CN108897867B - Data processing method, device, server and medium for knowledge question answering - Google Patents
Data processing method, device, server and medium for knowledge question answering Download PDFInfo
- Publication number
- CN108897867B CN108897867B CN201810712464.3A CN201810712464A CN108897867B CN 108897867 B CN108897867 B CN 108897867B CN 201810712464 A CN201810712464 A CN 201810712464A CN 108897867 B CN108897867 B CN 108897867B
- Authority
- CN
- China
- Prior art keywords
- target
- intention
- answer
- key information
- question
- 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
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
- G06Q40/03—Credit; Loans; Processing thereof
Landscapes
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Finance (AREA)
- Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- Technology Law (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention provides a data processing method, a device, a server and a medium for knowledge question answering, wherein the method comprises the following steps: the method comprises the steps of carrying out intention recognition on received target query voice to obtain a target intention and current key information, determining whether the current key information meets query requirements aiming at the target intention or not based on a pre-constructed intention frame, obtaining the target key information corresponding to the query requirements in a man-machine interaction mode if the current key information does not meet the query requirements, and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information. The embodiment of the invention solves the problem that the real intention of the user cannot be directly focused when the answer is provided for the user, and can improve the accuracy and the integrity of the intention identification of the user, so that the answer is found in the credit card knowledge base and fed back to the user, and the user requirement is accurately met.
Description
Technical Field
The embodiment of the invention relates to the technical field of computers, in particular to a data processing method, a data processing device, a data processing server and a data processing medium for knowledge question answering.
Background
Human-computer interaction is the science of studying the interactive relationships between systems and users. The system may be a variety of machines, and may be a computerized system and software. Various artificial intelligence systems can be realized through man-machine interaction, such as an intelligent customer service system, a voice control system and the like.
In various human-computer interaction products, such as a question-answering system for the credit card field, it is important to fully understand the user's true intention and true background information to give professional accurate answers and suggestions. Currently, it is common for a specialized expert to answer various credit card related questions of a user on a particular platform and then provide references to other users by way of a search. However, when the user's question is not clear enough, it is often necessary to clarify all questions by searching, so that the answer becomes very long and cannot directly focus on the user's true intention.
Disclosure of Invention
The embodiment of the invention provides a data processing method, a data processing device, a data processing server and a data processing medium for knowledge question answering, and aims to solve the problem that the real intention of a user cannot be directly focused when answers are provided for the user in the prior art.
In a first aspect, an embodiment of the present invention provides a data processing method for knowledge question answering, where the method includes:
carrying out intention recognition on the received target query voice to obtain a target intention and current key information;
determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame;
if the query requirement is not met, acquiring target key information corresponding to the query requirement in a man-machine interaction mode;
and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
In a second aspect, an embodiment of the present invention further provides a data processing apparatus for knowledge question answering, where the apparatus includes:
the intention identification module is used for carrying out intention identification on the received target query voice to obtain a target intention and current key information;
the key information determining module is used for determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein the intention frame defines the intention and a key information set corresponding to the requirement of meeting the intention;
the key information acquisition module is used for acquiring target key information corresponding to the query requirement in a man-machine interaction mode if the current key information does not meet the query requirement;
and the answer searching module is used for searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
In a third aspect, an embodiment of the present invention further provides a server, where the server includes:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors implement the data processing method for a knowledge question and answer according to any embodiment of the present invention.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the data processing method for knowledge question answering according to any one of the embodiments of the present invention.
The embodiment of the invention provides a data processing method for knowledge question answering, which comprises the steps of carrying out intention identification on received target query voice to obtain a target intention and current key information, determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein the intention frame defines an intention and a key information set corresponding to the query requirement, if the query requirement is not met, obtaining the target key information corresponding to the query requirement in a man-machine interaction mode, and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information. The embodiment of the invention solves the problem that the user cannot directly focus on the real intention of the user when providing answers for the user in the prior art, and can improve the accuracy and integrity of the intention identification of the user, so that the answers are found in the credit card knowledge base and fed back to the user, and the user requirements are accurately met.
Drawings
FIG. 1 is a flow chart of a data processing method for knowledge question answering according to one embodiment of the present invention;
FIG. 2 is a flow chart of a data processing method for knowledge question answering according to a second embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a data processing apparatus for knowledge question answering according to a third embodiment of the present invention;
fig. 4 is a schematic structural diagram of a server provided in the fourth embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples. It is to be understood that the specific embodiments described herein are merely illustrative of the invention and are not limiting of the invention. It should be further noted that, for the convenience of description, only some of the structures related to the present invention are shown in the drawings, not all of the structures.
Example one
Fig. 1 is a flowchart of a data processing method for a question and answer in an embodiment of the present invention, which is applicable to a case where an answer is found for a question in a human-computer interaction question and answer scenario, for example, a human-computer interaction question and answer scenario in the field of credit cards. The method may be performed by a data processing apparatus for knowledge question answering, which may be implemented in software and/or hardware, and may be integrated on any server having network communication functions.
As shown in fig. 1, the data processing method for knowledge question answering may include:
s101, performing intention recognition on the received target query voice to obtain a target intention and current key information.
Various artificial intelligence systems can be realized through man-machine interaction, such as an intelligent customer service system, a voice control system and the like. The user can ask questions by using a human-computer interaction platform similar to an intelligent customer service system, a voice control system and the like, and after the user asks the questions, the human-computer interaction platform can give corresponding answers through the server. In other words, the user inputs a corresponding "question" to the human-computer interaction platform, and then the human-computer interaction platform can output a corresponding "answer according to the" question "input by the user. As many as thousands of credit cards are introduced into various large banks and small banks in the market, and each credit card has its own characteristics, so that the user is unable to properly select the credit card, and at this time, the user needs to provide professional and accurate answers and suggestions by fully understanding the real intention and the real background information of the user.
In the embodiment of the present invention, the target query speech may be understood as a "question" input by the user, and the data processing apparatus for knowledge question answering may receive the target query speech input by the user and perform intent recognition on the received target query speech. By performing intention recognition on the received target query voice, the target intention and current key information of the user can be mined from the target query voice input by the user. The target intention may be understood as the user's true intention in the target query speech input by the user. Illustratively, the question and answer for the credit card field may be divided into a plurality of user intentions, such as "card application flow", "card type selection", "card application condition", etc., but the target query voice input by the user may not include all of the intentions, and for this purpose, the intention recognition processing needs to be performed on the target query voice input by the user to determine the real target intention of the user.
In the embodiment of the invention, if an accurate 'answer' is required to be output aiming at the target intention of the user, corresponding key information needs to be mined from the target query voice input by the user, so that when the received target query voice is subjected to intention recognition to obtain the target intention, the current key information contained in the target query voice can be mined. For example, taking questions and answers in the field of credit cards as an example, if the expressed target intention of the user in the target query voice input by the user is a "card declaration condition", the card declaration conditions of credit cards in different banks that do not pass through the card are different, in order to output an accurate "answer" to the user, the user is required to explain which credit card and other related information of which bank needs to be queried, and the information can be generally included in the target query voice input by the user, and corresponding current key information can be obtained only by mining and analyzing the target query voice.
In the embodiment of the invention, when the target query voice is received, the data processing device for knowledge question answering can convert the target query voice into the corresponding target query text information, perform voice analysis on the target query text information by adopting a preset semantic analysis model, and acquire the target intention of the user and the current key information contained in the target query voice from the target query text information.
S102, determining whether current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame.
In the embodiment of the present invention, after the target intent and the current key information are obtained from the target query speech, the current key information may or may not satisfy the query requirement for the target intent. For example, it is assumed that a target intention included in a target query voice input by a user is a "card application condition", and current key information is "applying for a credit card today", and at this time, the current key information does not satisfy a query requirement for the target intention, in other words, the information cannot be answered for the "card application condition" according to the "applying for a credit card today"; assuming that a target intention included in a target query voice input by a user is a "card application condition", and current key information is a "XXX bank yyyy credit card", the current key information can meet a query requirement for the target intention, in other words, the information of the "XXX bank YYY credit card" can be used for answering the user for the "card application condition". Based on the above, since it is unclear whether the current key information obtained from the target query speech is the query requirement information that satisfies the target intention, it can be determined whether the current key information satisfies the query requirement for the target intention based on the previously constructed intention framework. The intention framework defines the intention and the key information set corresponding to the requirement for satisfying the inquiry. Whether the current key information meets the query requirement of the target intention can be judged through the intention defined in the pre-constructed intention frame and the key information set corresponding to the requirement of meeting the query requirement.
S103, if the query requirement is not met, acquiring target key information corresponding to the query requirement in a man-machine interaction mode.
In the embodiment of the invention, if the current key information is determined to be incapable of meeting the query requirement aiming at the target intention, an inquiry is initiated to the user in a man-machine interaction mode, the inquiry is actively made to the user through a plurality of times of man-machine interaction, the key information which corresponds to the target intention and can meet the query requirement is inquired to be used as the target key information, the target intention and the target key information in the real intention of the user are determined through a plurality of times of on-line interaction, and the accuracy and the integrity of intention identification of the received target query voice are improved in a man-machine interaction mode of obtaining the target key information. If the current key information is determined to meet the query requirement aiming at the target intention, the current key information obtained by performing intention recognition on the received target query voice is used as the target key information corresponding to the query requirement aiming at the target intention, and at the moment, the user is not asked any more in a man-machine interaction mode.
And S104, searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
In the embodiment of the present invention, a pre-established knowledge base may store a pre-established knowledge question and answer data structure, and the knowledge question and answer data structure may store at least one intention, a corresponding answer, and key information required for an intention query. And mapping relations exist among all intentions, corresponding answers and key information required by the intention query. After determining the target intention and the target key information, the target intention and the target key information may be input into a pre-established knowledge base, and at least one corresponding answer matching the target intention and the target key information may be searched in the pre-established knowledge base as a target answer corresponding to the target query voice. That is, the knowledge base established in advance in the embodiment of the present invention may be understood as an answer set for answering a question posed by a user by inputting a target query voice, and in the case of clarifying a real intention of the user, authoritative answers to the question are given. The target answer corresponding to the target query voice found out can be fed back to the user in a voice or text mode, and the specific feedback is not limited in detail here.
In addition, taking the knowledge question answering in the credit card field as an example, the data processing method for the knowledge question answering provided by the embodiment of the invention can help the user to improve the acquisition efficiency of the related information of the credit card in the credit card field through a multi-round interactive conversation mode, and finally guides the user to apply for the credit card on a related platform. The related platform can be understood as a credit card knowledge question and answer platform integrating question and answer, encyclopedia and social interaction, such as a Baidu feed-bot platform.
The embodiment of the invention provides a data processing method for knowledge question answering, which comprises the steps of carrying out intention identification on received target query voice to obtain a target intention and current key information, determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein the intention frame defines an intention and a key information set corresponding to the query requirement, if the query requirement is not met, obtaining the target key information corresponding to the query requirement in a man-machine interaction mode, and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information. The embodiment of the invention solves the problem that the user cannot directly focus on the real intention of the user when providing answers for the user in the prior art, and can improve the accuracy and integrity of the intention identification of the user, so that the answers are found in the credit card knowledge base and fed back to the user, and the user requirements are accurately met.
Example two
Fig. 2 is a flowchart of a data processing method for a knowledge question and answer according to a second embodiment of the present invention, which is further optimized based on the above embodiments.
As shown in fig. 2, the data processing method for knowledge question answering may include:
s201, performing intention recognition on the received target query voice to obtain a target intention and current key information.
In the embodiment of the invention, a large number of training samples are adopted for training in a machine learning model or rule template-based mode, and an intention recognition model or an intention recognition rule template for performing intention recognition on target query voice input by a user is constructed, so that the intention recognition on the target query voice is realized through the constructed intention recognition model or the intention recognition rule template, and the target intention and the current key information are obtained. Illustratively, taking a question and answer in the field of credit cards as an example, an intention identification model facing to a credit card question and answer is constructed by adopting a large-scale training sample based on a UNIT platform (a Baidu self-built user intention understanding machine learning platform), so that intention identification of user input contents is realized. In addition, in order to ensure the accuracy of the intention recognition model or the intention recognition rule template, the intention recognition model or the intention recognition rule template may be verified by using verification data at a preset time, and parameters of the intention recognition model or the intention recognition rule template may be modified according to a verification result to obtain a modified intention recognition model or the intention recognition rule template.
S202, extracting at least one keyword corresponding to each intention from the historical intention sample.
In the embodiment of the invention, each intention needs one or more keywords as parameters for distinguishing different complaints of the user, the intention of the user can be more definite through the keywords corresponding to each intention, and the accurate answer can be conveniently found according to the content input by the user. Illustratively, also taking questions and answers in the field of credit cards as an example, the historical intention sample of the embodiment may be understood as what the user says statistically when asking for relevant questions of the credit card, the intention extracted from the historical intention sample may be understood as "card claiming conditions", and the corresponding keywords may be understood as "XXX bank" and "yyyy credit card", by which a more accurate query may be made for the intention of the user. The accuracy of the user intention query can be improved through the assistance of the keywords corresponding to the intentions, so that at least one keyword corresponding to each intention can be extracted from a historical intention sample, and the intention can be queried through the extracted keywords as the assistance.
S203, mining at least one keyword corresponding to each intention by a semantic clustering method to obtain a key information set corresponding to each intention, and constructing an intention frame.
In the embodiment of the present invention, when a user asks a question, the user may ask a question using a synonym similar to a keyword corresponding to an intention, and the synonym may be understood as an expression word that is different in terms but intended to express the same thing. At this time, since the keyword corresponding to each intention extracted from the historical intention sample may be limited, if the user asks a question using a synonym corresponding to the keyword, an accurate answer to the user's intention may not be made. Based on the above situation, after at least one keyword corresponding to each intention is extracted from the historical intention sample, a word vector of each keyword can be constructed, and then each keyword corresponding to each intention can be mined by a semantic clustering method, so that more keywords corresponding to each intention can be mined. And constructing a keyword set corresponding to each intention according to the mined more keywords corresponding to each intention, wherein each intention can correspond to one keyword set. After obtaining the key information sets corresponding to each intention, an intention framework can be constructed by using each intention and the key information sets corresponding to each intention. The intention frame comprises each intention and a key information set corresponding to the fact that the intention meets the query requirement of the intention respectively.
S204, determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame.
In the embodiment of the invention, the current key information obtained from the target query voice may or may not meet the query requirement for the target intention. Comparing the current key information with key words in a key information set which is defined in an intention frame and corresponds to the requirement for meeting the target intention query, and if the current key information is consistent with the key words in the key information set which corresponds to the requirement for meeting the target intention query, namely the current key information contains the key words which are necessary in the key information set which corresponds to the requirement for meeting the target intention query, considering that the current key information meets the query requirement aiming at the target intention; and if the current key information is inconsistent with the key words in the key information set which is defined in the intention frame and corresponds to the requirement for meeting the target intention query, namely the current key information does not have the necessary key words in the key information set which corresponds to the requirement for meeting the target intention query, the current key information is considered not to meet the query requirement for the target intention.
In addition, the key information set defined in the intent framework and corresponding to the requirement for satisfying the target intent query may contain multiple types of keywords, for example, the target intent is "card declaration condition", and the key information set corresponding to the requirement for satisfying the target intent query needs to contain two types of keywords, i.e., "XXX bank" and "yyyy credit card". At this time, when comparing the current key information with the keywords in the key information set defined in the intent frame and corresponding to the requirement for satisfying the target intent query, the keywords of each type included in the current key information need to be compared with the keywords of each type in the key information set corresponding to the requirement for satisfying the target intent query one by one. If the keywords of each type in the current key information lack any type of keywords in the key information set corresponding to the query requirement meeting the target intention, the current key information is considered to not meet the query requirement aiming at the target intention; and if the keywords of each type in the current key information hit the keywords of each type in the key information set corresponding to the query requirement meeting the target intention, the current key information is considered to meet the query requirement aiming at the target intention. It should be noted that, for each type of keyword in the key information set corresponding to the requirement of satisfying the target intent query, each type of keyword may have different meanings in the key information set, that is, may be a keyword with a synonym property.
For example, the keywords of each type in the key information set corresponding to the requirement of the target intention query may be understood as the keywords necessary for clarifying the true meaning of the target intention in the target query speech input by the user. Assuming that A, B, C three types of keywords are included in the key information set corresponding to the requirement of satisfying the target intention query, each type of keyword may be replaced by another term having the same expression as the keyword. And comparing the keywords of each type contained in the current key information with A, B, C types of keywords in the key information set corresponding to the requirement of satisfying the target intention query. If the current key information also contains A, B, C types of keywords, the current key information is considered to be consistent with the keywords in the key information set defined in the intention frame and corresponding to the requirement of satisfying the target intention query, and the current key information is determined to satisfy the query requirement aiming at the target intention; if the keywords of each type contained in the current key information lack A, B, C any of the three types of keywords, for example, the keywords of type a, then the current key information is considered to be inconsistent with the keywords in the key information set defined in the intent frame and corresponding to the requirement of satisfying the target intent query, and it is determined that the current key information does not satisfy the query requirement for the target intent.
And S205, if the query requirement is not met, acquiring the target key information corresponding to the query requirement in a man-machine interaction mode.
In the embodiment of the present invention, if any type of keyword among the types of keywords included in the current key information, which is necessary in the key information set corresponding to the query requirement for the target intent, is absent, for example, the type a keyword is absent, the current key information is considered to not satisfy the query requirement for the target intent. At the moment, the user can be asked questions through the conversation template, and keywords which are lacked in the current key information are obtained through multiple times of man-machine interaction, so that the target key information corresponding to the query requirement is obtained, and the effect of accurately identifying the user intention is achieved. Illustratively, the target intention is a "card declaration condition", two types of keywords, namely "XXX bank" and "YYY credit card" need to be included in a key information set corresponding to the requirement for satisfying the target intention query, and if the current key information only includes "XXX bank" but does not include "yy credit card", the current key information is considered to lack one type of keyword. At this time, question information "what type of credit card of XXX bank is asked for," YYY type credit card "can be sent to the user, and then the keyword in the current key information can be completed in a man-machine interaction manner, and the completed current key information is used as the target key information corresponding to the requirement for satisfying the query. Of course, the data processing device for knowledge question answering can also provide the user with selectable keywords for the user to refer to and select after sending the question information to the user. In addition, the current key information may not only lack one type of key words, but also lack multiple types of key words, and at this time, multiple times of questions are asked for the user through multiple times of man-machine interaction, so that the target key information corresponding to the query requirement can be obtained.
S206, searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
In the embodiment of the present invention, the knowledge base includes a historical question and answer pair and knowledge metadata, and accordingly, the construction operation of the knowledge base may include: the first construction mode is that the intention of each question in a historical question-answer pair is extracted, and the intention is mapped into an answer category; extracting a keyword set from each question of the historical question-answer pair according to a pre-established key information dictionary; and collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set. And/or, a second construction mode, namely, mapping the names of the knowledge metadata to obtain answer categories and keyword sets to which each knowledge metadata belongs; and collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set. The relevant question-answer pair set can comprise question-answer pairs and/or knowledge metadata, and the question-answer pairs and/or the knowledge metadata in the relevant question-answer pair set have the same answer category and keyword set.
In the embodiment of the invention, a knowledge base in the field of credit cards is constructed as an example, based on mass data of each large credit card issuing platform, the data sources of the knowledge base mainly comprise various credit card metadata, historical question and answer pair data and the like. The knowledge base constructed by combining the two construction modes can fully utilize the two types of data, and the recall rate of reliable answers is improved. The embodiment of the invention defines three field structures of 'answer category', 'keyword set' and 'relevant question and answer pair' to construct the knowledge base. The method for constructing the credit card metadata and the knowledge base of the historical question-answer pair data is respectively explained in detail below.
(1) For the first construction mode, the data source of the knowledge base is historical question and answer pair data.
Answer categories: and extracting intentions of all questions in the historical question-answer pairs, and mapping the extracted intentions into corresponding answer categories. Specifically, the answer category corresponding to each intention can be determined by identifying the intention of each question in the historical question-answer pair, classifying the answers in the historical question-answer pair according to the intention of each question, and mapping the intention to the answer category.
And (3) keyword set: and extracting a keyword set from each question of the historical question-answer pair according to a pre-established key information dictionary. Specifically, each intent requires one or more key information as a parameter to make explicit the user intent. For this purpose, a dictionary for maintaining each type of key information, such as "bank name dictionary", may be provided, and the keyword set is matched and extracted from each question of the history question-answer pair according to each type of key information dictionary.
The relevant question-answer pairs: and collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set. Specifically, after the answer category and the keyword set are obtained, all the historical question-answer pairs may be collected, so that each historical question-answer pair belonging to the same set has the same answer category and keyword set, and thus, a related question-answer pair may be obtained.
(2) For the second construction, the data source of the knowledge base is credit card metadata.
Answer categories: the answer category can be directly obtained through the mapping relation between the name of the metadata and the intention category. For example, the metadata of "year fee policy" is included in the metadata of "buying bank YOUNG card platinum card", and the "year fee policy" itself is also a kind of intentions that users are very interested in, and can be directly mapped to answer categories.
And (3) keyword set: the name of the metadata can be directly used as a keyword set and can be directly mapped without mining.
The relevant question-answer pairs: for the answer of the credit card metadata source, since the answer itself is the attribute value and has uniqueness, the relevant question is null, and there is only one answer. After the user question is analyzed into the intention and the key information set, if the metadata source answer can be hit, the metadata source answer can be directly extracted without further calculation and similarity of the question.
It should be noted that, the first construction method may be selected for constructing the knowledge base of the knowledge-based questioning and answering, or a combination of the two construction methods may be selected, and when the two construction methods are selected, the second construction method needs to be constructed according to data required by the actual knowledge-based questioning and answering field, which is not limited specifically here. For example, when a knowledge question and answer in a specific field is aimed, for example, the knowledge base construction in the credit card field can simultaneously select the combination of the first construction mode and the second construction mode, wherein the second construction mode selects the credit card metadata for construction.
On the basis of the foregoing embodiment, optionally, based on the target intention and the target key information, searching a target answer corresponding to the target query speech in a pre-established knowledge base, may include:
determining a corresponding target related question-answer pair set in a knowledge base based on the target intention and the target key information; taking an answer corresponding to a question of which the similarity of text information corresponding to the target query voice meets a preset threshold value as a target answer according to a text similarity algorithm from the target related question-answer pair set; and if the questions which meet the preset threshold value do not exist, using the knowledge metadata in the target-related question-answer pair set as the target answers.
Specifically, when a target answer corresponding to the target query speech is searched in a pre-established knowledge base, there may be more than one target answer. At this time, answers may be preferentially found in the question-answer pairs in the relevant question-answer pair set. If the question-answer pairs in the relevant question-answer pair set have answers meeting the conditions, the answers searched in the question-answer pairs can be used as final target answers; if no answer satisfying the condition exists in the question-answer pairs in the relevant question-answer pair set, the answer can be searched in the knowledge metadata in the relevant question-answer pair set. For example, when finding answers in question-answer pairs in the relevant question-answer pair set, it may be determined whether the similarity between the question of one question-answer pair and the target query speech input by the current user reaches a preset threshold through a text similarity comparison method, and if yes, the answer to the question of the question-answer pair is taken as the final target answer.
The embodiment of the invention provides a data processing method for knowledge question answering, which comprises the steps of carrying out intention identification on received target query voice to obtain a target intention and current key information, extracting at least one keyword corresponding to each intention from a historical intention sample, mining at least one keyword corresponding to each intention by a semantic clustering method to obtain a key information set corresponding to each intention, constructing an intention frame, determining whether the current key information meets the query requirement aiming at the target intention or not based on the pre-constructed intention frame, wherein the intention frame defines the intention and the key information set corresponding to the query requirement, if the intention frame does not meet the query requirement, obtaining the target key information corresponding to the query requirement by a human-computer interaction mode, and based on the target intention and the target key information, and searching a target answer corresponding to the target query voice in a pre-established knowledge base. The embodiment of the invention solves the problem that the real intention of the user cannot be directly hit when the answer is provided for the user in the prior art, not only can the model be used for judging whether the current key information meets the requirement or not, but also the accuracy and the integrity of the intention identification of the user can be improved on line through a multi-round clarification method through the construction of an intention frame, so that the answer is found in a credit card knowledge base and fed back to the user, and the requirement of the user is accurately met.
EXAMPLE III
Fig. 3 is a schematic structural diagram of a data processing apparatus for trivia provided in a third embodiment of the present invention, and this embodiment is applicable to a case of data processing for trivia, and the apparatus may be implemented in a software and/or hardware manner, and may be configured on any server with a network communication function.
As shown in fig. 3, the data processing apparatus for knowledge question answering may include: an intention identifying module 301, a key information determining module 302, a key information obtaining module 303 and an answer searching module 304, wherein:
an intention identifying module 301, configured to perform intention identification on the received target query speech to obtain a target intention and current key information;
a key information determining module 302, configured to determine whether the current key information meets a query requirement for a target intent based on a pre-constructed intent frame, where an intent and a key information set corresponding to meeting the query requirement are defined in the intent frame;
a key information obtaining module 303, configured to obtain, in a human-computer interaction manner, target key information corresponding to the query requirement if the current key information does not meet the query requirement;
and the answer searching module 304 is configured to search a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
On the basis of the above scheme, optionally, the data processing apparatus for knowledge question answering may further include an intention framework building module 305, and the intention framework building module 305 may specifically include: keyword extraction unit and keyword mining unit, wherein:
the keyword extraction unit is used for extracting at least one keyword corresponding to each intention from a historical intention sample;
and the keyword mining unit is used for mining at least one keyword corresponding to each intention by a semantic clustering method to obtain a key information set corresponding to each intention.
On the basis of the above scheme, optionally, the knowledge base may include historical question-answer pairs and knowledge metadata; accordingly, the data processing apparatus for knowledge question answering may further include: the knowledge base building module 306, wherein the knowledge base building module 306 specifically includes: the device comprises a first extraction unit, a second extraction unit, a mapping unit and a collection unit, wherein:
the first extraction unit is used for extracting the intention of each question in the historical question-answer pair and mapping the intention into an answer category;
the second extraction unit is used for extracting a keyword set from each question of the historical question-answer pair according to a pre-established key information dictionary;
the mapping unit is used for mapping the names of the knowledge metadata to obtain answer categories and keyword sets to which each knowledge metadata belongs;
and the collection unit is used for collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set, wherein the question-answer pairs and/or the knowledge metadata in the relevant question-answer pair set have the same answer categories and keyword sets.
Based on the above solution, optionally, the answer searching module 304 may include: a target-related question-answer pair set determining unit, a first target answer determining unit and a second target answer determining unit, wherein:
the target related question-answer pair set determining unit is used for determining a corresponding target related question-answer pair set in the knowledge base based on the target intention and the target key information;
a first target answer determining unit, configured to use, as the target answer, an answer corresponding to a question that a similarity of text information corresponding to the target query voice meets a preset threshold according to a text similarity algorithm from the target-related question-answer pair set;
and the second target answer determining unit is used for taking the knowledge metadata in the target-related question-answer pair set as the target answer if the question meeting the preset threshold value does not exist.
On the basis of the above scheme, optionally, the key information obtaining module 302 may be further configured to, if the current key information meets the query requirement, use the current key information as the target key information.
The data processing device for knowledge question answering provided by the embodiment of the invention can execute the data processing method for knowledge question answering provided by any embodiment of the invention, and has corresponding functional modules and beneficial effects of the execution method.
Example four
Fig. 4 is a schematic structural diagram of a server provided in the fourth embodiment of the present invention. FIG. 4 illustrates a block diagram of an exemplary server 412 suitable for use in implementing embodiments of the present invention. The server 412 shown in fig. 4 is only an example and should not bring any limitations to the function and scope of use of the embodiments of the present invention.
As shown in FIG. 4, server 412 is in the form of a general purpose server. Components of server 412 may include, but are not limited to: one or more processors 416, a storage device 428, and a bus 418 that couples the various system components including the storage device 428 and the processors 416.
A program/utility 440 having a set (at least one) of program modules 442 may be stored, for instance, in storage 428, such program modules 442 including, but not limited to, an operating system, one or more application programs, other program modules, and program data, each of which examples or some combination thereof may comprise an implementation of a network environment. The program modules 442 generally perform the functions and/or methodologies of the described embodiments of the invention.
The server 412 may also communicate with one or more external devices 414 (e.g., keyboard, pointing terminal, display 424, etc.), with one or more terminals that enable a user to interact with the server 412, and/or with any terminals (e.g., network card, modem, etc.) that enable the server 412 to communicate with one or more other computing terminals. Such communication may occur via input/output (I/O) interfaces 422. Further, server 412 may communicate with one or more networks (e.g., a Local Area Network (LAN), Wide Area Network (WAN), and/or a public Network, such as the Internet) via Network adapter 420. As shown in FIG. 4, network adapter 420 communicates with the other modules of server 412 via bus 418. It should be appreciated that although not shown in the figures, other hardware and/or software modules may be used in conjunction with the server 412, including but not limited to: microcode, end drives, Redundant processors, external disk drive Arrays, RAID (Redundant Arrays of Independent Disks) systems, tape drives, and data backup storage systems, among others.
The processor 416 executes various functional applications and data processing by executing programs stored in the storage device 428, for example, implementing a data processing method for question answering provided in any embodiment of the present invention, which may include:
carrying out intention recognition on the received target query voice to obtain a target intention and current key information;
determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame;
if the query requirement is not met, acquiring target key information corresponding to the query requirement in a man-machine interaction mode;
and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
EXAMPLE five
An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, implements a data processing method for knowledge question answering as provided in any embodiment of the present invention, and the method may include:
carrying out intention recognition on the received target query voice to obtain a target intention and current key information;
determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame;
if the query requirement is not met, acquiring target key information corresponding to the query requirement in a man-machine interaction mode;
and searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information.
Computer storage media for embodiments of the invention may employ any combination of one or more computer-readable media. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples (a non-exhaustive list) of the computer readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the context of this document, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
A computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.
Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
Computer program code for carrying out operations for aspects of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, Smalltalk, C + + or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or terminal. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet service provider.)
It is to be noted that the foregoing is only illustrative of the preferred embodiments of the present invention and the technical principles employed. It will be understood by those skilled in the art that the present invention is not limited to the particular embodiments described herein, but is capable of various obvious changes, rearrangements and substitutions as will now become apparent to those skilled in the art without departing from the scope of the invention. Therefore, although the present invention has been described in greater detail by the above embodiments, the present invention is not limited to the above embodiments, and may include other equivalent embodiments without departing from the spirit of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims (8)
1. A data processing method for question answering, the method comprising:
carrying out intention recognition on the received target query voice to obtain a target intention and current key information;
determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein an intention and a key information set corresponding to meeting the query requirement are defined in the intention frame;
if the query requirement is not met, acquiring target key information corresponding to the query requirement in a man-machine interaction mode;
searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information;
the knowledge base comprises historical question-answer pairs and knowledge metadata; correspondingly, the construction operation of the knowledge base comprises the following steps: extracting the intention of each question in the historical question-answer pair, and mapping the intention into an answer category; extracting a keyword set from each question of the historical question-answer pair according to a pre-established key information dictionary; mapping the names of the knowledge metadata to obtain answer categories and keyword sets to which each knowledge metadata belongs; collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set, wherein the question-answer pairs and/or the knowledge metadata in the relevant question-answer pair set have the same answer categories and keyword sets;
searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information, wherein the target answer comprises: determining a corresponding target related question-answer pair set in the knowledge base based on the target intention and the target key information; taking an answer corresponding to a question of which the similarity of the text information corresponding to the target query voice meets a preset threshold value as the target answer according to a text similarity algorithm from the target related question-answer pair set; and if the questions which meet the preset threshold value do not exist, using the knowledge metadata in the target-related question-answer pair set as the target answers.
2. The method of claim 1, wherein the construction operation of the intent framework comprises:
extracting at least one keyword corresponding to each intention from a historical intention sample;
and mining at least one keyword corresponding to each intention by a semantic clustering method to obtain a key information set corresponding to each intention.
3. The method of claim 1, further comprising:
and if the query requirement is met, taking the current key information as the target key information.
4. A data processing apparatus for knowledgeable question answering, the apparatus comprising:
the intention identification module is used for carrying out intention identification on the received target query voice to obtain a target intention and current key information;
the key information determining module is used for determining whether the current key information meets the query requirement aiming at the target intention or not based on a pre-constructed intention frame, wherein the intention frame defines the intention and a key information set corresponding to the requirement of meeting the intention;
the key information acquisition module is used for acquiring target key information corresponding to the query requirement in a man-machine interaction mode if the current key information does not meet the query requirement;
the answer searching module is used for searching a target answer corresponding to the target query voice in a pre-established knowledge base based on the target intention and the target key information;
the knowledge base comprises historical question-answer pairs and knowledge metadata; correspondingly, the device further comprises a knowledge base building module, which comprises: the first extraction unit is used for extracting the intention of each question in the historical question-answer pair and mapping the intention into an answer category; the second extraction unit is used for extracting a keyword set from each question of the historical question-answer pair according to a pre-established key information dictionary; the mapping unit is used for mapping the names of the knowledge metadata to obtain answer categories and keyword sets to which each knowledge metadata belongs; the collection unit is used for collecting historical question-answer pairs and knowledge metadata according to the answer categories and the keyword sets to obtain at least one relevant question-answer pair set, wherein the question-answer pairs and/or the knowledge metadata in the relevant question-answer pair set have the same answer categories and keyword sets;
wherein, the answer searching module comprises: the target related question-answer pair set determining unit is used for determining a corresponding target related question-answer pair set in the knowledge base based on the target intention and the target key information; a first target answer determining unit, configured to use, as the target answer, an answer corresponding to a question that a similarity of text information corresponding to the target query voice meets a preset threshold according to a text similarity algorithm from the target-related question-answer pair set; and the second target answer determining unit is used for taking the knowledge metadata in the target-related question-answer pair set as the target answer if the question meeting the preset threshold value does not exist.
5. The apparatus of claim 4, further comprising an intent framework building module comprising:
the keyword extraction unit is used for extracting at least one keyword corresponding to each intention from a historical intention sample;
and the keyword mining unit is used for mining at least one keyword corresponding to each intention by a semantic clustering method to obtain a key information set corresponding to each intention.
6. The apparatus according to claim 4, wherein the key information obtaining module is further configured to, if the current key information meets the query requirement, take the current key information as the target key information.
7. A server, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the data processing method for trivia as recited in any one of claims 1-3.
8. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, implements the data processing method for question answering according to any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810712464.3A CN108897867B (en) | 2018-06-29 | 2018-06-29 | Data processing method, device, server and medium for knowledge question answering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810712464.3A CN108897867B (en) | 2018-06-29 | 2018-06-29 | Data processing method, device, server and medium for knowledge question answering |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108897867A CN108897867A (en) | 2018-11-27 |
| CN108897867B true CN108897867B (en) | 2021-02-09 |
Family
ID=64347407
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810712464.3A Active CN108897867B (en) | 2018-06-29 | 2018-06-29 | Data processing method, device, server and medium for knowledge question answering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108897867B (en) |
Families Citing this family (43)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109783605B (en) * | 2018-12-14 | 2021-05-11 | 天津大学 | Scientific and technological service docking method based on Bayesian inference technology |
| CN111427444B (en) * | 2018-12-24 | 2022-05-10 | 中移（杭州）信息技术有限公司 | Control method and device of intelligent device |
| CN111368040B (en) * | 2018-12-25 | 2021-01-26 | 马上消费金融股份有限公司 | Dialogue processing method, model training method and related equipment |
| CN109840276A (en) * | 2019-02-12 | 2019-06-04 | 北京健康有益科技有限公司 | Intelligent dialogue method, apparatus and storage medium based on text intention assessment |
| CN110442695B (en) * | 2019-02-26 | 2022-03-01 | 杭州蓦然认知科技有限公司 | Intelligent interaction platform training method, device and equipment |
| CN111666006B (en) * | 2019-03-05 | 2022-01-14 | 京东方科技集团股份有限公司 | Method and device for drawing question and answer, drawing question and answer system and readable storage medium |
| CN110046177A (en) * | 2019-04-24 | 2019-07-23 | 北京思特奇信息技术股份有限公司 | A kind of method, system, storage medium and computer equipment that knowledge is extracted |
| CN111858829A (en) * | 2019-04-30 | 2020-10-30 | 广东小天才科技有限公司 | Method and device for sorting multiple intents and storage medium |
| CN112069267A (en) * | 2019-06-10 | 2020-12-11 | 阿里巴巴集团控股有限公司 | Data processing method and device |
| CN110232920B (en) * | 2019-06-21 | 2021-11-19 | 阿波罗智联(北京)科技有限公司 | Voice processing method and device |
| CN112115241B (en) * | 2019-06-21 | 2023-09-05 | 百度在线网络技术（北京）有限公司 | Question answering method, device and equipment |
| CN110321559B (en) * | 2019-06-24 | 2024-09-24 | 腾讯科技（深圳）有限公司 | Answer generation method, device and storage medium for natural language questions |
| CN112148939A (en) * | 2019-06-28 | 2020-12-29 | 阿里巴巴集团控股有限公司 | Data processing method and device and electronic equipment |
| CN110349463A (en) * | 2019-07-10 | 2019-10-18 | 南京硅基智能科技有限公司 | A kind of reverse tutoring system and method |
| CN112241448A (en) * | 2019-07-18 | 2021-01-19 | 阿里巴巴集团控股有限公司 | Response information generation method, device, equipment and storage medium |
| CN111881266B (en) * | 2019-07-19 | 2024-06-07 | 马上消费金融股份有限公司 | Response method and device |
| CN110444292B (en) * | 2019-07-29 | 2022-04-08 | 北京爱医生智慧医疗科技有限公司 | Information question-answering method and system |
| CN110442697B (en) * | 2019-08-06 | 2023-09-12 | 中电金信软件(上海)有限公司 | Man-machine interaction method, system, computer equipment and storage medium |
| CN110597951B (en) * | 2019-08-13 | 2024-01-05 | 平安科技（深圳）有限公司 | Text parsing method, text parsing device, computer equipment and storage medium |
| CN112395398A (en) * | 2019-08-16 | 2021-02-23 | 阿里巴巴集团控股有限公司 | Question and answer processing method, device and equipment |
| CN110784523B (en) * | 2019-10-11 | 2022-08-02 | 北京地平线机器人技术研发有限公司 | Target object information pushing method and device |
| CN112711653A (en) * | 2019-10-24 | 2021-04-27 | 华为技术有限公司 | Man-machine interaction method and electronic equipment |
| CN112925888B (en) * | 2019-12-06 | 2024-06-28 | 上海大岂网络科技有限公司 | Training method and device for question-answer response and small sample text matching model |
| CN111161068B (en) * | 2019-12-23 | 2022-06-14 | 支付宝(杭州)信息技术有限公司 | Transaction element checking method and device |
| CN111104501A (en) * | 2019-12-23 | 2020-05-05 | 中国银行股份有限公司 | Knowledge determination method and device for test service |
| CN112035631A (en) * | 2019-12-31 | 2020-12-04 | 北京来也网络科技有限公司 | Dialogue question-answering method, device, equipment and storage medium combining RPA and AI |
| CN111460095B (en) * | 2020-03-17 | 2023-06-27 | 北京百度网讯科技有限公司 | Question-answering processing method and device, electronic equipment and storage medium |
| CN111522927B (en) * | 2020-04-15 | 2023-07-14 | 北京百度网讯科技有限公司 | Entity query method and device based on knowledge graph |
| CN111858876B (en) * | 2020-05-14 | 2024-05-10 | 北京捷通华声科技股份有限公司 | Knowledge base generation method, text searching method and device |
| CN112035610B (en) * | 2020-08-28 | 2023-04-07 | 康键信息技术（深圳）有限公司 | Medical field question and answer pair generation method and device, computer equipment and medium |
| CN112035647B (en) * | 2020-09-02 | 2023-11-24 | 中国平安人寿保险股份有限公司 | Question and answer method, device, equipment and medium based on man-machine interaction |
| CN112035655B (en) * | 2020-09-02 | 2023-11-21 | 中国银行股份有限公司 | Bank manpower resource system and working method |
| CN112071436A (en) * | 2020-09-24 | 2020-12-11 | 上海依图网络科技有限公司 | Intelligent interaction method and device for epidemic situation and storage medium |
| CN112182177B (en) * | 2020-09-25 | 2024-10-01 | 中国建设银行股份有限公司 | User problem processing method and device, electronic equipment and storage medium |
| CN112182176A (en) * | 2020-09-25 | 2021-01-05 | 北京字节跳动网络技术有限公司 | Intelligent question answering method, device, equipment and readable storage medium |
| CN112527979A (en) * | 2020-11-10 | 2021-03-19 | 联想(北京)有限公司 | Information interaction method and device and readable storage medium |
| CN112287088A (en) * | 2020-11-20 | 2021-01-29 | 四川长虹电器股份有限公司 | Intelligent man-machine interaction query method, system, computer equipment and storage medium |
| CN112463920A (en) * | 2020-11-25 | 2021-03-09 | 联想(北京)有限公司 | Information response method and device |
| CN112394816B (en) * | 2020-11-26 | 2023-06-16 | 浙江连信科技有限公司 | Psychological service method based on man-machine interaction and robot |
| CN112463949B (en) * | 2020-12-01 | 2022-09-16 | 贝壳技术有限公司 | Data recall method and system, interaction method and interaction system |
| CN112597500A (en) * | 2020-12-08 | 2021-04-02 | 国汽(北京)智能网联汽车研究院有限公司 | Automobile information security risk assessment method and device, electronic equipment and storage medium |
| CN115086314B (en) * | 2022-06-09 | 2024-06-21 | 中国银行股份有限公司 | Interactive data processing method and related device |
| CN117151662B (en) * | 2023-08-31 | 2024-09-20 | 北京五八信息技术有限公司 | Position information processing method, device, equipment and storage medium based on AI |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015036945A (en) * | 2013-08-15 | 2015-02-23 | 株式会社インテリジェントウェイブ | Question answering control program, question answering control server, and question answering control method |
| CN107918634A (en) * | 2017-06-27 | 2018-04-17 | 上海壹账通金融科技有限公司 | Intelligent answer method, apparatus and computer-readable recording medium |
| CN107977415A (en) * | 2017-11-22 | 2018-05-01 | 北京寻领科技有限公司 | Automatic question-answering method and device |
| CN108170859A (en) * | 2018-01-22 | 2018-06-15 | 北京百度网讯科技有限公司 | Method, apparatus, storage medium and the terminal device of speech polling |
-
2018
- 2018-06-29 CN CN201810712464.3A patent/CN108897867B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015036945A (en) * | 2013-08-15 | 2015-02-23 | 株式会社インテリジェントウェイブ | Question answering control program, question answering control server, and question answering control method |
| CN107918634A (en) * | 2017-06-27 | 2018-04-17 | 上海壹账通金融科技有限公司 | Intelligent answer method, apparatus and computer-readable recording medium |
| CN107977415A (en) * | 2017-11-22 | 2018-05-01 | 北京寻领科技有限公司 | Automatic question-answering method and device |
| CN108170859A (en) * | 2018-01-22 | 2018-06-15 | 北京百度网讯科技有限公司 | Method, apparatus, storage medium and the terminal device of speech polling |
Non-Patent Citations (1)
| Title |
|---|
| 自动问答系统的研究与应用;王肖磊;《中国优秀硕士学位论文全文数据库》;20150316;第1-57页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108897867A (en) | 2018-11-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108897867B (en) | Data processing method, device, server and medium for knowledge question answering | |
| US10417581B2 (en) | Question answering system-based generation of distractors using machine learning | |
| US11392775B2 (en) | Semantic recognition method, electronic device, and computer-readable storage medium | |
| US20200098352A1 (en) | Techniques for model training for voice features | |
| US20190163691A1 (en) | Intent Based Dynamic Generation of Personalized Content from Dynamic Sources | |
| US20170330084A1 (en) | Clarification of Submitted Questions in a Question and Answer System | |
| CN112417090B (en) | Using uncommitted user input data to improve task performance | |
| CN106951503B (en) | Information providing method, device, equipment and storage medium | |
| US20180373702A1 (en) | Interactive method and apparatus based on test-type application | |
| US20190377824A1 (en) | Schemaless systems and methods for automatically building and utilizing a chatbot knowledge base or the like | |
| US10754882B2 (en) | Method of retrieving information from a health report through a machine assisted interrogation process | |
| US11699034B2 (en) | Hybrid artificial intelligence system for semi-automatic patent infringement analysis | |
| CN112559865B (en) | Information processing system, computer-readable storage medium, and electronic device | |
| CN116521841B (en) | Method, device, equipment and medium for generating reply information | |
| US11226832B2 (en) | Dynamic generation of user interfaces based on dialogue | |
| CN111339768A (en) | Sensitive text detection method, system, electronic device and medium | |
| WO2021174829A1 (en) | Crowdsourced task inspection method, apparatus, computer device, and storage medium | |
| US11907863B2 (en) | Natural language enrichment using action explanations | |
| US11423219B2 (en) | Generation and population of new application document utilizing historical application documents | |
| US10354013B2 (en) | Dynamic translation of idioms | |
| CN112749238A (en) | Search ranking method and device, electronic equipment and computer-readable storage medium | |
| KR102681147B1 (en) | Method and apparatus for generating appropriate responses based on the user intent in an ai chatbot through retrieval-augmented generation | |
| CN114186040A (en) | Operation method of intelligent robot customer service | |
| CN110263346B (en) | Semantic analysis method based on small sample learning, electronic equipment and storage medium | |
| CN113050933B (en) | Brain graph data processing method, device, equipment and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |