CN111339269A - Knowledge graph question-answer training and application service system with automatically generated template - Google Patents
Knowledge graph question-answer training and application service system with automatically generated template Download PDFInfo
- Publication number
- CN111339269A CN111339269A CN202010104143.2A CN202010104143A CN111339269A CN 111339269 A CN111339269 A CN 111339269A CN 202010104143 A CN202010104143 A CN 202010104143A CN 111339269 A CN111339269 A CN 111339269A
- Authority
- CN
- China
- Prior art keywords
- question
- query
- template
- backbone
- answer
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation or dialogue systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a knowledge graph question-answer training system with automatically generated templates, which comprises: the construction module of the predicate dictionary and the category dictionary is used for respectively constructing the predicate dictionary and the category dictionary in a remote supervision mode; the backbone query generation module is used for acquiring a sub-graph of the subject entity and the answer entity of each training question-answer pair in the knowledge graph, and replacing answer nodes in the sub-graph with variables to form a backbone query; a semantic alignment module; for aligning question phrases and backbone query semantic elements using dependency parsing and shaping linear alignment techniques; the template bloom module is used for storing the dependency syntax tree, the backbone query and the corresponding relation as templates into a template library; and the ranking model training module is used for performing classification learning on every two matching templates by using a machine learning two-classifier according to the matching degree to obtain a question template ranking model, so that the problems of high labor cost and low problem coverage rate in the prior art are solved.
Description
Technical Field
The application relates to the field of intelligent application, in particular to a knowledge-graph question-answer training system with an automatically generated template and a knowledge-graph question-answer application service system with an automatically generated template.
Background
The method based on the question-answering template plays an important role in knowledge-graph question-answering, semantic features of natural language questions of a user are extracted by using the modes of word segmentation, named entity recognition, predicate detection, category detection, question type classification, entity linking and the like, and the obtained semantic features are matched with question templates in a template library through a similarity or sequencing algorithm. After the template matching is successful, the query template (usually, SPARQL query statement) is instantiated by using the information of entities, categories and the like in the natural language question statement, then the knowledge query is executed, and the result is returned.
The knowledge-graph question-answering method based on the question-answering template can track the whole question-answering process clearly and can also realize question-answering of complex questions, but the traditional knowledge-graph question-answering method based on the template has the following two problems:
1. relying on manual template making requires a significant amount of labor cost.
2. It is difficult to ensure coverage of the problem.
Disclosure of Invention
The application provides a knowledge graph question-answer training and application service system with automatically generated templates, and solves the problems that in the prior art, labor cost is high and problem coverage rate is low.
The application provides a knowledge-graph question-answer training system of automatic generation of template, its characterized in that includes:
the construction module of the predicate dictionary and the category dictionary is used for respectively constructing the predicate dictionary and the category dictionary in a remote supervision mode;
the backbone query generation module is used for acquiring a sub-graph of the subject entity and the answer entity of each training question-answer pair in the knowledge graph, and replacing answer nodes in the sub-graph with variables to form a backbone query module;
the dependency syntax analysis and semantic role alignment module is used for analyzing sentences into a dependency syntax tree and describing the dependency relationship among all words; and the semantic role alignment module is used for mapping the phrases in the question sentence to the entities, the relations or the categories mentioned in the backbone query to form corresponding relations.
The template bloom module is used for removing the question dependency tree nodes and the backbone query semantic elements which are not mapped after semantic roles are aligned according to the corresponding relation among the dependency syntax tree, the backbone query, the question elements and the backbone query elements, and storing the dependency syntax tree, the backbone query and the corresponding relation as templates into a template library;
and the ranking model training module is used for performing classification learning on every two matching templates by using a machine learning two-classifier according to the matching degree to obtain a question template ranking model.
Preferably, the constructing of the predicate dictionary by using a remote supervision mode comprises the following steps:
for a relation p in the knowledge graph, representing all triples related to p in the knowledge graph by C (p) { (s, o): (s, p, o) ∈ K }, wherein K represents the knowledge graph;
if the two entities s and o in the step C (p) are detected in the same natural language description, extracting the intermediate language description r of the two entities in the text;
according to the assumption of remote supervision that if (s, p, o) is a triple in the knowledge-graph, then r represents p, a mapping (r → p) is added to the predicate dictionary LpPerforming the following steps;
and taking the quotient of the occurrence times of the mapping and the sum of the detected times of all the relations in the corpus as the weight of the mapping.
Preferably, the category dictionary is constructed by using a remote supervision mode, and the method comprises the following steps:
for class c in the knowledge graph, all entities of class c in the knowledge graph are represented by c (c) ═ { e (e type c) ∈ K };
the system searches on the corpus, and if an entity or other nominal phrases are detected, a mapping (np → c) condition dictionary library is obtained;
and taking the quotient of the occurrence times of the mapping and the sum of the detected times of all the relations in the corpus as the weight of the mapping.
Preferably, the backbone query generation module is configured to obtain a subgraph of the topic entity and the answer entity of each training question-answer pair in the knowledge graph, and replace an answer node in the subgraph with a variable to form the backbone query module, and includes:
for each training question-answer pair, detecting entity mentions in the question sentence by using a named entity recognition technology;
detecting subject entities mentioned by the entities in the knowledge graph through entity links;
obtaining a subgraph M of a subject entity and an answer entity in a knowledge graph through a shortest path algorithm;
adding the type nodes of all answer nodes into the subgraph M;
and replacing the answer nodes in the subgraph M by using variables to obtain a backbone query module in a SPARQL form.
Preferably, the method further comprises the following steps: dependency syntax analysis and semantic role alignment for obtaining the corresponding relationship between question phrases and backbone query semantic elements according to dependency syntax trees and shaping linear alignment, comprising:
performing dependency syntax analysis on the question to obtain a question dependency syntax analysis tree;
acquiring a question phrase permutation combination and a backbone query semantic element combination;
acquiring the weight of the question phrase by using the dictionary;
and aligning the question phrase and the backbone query semantic elements according to the shaping linear alignment.
Preferably, the template bloom module is configured to remove nodes that are not mapped after semantic roles of the question dependency tree are aligned according to a correspondence between the dependency syntax tree, the backbone query, the question element, and the backbone query element, and to store the dependency syntax tree, the backbone query, and the correspondence as templates in a template library after removing class nodes that are not mentioned in the correspondence in the backbone query, and includes:
according to the corresponding relation, removing nodes which are not mapped on the question dependency tree after semantic roles are aligned, replacing the mapping provided by the specific vocabulary on the dependency syntax tree according to the corresponding relation with semantic annotations, and reserving the part of speech information of the vocabulary and the side information of the tree;
removing class nodes which are not mentioned in the corresponding relation in the backbone query according to the corresponding relation, and replacing semantic element information with semantic annotations;
and storing the processed dependency syntax tree, the backbone query and the corresponding relation into a template library as templates.
Preferably, the ranking model training module is configured to use a machine learning two-classifier to perform classification learning on every two matching templates according to the matching degree, and obtain a question template ranking model, and includes:
and acquiring training characteristics, semantic role alignment characteristics, semantic characteristics and template characteristics.
Data training is performed using a machine learning model.
The application also provides a knowledge-graph question-answer application service system with automatically generated templates, which comprises:
the template matching module is used for analyzing the new question according to the dependency syntax to generate a question dependency syntax tree, and if the templates in the template library are subtrees of the question dependency syntax tree, the template matching is successful;
the template instantiation unit is used for instantiating the template according to the given question and the template set successfully matched;
the template sorting module is used for performing sorting prediction on the obtained question query pair serving as input data by using a sorting model and taking the query template with the highest score as an optimal template;
and the template query module is used for carrying out data query of the knowledge graph on the obtained optimal instantiation query statement and returning an answer.
Preferably, the template matching module is configured to analyze the dependency syntax of the new question to generate a question dependency syntax tree, and if the template in the template library is a subtree of the question dependency syntax tree, the template matching is successful, including:
analyzing the new question according to the dependency syntax to generate a question dependency syntax tree;
and if the templates in the template library are subtrees of the question-dependent syntax tree under the condition that only part of speech and side information are considered, the templates are successfully matched, and the templates are candidate templates of the question.
The application also provides a knowledge-graph question-answer system capable of automatically generating the template, which comprises a knowledge-graph question-answer training system capable of automatically generating the template and a knowledge-graph question-answer application service system capable of automatically generating the template.
The application provides a knowledge graph question-answer training and application service system capable of automatically generating templates, question templates and query templates are generated by automatically learning related simple question-answers of specific knowledge graphs on the basis of dependency syntax analysis, and meanwhile complex questions are answered by using the combination characteristics of dependency syntax analysis results and templates learned from the simple questions, so that the complete complex questions do not need to be captured by any specific template, and the problems of high labor cost and low problem coverage rate in the prior art are solved.
Drawings
Fig. 1 is a schematic structural diagram of a knowledge-graph question-answer training and application service system with an automatically generated template according to the present application.
Detailed Description
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. This application is capable of implementation in many different ways than those herein set forth and of similar import by those skilled in the art without departing from the spirit of this application and is therefore not limited to the specific implementations disclosed below.
In the schematic architecture diagram of a knowledge-graph question-answer training and application service system with automatically generated templates provided in fig. 1, a training system includes: the system comprises a predicate dictionary and category dictionary building module, a backbone query generating module, a dependency syntax analysis and semantic role alignment module, a template bloom module and a sequencing model training module.
And the predicate dictionary and category dictionary building module is used for building a predicate dictionary and a category dictionary respectively in a remote supervision mode. Because there are diversified natural language representations of the relationships or categories in the knowledge graph, the natural language description forms of all the relationships or categories cannot be collected in a manual enumeration manner. The system respectively constructs a predicate dictionary (relation dictionary) and a category dictionary by using a remote supervision mode.
Predicate dictionary LpRepresenting all triples related to p in the knowledge graph by C (p) { (s, o): s, p, o) ∈ K } aiming at the relation p in the knowledge graph, wherein K represents the knowledge graph, searching on the corpus by the system, extracting two entity intermediate language descriptions r in the text of a sentence if two entities of the s and the o in the C (p) are simultaneously detected in the same natural language description, and adding the mapping (r → p) to a predicate L by adding the mapping (r → p) to the predicate according to the assumption of remote supervision if the (s, p, o) is the triples in the knowledge graph, wherein the r represents the intermediate language description r in the two entities in the text of the sentencepPerforming the following steps; since the assumption of remote supervision is not always true, the system takes the quotient of the number of times the mapping occurs and the sum of the number of times all relationships in the corpus are detected as the weight of the mapping.
Category dictionary LcFor class c in the knowledge graph, all entities of class c in the knowledge graph are represented by C (c) { e (e type c) ∈ K }, the system searches on the corpus, and if entities or other nominal phrases such as 'e and other np' (where np represents nominal phrase) are detected, the mapping (np → c) is mapped to the conditional dictionary library Lc(ii) a Method for calculating weight of each entry and predicate dictionary Lp。
And the backbone query generation module is used for acquiring a sub-graph of the subject entity and the answer entity of each training question-answer pair in the knowledge graph, and replacing answer nodes in the sub-graph with variables to form the backbone query module. For each training question-answer pair, firstly, the entity mention in the question sentence is detected by using the technology of Named Entity Recognition (NER) and the like, and then the subject entity mentioned about the entity in the knowledge graph is detected through the entity link. And obtaining a subgraph M of the subject entity and the answer entity in the question in the knowledge graph through a shortest path algorithm. Since the type of answer entity plays a key role in the knowledge-graph question-answer, the present system adds the type nodes of all answer nodes to the subgraph M. And replacing the answer nodes in the subgraph M by using variables to obtain the backbone query in a SPARQL form.
And the dependency syntax analysis module is used for analyzing the sentences into a dependency syntax tree and describing the dependency relationship among the words, namely indicating the syntactic collocation relationship among the words, and the collocation relationship is associated with semantics. The method is suitable for the system based on machine learning and neural network dependency syntactic analysis; the semantic role alignment module is used to map the phrases in the question sentence to the mentioned entities, relationships, or categories in the backbone query. The purpose of semantic role alignment is to map phrases in a question to referenced entities, relationships, or categories in the backbone query. This alignment will determine which phrases in the question sentence can be mapped into the backbone query, which answer type constraints in the backbone query will be preserved, and the mapping relationship between the question sentence phrases and the backbone query semantic elements. Dependency syntax analysis and semantic role alignment for obtaining the corresponding relationship between question phrases and backbone query semantic elements according to dependency syntax trees and shaping linear alignment, comprising: performing dependency syntax analysis on the question to obtain a question dependency syntax analysis tree; acquiring a question phrase permutation combination and a backbone query semantic element combination; acquiring the weight of the question phrase by using the dictionary; and aligning the question phrase and the backbone query semantic elements according to the shaping linear alignment.
First, all sub-phrases in the question are used to form a phrase set Ph ═ { Ph ═ Ph1,ph2,...,phiDefining all element sets S in backbone queryq. Slave predicate dictionary LpAnd a category dictionary LcMiddle capture of phi∈ Ph to sj∈SqAll possible mappings of (g, g), mapping (ph)i→sj) W for the weight ofijIndicating that the weight is taken from a dictionary, XijRepresents the mappingWhether the value is kept to be 0 or 1. The goal of semantic role alignment is to maximize:
the constraints are:
where equation (2) represents that each backbone query semantic term is aligned by at most one phrase.
Equation (3) represents that once each phrase is aligned with an entity, any vocabulary in the phrase cannot be aligned with other backbone query elements, EjThe representative element j is an entity.
Equation (4) represents that the answer type constraint is at most one for intelligent retention, and s (v) represents the type element in the backbone query.
The short language elements in the question sentence and the semantic elements in the backbone query can be obtained by the formula and an integer linear programming means for optimal alignment.
The template bloom module is used for processing the components according to the corresponding relation mt among the dependency syntax tree, the backbone query, the question phrase and the backbone query semantic elements in order to enable the template to have the bloom capacity as follows:
A. removing nodes which are not mapped after semantic roles are aligned on the question dependency tree, replacing the mapping provided by the specific vocabulary on the dependency syntax tree according to mt with semantic annotations (entities, relations and categories), and keeping the part-of-speech information of the vocabulary and the side information of the tree.
B. Removing class nodes which are not mentioned in the corresponding relation mt in the backbone query, and replacing semantic element information with semantic annotations.
C. And storing the dependency syntax tree, the backbone query and the corresponding relation after the processing as templates into a template library.
And the ranking model training module is used for performing classification learning on every two matching templates by using a machine learning two-classifier according to the matching degree to obtain a question template ranking model. The training features include semantic role alignment features, semantic features, template features, and the like. Algorithms such as random forest, support vector machine, neural network and the like are all suitable for the system.
In a knowledge-graph question-answer training system with automatically generated templates, new question sentences are received and responded after training is finished, so that the application service system finishes responding to the new question sentences based on the knowledge-graph question-answer training system with automatically generated templates.
As shown in fig. 1, a knowledge-graph question-answering application service system with automatically generated templates includes: the template matching module, the template instantiation module, the template sorting module and the template inquiry module.
The template matching module is used for analyzing the new question according to the dependency syntax to generate a question dependency syntax tree; and if the templates in the template library are subtrees of the question-dependent syntax tree under the condition that only part of speech and side information are considered, the templates are successfully matched, and the templates are candidate templates of the question.
The template instantiation module is used for instantiating the template according to the given question and the template set successfully matched; the instantiation is to instantiate a query statement according to the mapping rule mt and the dictionary in the template, and replace semantic annotations (entities, relations and categories) in the query template in the template with specific semantic elements. The template sorting unit is used for performing sorting prediction on the obtained question query pair serving as input data by using the sorting model and taking the query template with the highest score as the optimal template;
and the template query module is used for carrying out data query of the knowledge graph on the obtained optimal instantiation query statement and returning an answer.
A knowledge-graph question-answer training system and an application service system which are automatically generated by a template form a knowledge-graph question-answer system which is automatically generated by the template.
The application provides a knowledge graph question-answer training and application service system capable of automatically generating templates, question templates and query templates are generated by automatically learning related simple question-answers of specific knowledge graphs on the basis of dependency syntax analysis, and meanwhile complex questions are answered by using the combination characteristics of dependency syntax analysis results and templates learned from the simple questions, so that the complete complex questions do not need to be captured by any specific template, and the problems of high labor cost and low problem coverage rate in the prior art are solved.
Finally, it should be noted that: although the present invention has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that various changes may be made and equivalents may be substituted for elements thereof without departing from the spirit and scope of the invention.
Claims (10)
1. A knowledge-graph question-answer training system with automatically generated templates is characterized by comprising:
the construction module of the predicate dictionary and the category dictionary is used for respectively constructing the predicate dictionary and the category dictionary in a remote supervision mode;
the backbone query generation module is used for acquiring a sub-graph of the subject entity and the answer entity of each training question-answer pair in the knowledge graph, and replacing answer nodes in the sub-graph with variables to form a backbone query module;
the dependency syntax analysis and semantic role alignment module is used for analyzing sentences into a dependency syntax tree and describing the dependency relationship among all words; and the semantic role alignment module is used for mapping the phrases in the question sentence to the entities, the relations or the categories mentioned in the backbone query to form corresponding relations.
The template bloom module is used for removing the question dependency tree nodes and the backbone query semantic elements which are not mapped after semantic roles are aligned according to the corresponding relation among the dependency syntax tree, the backbone query, the question elements and the backbone query elements, and storing the dependency syntax tree, the backbone query and the corresponding relation as templates into a template library;
and the ranking model training module is used for performing classification learning on every two matching templates by using a machine learning two-classifier according to the matching degree to obtain a question template ranking model.
2. The system of claim 1, wherein building a predicate dictionary using remote supervision comprises:
for a relation p in the knowledge graph, representing all triples related to p in the knowledge graph by C (p) { (s, o): (s, p, o) ∈ K }, wherein K represents the knowledge graph;
if the two entities s and o in the step C (p) are detected in the same natural language description, extracting the intermediate language description r of the two entities in the text;
according to the assumption of remote supervision that if (s, p, o) is a triple in the knowledge-graph, then r represents p, a mapping (r → p) is added to the predicate dictionary LpPerforming the following steps;
and taking the quotient of the occurrence times of the mapping and the sum of the detected times of all the relations in the corpus as the weight of the mapping.
3. The system of claim 1, wherein building a class dictionary using remote supervision comprises:
for class c in the knowledge graph, all entities of class c in the knowledge graph are represented by c (c) ═ { e (e type c) ∈ K };
the system searches on the corpus, and if an entity or other nominal phrases are detected, a mapping (np → c) condition dictionary library is obtained;
and taking the quotient of the occurrence times of the mapping and the sum of the detected times of all the relations in the corpus as the weight of the mapping.
4. The system of claim 1, wherein the backbone query generation module is configured to obtain a subgraph of the topic entities and the answer entities of each training question-answer pair in the knowledge graph, and replace answer nodes in the subgraph with variables to form the backbone query module, and the backbone query generation module comprises:
for each training question-answer pair, detecting entity mentions in the question sentence by using a named entity recognition technology;
detecting subject entities mentioned by the entities in the knowledge graph through entity links;
obtaining a subgraph M of a subject entity and an answer entity in a knowledge graph through a shortest path algorithm;
adding the type nodes of all answer nodes into the subgraph M;
and replacing the answer nodes in the subgraph M by using variables to obtain a backbone query module in a SPARQL form.
5. The system of claim 1, further comprising: dependency syntax analysis and semantic role alignment for obtaining the corresponding relationship between question phrases and backbone query semantic elements according to dependency syntax trees and shaping linear alignment, comprising:
performing dependency syntax analysis on the question to obtain a question dependency syntax analysis tree;
acquiring a question phrase permutation combination and a backbone query semantic element combination;
acquiring the weight of the question phrase by using the dictionary;
and aligning the question phrase and the backbone query semantic elements according to the shaping linear alignment.
6. The system according to claim 1, wherein the template bloom module is configured to remove nodes that are not mapped after semantic roles of the question dependency tree are aligned and remove category nodes that are not mentioned in the corresponding relationship in the backbone query according to the corresponding relationship among the dependency syntax tree, the backbone query, the question element, and the backbone query element, and store the dependency syntax tree, the backbone query, and the corresponding relationship as templates in the template library, including:
according to the corresponding relation, removing nodes which are not mapped on the question dependency tree after semantic roles are aligned, replacing the mapping provided by the specific vocabulary on the dependency syntax tree according to the corresponding relation with semantic annotations, and reserving the part of speech information of the vocabulary and the side information of the tree;
removing class nodes which are not mentioned in the corresponding relation in the backbone query according to the corresponding relation, and replacing semantic element information with semantic annotations;
and storing the processed dependency syntax tree, the backbone query and the corresponding relation into a template library as templates.
7. The system according to claim 1, wherein the ranking model training module is configured to perform classification learning on every two matching templates according to the matching degree by using a machine learning two-classifier, and obtain the question template ranking model, and includes:
acquiring training characteristics, semantic role alignment characteristics, semantic characteristics and template characteristics;
data training is performed using a machine learning model.
8. A knowledge graph question-answering application service system with automatically generated templates is characterized by comprising:
the template matching module is used for analyzing the new question according to the dependency syntax to generate a question dependency syntax tree, and if the templates in the template library are subtrees of the question dependency syntax tree, the template matching is successful;
the template instantiation module is used for instantiating the template according to the given question and the template set successfully matched;
the template sorting module is used for performing sorting prediction on the obtained question query pair serving as input data by using a sorting model and taking the query template with the highest score as an optimal template;
and the template query module is used for carrying out data query of the knowledge graph on the obtained optimal instantiation query statement and returning an answer.
9. The system according to claim 8, wherein the template matching module, configured to generate a question-dependent syntax tree for a new question based on dependency syntax analysis, and if the templates in the template library are subtrees of the question-dependent syntax tree, the template matching is successful, includes:
analyzing the new question according to the dependency syntax to generate a question dependency syntax tree;
and if the templates in the template library are subtrees of the question-dependent syntax tree under the condition that only part of speech and side information are considered, the templates are successfully matched, and the templates are candidate templates of the question.
10. A knowledge-graph question-answer system with an automatically generated template, which is characterized by comprising a knowledge-graph question-answer training system with an automatically generated template according to claim 1 and a knowledge-graph question-answer application service system with an automatically generated template according to claim 8.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010104143.2A CN111339269B (en) | 2020-02-20 | 2020-02-20 | Knowledge graph question-answering training and application service system capable of automatically generating templates |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010104143.2A CN111339269B (en) | 2020-02-20 | 2020-02-20 | Knowledge graph question-answering training and application service system capable of automatically generating templates |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111339269A true CN111339269A (en) | 2020-06-26 |
CN111339269B CN111339269B (en) | 2023-09-26 |
Family
ID=71185432
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010104143.2A Active CN111339269B (en) | 2020-02-20 | 2020-02-20 | Knowledge graph question-answering training and application service system capable of automatically generating templates |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111339269B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111897944A (en) * | 2020-08-17 | 2020-11-06 | 哈尔滨工业大学 | Knowledge map question-answering system based on semantic space sharing |
CN112632237A (en) * | 2020-12-07 | 2021-04-09 | 厦门渊亭信息科技有限公司 | Knowledge graph-based question-answer template automatic generation method and device |
CN112765312A (en) * | 2020-12-31 | 2021-05-07 | 湖南大学 | Knowledge graph question-answering method and system based on graph neural network embedding matching |
CN112989005A (en) * | 2021-04-16 | 2021-06-18 | 重庆中国三峡博物馆 | Knowledge graph common sense question-answering method and system based on staged query |
CN113066358A (en) * | 2021-04-14 | 2021-07-02 | 吴光银 | Science teaching auxiliary system |
CN113947084A (en) * | 2021-09-24 | 2022-01-18 | 盛景智能科技(嘉兴)有限公司 | Question-answer knowledge retrieval method, device and equipment based on graph embedding |
CN114417807A (en) * | 2022-01-24 | 2022-04-29 | 中国电子科技集团公司第五十四研究所 | Human-like language description expression method oriented to presence or absence of human collaboration scene |
CN114579710A (en) * | 2022-03-15 | 2022-06-03 | 西南交通大学 | Method for generating problem query template of high-speed train |
CN117743652A (en) * | 2023-12-16 | 2024-03-22 | 华南理工大学 | Weak supervision time sequence map question-answering method based on deep learning |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101908042A (en) * | 2010-08-09 | 2010-12-08 | 中国科学院自动化研究所 | Tagging method of bilingual combination semantic role |
CN105701253A (en) * | 2016-03-04 | 2016-06-22 | 南京大学 | Chinese natural language interrogative sentence semantization knowledge base automatic question-answering method |
CN107451230A (en) * | 2017-07-24 | 2017-12-08 | 宗晖(上海)机器人有限公司 | A kind of answering method and question answering system |
CN108363743A (en) * | 2018-01-24 | 2018-08-03 | 清华大学深圳研究生院 | A kind of intelligence questions generation method, device and computer readable storage medium |
CN110147436A (en) * | 2019-03-18 | 2019-08-20 | 清华大学 | A kind of mixing automatic question-answering method based on padagogical knowledge map and text |
CN110399457A (en) * | 2019-07-01 | 2019-11-01 | 吉林大学 | A kind of intelligent answer method and system |
CN110502621A (en) * | 2019-07-03 | 2019-11-26 | 平安科技(深圳)有限公司 | Answering method, question and answer system, computer equipment and storage medium |
CN110532358A (en) * | 2019-07-05 | 2019-12-03 | 东南大学 | A kind of template automatic generation method towards knowledge base question and answer |
CN110807091A (en) * | 2019-03-01 | 2020-02-18 | 王涵 | Hotel intelligent question-answer recommendation and decision support analysis method and system |
-
2020
- 2020-02-20 CN CN202010104143.2A patent/CN111339269B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101908042A (en) * | 2010-08-09 | 2010-12-08 | 中国科学院自动化研究所 | Tagging method of bilingual combination semantic role |
CN105701253A (en) * | 2016-03-04 | 2016-06-22 | 南京大学 | Chinese natural language interrogative sentence semantization knowledge base automatic question-answering method |
CN107451230A (en) * | 2017-07-24 | 2017-12-08 | 宗晖(上海)机器人有限公司 | A kind of answering method and question answering system |
CN108363743A (en) * | 2018-01-24 | 2018-08-03 | 清华大学深圳研究生院 | A kind of intelligence questions generation method, device and computer readable storage medium |
CN110807091A (en) * | 2019-03-01 | 2020-02-18 | 王涵 | Hotel intelligent question-answer recommendation and decision support analysis method and system |
CN110147436A (en) * | 2019-03-18 | 2019-08-20 | 清华大学 | A kind of mixing automatic question-answering method based on padagogical knowledge map and text |
CN110399457A (en) * | 2019-07-01 | 2019-11-01 | 吉林大学 | A kind of intelligent answer method and system |
CN110502621A (en) * | 2019-07-03 | 2019-11-26 | 平安科技(深圳)有限公司 | Answering method, question and answer system, computer equipment and storage medium |
CN110532358A (en) * | 2019-07-05 | 2019-12-03 | 东南大学 | A kind of template automatic generation method towards knowledge base question and answer |
Non-Patent Citations (2)
Title |
---|
JOHN DENERO: "The Complexity of Phrase Alignment Problems" * |
阮翀: "基于多译文的中文转述语料库建设及转述评价方案" * |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111897944A (en) * | 2020-08-17 | 2020-11-06 | 哈尔滨工业大学 | Knowledge map question-answering system based on semantic space sharing |
CN111897944B (en) * | 2020-08-17 | 2024-03-22 | 哈尔滨工业大学 | Knowledge graph question-answering system based on semantic space sharing |
CN112632237A (en) * | 2020-12-07 | 2021-04-09 | 厦门渊亭信息科技有限公司 | Knowledge graph-based question-answer template automatic generation method and device |
CN112765312A (en) * | 2020-12-31 | 2021-05-07 | 湖南大学 | Knowledge graph question-answering method and system based on graph neural network embedding matching |
CN113066358A (en) * | 2021-04-14 | 2021-07-02 | 吴光银 | Science teaching auxiliary system |
CN112989005B (en) * | 2021-04-16 | 2022-07-12 | 重庆中国三峡博物馆 | Knowledge graph common sense question-answering method and system based on staged query |
CN112989005A (en) * | 2021-04-16 | 2021-06-18 | 重庆中国三峡博物馆 | Knowledge graph common sense question-answering method and system based on staged query |
CN113947084A (en) * | 2021-09-24 | 2022-01-18 | 盛景智能科技(嘉兴)有限公司 | Question-answer knowledge retrieval method, device and equipment based on graph embedding |
CN114417807A (en) * | 2022-01-24 | 2022-04-29 | 中国电子科技集团公司第五十四研究所 | Human-like language description expression method oriented to presence or absence of human collaboration scene |
CN114417807B (en) * | 2022-01-24 | 2023-09-22 | 中国电子科技集团公司第五十四研究所 | Human-like language description expression method for collaboration scene of presence or absence |
CN114579710B (en) * | 2022-03-15 | 2023-04-25 | 西南交通大学 | Method for generating problem query template of high-speed train |
CN114579710A (en) * | 2022-03-15 | 2022-06-03 | 西南交通大学 | Method for generating problem query template of high-speed train |
CN117743652A (en) * | 2023-12-16 | 2024-03-22 | 华南理工大学 | Weak supervision time sequence map question-answering method based on deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN111339269B (en) | 2023-09-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111339269B (en) | Knowledge graph question-answering training and application service system capable of automatically generating templates | |
CN112989005B (en) | Knowledge graph common sense question-answering method and system based on staged query | |
CN110377715A (en) | Reasoning type accurate intelligent answering method based on legal knowledge map | |
CN111325029B (en) | Text similarity calculation method based on deep learning integrated model | |
CN109902159A (en) | A kind of intelligent O&M statement similarity matching process based on natural language processing | |
CN108932342A (en) | A kind of method of semantic matches, the learning method of model and server | |
CN109710744B (en) | Data matching method, device, equipment and storage medium | |
CN109857846B (en) | Method and device for matching user question and knowledge point | |
CN112328800A (en) | System and method for automatically generating programming specification question answers | |
EP3864565A1 (en) | Method of searching patent documents | |
CN112149427B (en) | Verb phrase implication map construction method and related equipment | |
CN113569023A (en) | Chinese medicine question-answering system and method based on knowledge graph | |
CN113196277A (en) | System for retrieving natural language documents | |
CN112417170B (en) | Relationship linking method for incomplete knowledge graph | |
CN110245238A (en) | The figure embedding grammar and system of Process Based and syntax schema | |
CN112632250A (en) | Question and answer method and system under multi-document scene | |
Ren et al. | Detecting the scope of negation and speculation in biomedical texts by using recursive neural network | |
CN113821605A (en) | Event extraction method | |
CN116402066A (en) | Attribute-level text emotion joint extraction method and system for multi-network feature fusion | |
CN113535897A (en) | Fine-grained emotion analysis method based on syntactic relation and opinion word distribution | |
CN116127099A (en) | Combined text enhanced table entity and type annotation method based on graph rolling network | |
CN111666374A (en) | Method for integrating additional knowledge information into deep language model | |
CN114372454B (en) | Text information extraction method, model training method, device and storage medium | |
CN113705207A (en) | Grammar error recognition method and device | |
CN116680407A (en) | Knowledge graph construction method and device |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |