obtained question habit model of the corresponding questioner and the real-time question data of the corresponding questioner are imported into the constructed question habit model to obtain the constituent character data of the corresponding real-time question data, and the constituent character data of the question data corresponding to the final answer data obtained after the historical questioning of the corresponding questioner is obtained; Based on the number of occurrences and the time of occurrence of the constituent character data corresponding to the real-time question data in the constituent character data of the question data corresponding to the final answer data obtained after the historical question of the corresponding questioner, the importance of the character habit is evaluated; The component character data corresponding to the real-time question data are arranged according to the importance of the character habits, and a number of relatively large component characters are selected as keywords corresponding to the real-time question data.

5. The deep learning-based intelligent question-answering method according to claim 4, characterized in that: The searching for answers to questions based on the obtained real-time question keywords includes the following specific contents: obtaining keywords corresponding to the real-time question data, and querying the questions according to the order of importance of the characters of the keywords; Get answers to the questions based on the remaining questions obtained.

6. A deep learning-based intelligent question-answering method as claimed in claim 5, characterized in that: The specific content of the query of the problem is: Keywords are sorted by their character habit importance, and questions are searched in order of character habit importance, that is, the keyword with the greatest character habit importance is searched first, and then the keyword with the second greatest character habit importance is searched based on the search results, and so on, eventually obtaining the remaining questions.

7. A smart question-answering system based on deep learning, which is implemented based on the smart question-answering method based on deep learning as claimed in any one of claims 1 to 6, characterized in that: It specifically includes a data acquisition module, a question habit analysis module, a keyword acquisition module, a question answer answer module and an answer return module; wherein the data acquisition module is used to obtain the question data of the questioner, and at the same time obtain the questioner's historical question data and answer acquisition data; the question habit analysis module builds a deep learning neural network model based on the acquired questioner's historical question data and answer acquisition data, and analyzes the questioner's questioning habits; the keyword acquisition module acquires real-time question keywords based on the questioner's questioning habits and the questioner's real-time question data obtained by analysis; the question answer answer module is used to search for question answers based on the obtained real-time question keywords, and answer the corresponding question answers, and the answer return module is used to transmit the corresponding question answers to the questioner.

8. An electronic device comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; It is characterized in that the processor executes the deep learning-based intelligent question-answering method as described in any one of claims 1 to 6 by calling the computer program stored in the memory.

9. A computer-readable storage medium, characterized in that: Instructions are stored, and when the instructions are executed on a computer, the computer executes a smart question-answering method based on deep learning as described in any one of claims 1 to 6.

Images