CN111767376B - Question-answering system and method based on dynamic knowledge graph - Google Patents
Question-answering system and method based on dynamic knowledge graph Download PDFInfo
- Publication number
- CN111767376B CN111767376B CN202010495779.4A CN202010495779A CN111767376B CN 111767376 B CN111767376 B CN 111767376B CN 202010495779 A CN202010495779 A CN 202010495779A CN 111767376 B CN111767376 B CN 111767376B
- Authority
- CN
- China
- Prior art keywords
- knowledge graph
- entity
- module
- subspace
- updated
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000003058 natural language processing Methods 0.000 claims description 20
- 230000003993 interaction Effects 0.000 claims description 11
- 238000005295 random walk Methods 0.000 claims description 9
- 230000002457 bidirectional effect Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 6
- 150000001875 compounds Chemical class 0.000 claims description 4
- 238000013519 translation Methods 0.000 abstract description 6
- 230000002708 enhancing effect Effects 0.000 abstract description 2
- 230000003068 static effect Effects 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 239000013598 vector Substances 0.000 description 3
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000008520 organization Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/901—Indexing; Data structures therefor; Storage structures
- G06F16/9024—Graphs; Linked lists
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Human Computer Interaction (AREA)
- Bioinformatics & Computational Biology (AREA)
- Software Systems (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a question-answering system and method based on a dynamic knowledge graph, which are characterized in that a subspace is constructed by associating entity relations to fuse new entities in the subspace, so that entity updating in the knowledge graph does not depend on global configuration any more, and the entity updating can be respectively updated in different subspaces, thereby eliminating the need of the prior translation model for super-parameter adjustment, relieving congestion of entities and relations in the subspaces, simultaneously easily realizing parallelization of updating, enhancing the adaptability to dynamic data, particularly the new entities, and improving the precision and timeliness of the question-answering system.
Description
Technical Field
The invention belongs to the technical field of artificial intelligence, and particularly relates to a question-answering system and a question-answering method based on a dynamic knowledge graph.
Background
The knowledge graph is a large-scale semantic network, which is composed of concept entities and semantic relations, and describes various entities or concepts existing in the real world and relations thereof through node representation entities or concepts and edge representation relations, and is generally represented by triples, namely: head entity, relationship, and tail entity. As an important component of artificial intelligence technology, knowledge maps have been widely applied in intelligent search, man-machine question answering, personalized recommendation and other directions due to their strong interconnected organization, information retrieval and knowledge reasoning capabilities, and provide a technical basis for the intellectual organization and intelligent application in a plurality of fields such as medical treatment, finance and the like.
The knowledge graph is divided into a static knowledge graph and a dynamic knowledge graph. The static knowledge graph is a closed knowledge graph, and no new entity exists in the knowledge graph, and the existing entity cannot be updated greatly. The dynamic knowledge graph is an open knowledge graph, and new entities can be added into the knowledge graph and new relations can be generated.
The question-answering system is a high-level form of information retrieval system, and can answer questions posed by users in natural language with accurate and concise natural language. The main reason for the rise of the research is the demand of people for quickly and accurately acquiring information, and the question-answering system is a research direction which is concerned by people and has wide development prospect in the fields of artificial intelligence and natural language processing at present. The existing question-answering systems mainly include the following two types: the knowledge graph question-answering system based on the template matching technology adopts methods of natural language processing, entity matching, relation matching and the like to convert a question sentence of a user into a structured query sentence, then adopts the structured query sentence to search in a knowledge graph, and returns a final result; in addition, the knowledge map question-answering system based on naive Bayes classification firstly carries out word segmentation on natural language input by a user, and the system uses naive Bayes classification to carry out classification according to word segmentation processing results, thereby conjecturing the problem that the user wants to consult, and finally the system extracts answers from a database and displays the answers to the user.
In summary, the defects of the existing knowledge-graph question-answering system mainly include the following points: 1. the construction of the knowledge graph is static, the knowledge in the real society is constantly changed, namely the knowledge is time-efficient, and most of the existing knowledge graph data is static; 2. the knowledge-graph lacks adaptability to dynamic data and lacks support for incremental updates in the model, particularly for newly added entities.
Disclosure of Invention
In view of this, the invention provides a question-answering system and method based on a dynamic knowledge graph, which realize dynamic update of the knowledge graph and improve the precision and timeliness of the question-answering system.
The invention provides a question-answering method based on a dynamic knowledge graph, which adopts a natural language processing method to convert user questions into query sentences, uses the query sentences to search in the knowledge graph and returns corresponding answers, and the updating process of the knowledge graph comprises the following steps:
step 4, performing knowledge graph completion on the updated subspace to form an updated knowledge graph; and combining the updated knowledge graph and the original knowledge graph to form a new knowledge graph.
Further, the method further comprises a process of predicting according to historical answers, and specifically comprises the following steps:
forming a high-grade answer table according to the scores of the historical answers; and constructing a query statement according to the entity relationship contained in the high-score answer table, and searching and returning an answer in a knowledge graph by using the query statement.
Further, in the step 2, the similarity between entity relationships is calculated by using cosine similarity.
Further, the system comprises a user interaction module, a natural language processing module, a knowledge graph storage module and a knowledge graph updating module;
the user interaction module is used for sending the received user question to the natural language processing module;
the natural language processing module is used for converting the user question into a query statement and sending the query statement to the knowledge graph storage module; converting the updated file into a structured newly added entity, and sending the newly added entity to the knowledge graph updating module;
the knowledge graph updating module is used for clustering the newly added entities to form a core entity and determining an entity relationship with the similarity of the entity relationship with the core entity larger than a set threshold as an associated entity relationship; selecting an entity containing the associated entity relationship from the original knowledge graph as an associated entity; adopting a bidirectional random walk model, dividing to form subspaces by taking the associated entity relationship as a semantic focus and the associated entity as a starting point; fusing the newly added entity with the subspace to form an update subspace; performing knowledge graph completion on the updated subspace to form an updated knowledge graph; sending the updated knowledge graph to the knowledge graph storage module;
the knowledge graph storage module is used for storing a knowledge graph and returning a corresponding answer according to the query statement; and combining the updated knowledge-graph with the original knowledge-graph to form a new knowledge-graph.
Further, the user interaction module is also used for scoring the answers and forming a high-score answer table according to the scores.
Further, the system further comprises a prediction recommendation module, wherein the prediction recommendation module is used for constructing a compound query statement according to the entity relationship contained in the high-score answer table and sending the compound query statement to the knowledge graph storage module.
Has the advantages that:
according to the invention, the newly added entity is fused in the subspace through the associated entity relationship construction subspace, so that entity updating in the knowledge graph does not depend on global configuration any more, and can be respectively updated in different subspaces, thereby eliminating the need of the prior translation model for hyper-parameter adjustment, relieving congestion of the entity and the relationship in the subspace, simultaneously easily realizing updating parallelization, enhancing the adaptability to dynamic data, especially the newly added entity, and improving the precision and timeliness of the question-answering system.
Drawings
Fig. 1 is a flow chart of knowledge graph update of the question answering method based on dynamic knowledge graph according to the present invention.
Fig. 2 is a flow chart of a natural language processing procedure of the question answering method based on the dynamic knowledge graph provided by the invention.
Fig. 3 is a knowledge graph updating flow chart of the question-answering method based on the dynamic knowledge graph provided by the invention.
Fig. 4 is a schematic structural diagram of the question-answering system based on the dynamic knowledge graph provided by the invention.
FIG. 5 is a flow diagram of a knowledge graph update module of the dynamic knowledge graph-based question answering system provided by the present invention.
FIG. 6 is a schematic diagram of a knowledge-graph storage structure of a question-answering system based on a dynamic knowledge graph according to the present invention
Detailed Description
The invention is described in detail below by way of example with reference to the accompanying drawings.
The invention provides a question-answering method based on a dynamic knowledge graph, which has the core idea that: the method comprises the steps of converting a user question into an inquiry statement by adopting a natural language processing method, searching in a knowledge graph by using the inquiry statement and returning a corresponding answer, wherein the knowledge graph can be dynamically updated according to an external update file.
The process of knowledge graph updating in the question-answering method based on the dynamic knowledge graph, as shown in fig. 1, specifically comprises the following steps:
The method provided by the invention can be realized by adopting a knowledge representation mode based on a translation model (Trans series model). The Trans series model is one of the classic static knowledge map completion models, and represents an entity as a triple entity: (head, relation, tail), wherein head and tail represent attributes of the entity, and relation represents translation from the head of the entity to the tail of the entity, namely the relationship between the entities. On the basis, the entity in the invention is expressed as a triple (head, relation, tail), and the update file is converted into a structured newly-added entity triple through a natural language processing method. As shown in fig. 2, in the present invention, a natural language processing method is used to complete the conversion of user questions to form query statements, and meanwhile, a natural language processing method is used to complete the processing of update files to form structured entity data.
And 2, clustering the newly added entities to form a core entity, selecting an entity relationship with the similarity of the entity relationship with the core entity larger than a set threshold value as an associated entity relationship, and selecting an entity containing the associated entity relationship from the original knowledge graph as the associated entity.
And clustering the newly added entities by adopting a clustering method to form a plurality of clustering centers, wherein the clustering centers are the core entities.
The present invention can use cosine similarity to calculate the similarity between entity relationships, where cosine similarity is the cosine of the angle between two n-dimensional vectors in an n-dimensional space, which is the product of the dot product of the two vectors divided by the length (or amplitude) of the two vectors. Selecting entity relations with the similarity of the entity relations with the core entity larger than a set threshold value as associated entity relations, and defining all entities containing the associated entity relations as associated entities.
And in the original knowledge graph, selecting the associated entities to generate a set containing all the associated entities according to the calculated associated entity relationship.
The invention adopts a bidirectional Random Walk (RW) model, takes the associated entity in the selected entity set as a starting point and takes the associated entity relationship as a semantic focus to divide the entity set into subspaces, thereby realizing the mutual association between the newly added entity and the entity in the original knowledge graph. At this time, a situation that the size of the divided subspace may be too large may occur, and the size of the subspace may be controlled by setting a hyper-parameter.
On the basis, an entity linking method (entity linking) is adopted to fuse the triple sets corresponding to the newly added entities and the triple sets respectively corresponding to the subspaces generated by the bidirectional random walk, so that a complete updating subspace is formed.
Step 4, complementing the knowledge graph of the updated subspace to form an updated knowledge graph; and combining the updated knowledge graph and the original knowledge graph to form a new knowledge graph.
And (4) performing knowledge graph completion on the updated subspace by adopting a classical translation model to form an updated knowledge graph. Since the update file is divided into a plurality of subspaces, it is not necessary to rely on the global configuration of the algorithm, and only different values of the hyper-parameters (margin and learning rate) need to be set locally for each subspace, which greatly simplifies the operation cost of the method. And merging the updated knowledge graph into the original knowledge graph to form the knowledge graph containing the newly added entity.
In addition, the question-answering method based on the dynamic knowledge graph provided by the invention can also predict according to historical answers, as shown in fig. 3, firstly, a high-score answer table is established according to scores of historical answers by users, then, query sentences are constructed according to entity relations contained in the high-score answer table, and the generated query sentences are used for searching in the knowledge graph and returning answers. The answer prediction is carried out based on the high-score answer sheet, the intelligence of the question-answering method is improved, the answer which the user wants to know can be presumed according to the query habit of the user, meanwhile, the question which the user wants to ask can be prejudged, and the answer which the user possibly wants is actively provided for the user.
The invention provides a question-answering system based on a dynamic knowledge graph, which comprises a user interaction module, a natural language processing module, a prediction recommendation module, a knowledge graph storage module and a knowledge graph updating module, as shown in figure 4.
And the user interaction module is used for receiving the user question, transmitting the user question to the natural language processing module, displaying the query result (answer) returned by the knowledge graph storage module, grading the query result by the user, and storing the recent high-score evaluation entity record into a high-score cache table, wherein the cache table adopts a first-in first-out (FIFO) strategy. And after the cache table is updated, the user interaction module sends the cache table to the prediction recommendation module and displays the prediction recommendation result returned by the prediction recommendation module.
And the natural language processing module is used for receiving the user questions and analyzing the user questions in the natural language form. Firstly, performing entity attribute identification and entity relationship extraction on a problem based on a bi-LSTM neural network model; and then, completing subsequent semantic matching according to the training model to generate a Cypher query statement, and sending the Cypher query statement to the knowledge spectrogram storage module for query. In the dynamic updating process of the knowledge base, the natural language processing module receives an external updating file, performs entity identification on data, refines entity attributes, extracts entity relationships and transmits the refined structured data (comprising an entity attribute list and an entity relationship list) to the knowledge graph updating module.
And the prediction recommendation module is used for extracting a plurality of entities with the highest occurrence frequency according to the high-resolution cache list transmitted by the user interaction module, inquiring N most related entities from the knowledge map repository, replacing the entities in the recommended cache list by the N entities by adopting a time-sensing least-recently-used (TLRU) replacement strategy, and returning the module to the user interaction module and displaying the module to the user as long as the data of the recommended cache list is updated.
The invention can be realized by adopting a Neo4j knowledge base, wherein Neo4j is a high-performance NOSQL graph database and stores structured data on a network instead of a table, and Neo4j can also be regarded as a high-performance graph engine which has all the characteristics of a mature database. The invention uses Cypher as the query language of the graph database, the storage mode of the invention is that different types of entities and attributes thereof are separately stored in one csv file, the entity relationship is in the other csv file, and the specific storage structure is shown in FIG. 6.
And the knowledge graph updating module is used for updating the knowledge graph stored in the knowledge graph storage module according to the update file, and the flow of the knowledge graph updating module is shown in fig. 5. Firstly, a knowledge graph updating module receives entity relations transmitted from a natural language processing module, clusters the entity relations, forms n clustering centers as shown in the figure, divides parallel subspaces for each clustering center in a random walk mode, and completes the knowledge graph in the parallel subspaces by using a translation model to form an updated knowledge graph; and finally, the updated knowledge graph is merged into the original knowledge graph to form the knowledge graph containing the new entity.
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. A question-answering method based on a dynamic knowledge graph adopts a natural language processing method to convert user questions into query sentences, and the query sentences are used for searching in the knowledge graph and returning corresponding answers, and is characterized in that the updating process of the knowledge graph comprises the following steps:
step 1, converting an update file into a structured newly added entity, wherein the newly added entity comprises entity attributes and entity relationships;
step 2, clustering the newly added entities to form a core entity; selecting the entity relation with the similarity of the entity relation with the core entity larger than a set threshold value as an associated entity relation; selecting an entity containing the associated entity relationship from the original knowledge graph as an associated entity;
step 3, adopting a bidirectional random walk model, taking the associated entity relationship as a semantic focus, and taking the associated entity as a starting point to divide and form a subspace; fusing the newly added entity with the subspace to form an update subspace;
step 4, complementing the knowledge graph of the updated subspace to form an updated knowledge graph; and combining the updated knowledge graph and the original knowledge graph to form a new knowledge graph.
2. The method according to claim 1, further comprising a process of predicting based on historical answers, including the steps of:
forming a high-grade answer table according to the scores of the historical answers; and constructing a query statement according to the entity relationship contained in the high-score answer table, and searching and returning an answer in the knowledge graph by using the query statement.
3. The method according to claim 1, wherein the similarity between entity relationships is calculated in step 2 by using cosine similarity.
4. A question-answering system based on a dynamic knowledge graph is characterized by comprising a user interaction module, a natural language processing module, a knowledge graph storage module and a knowledge graph updating module;
the user interaction module is used for sending the received user question to the natural language processing module;
the natural language processing module is used for converting the user question into a query statement and sending the query statement to the knowledge graph storage module; converting the updated file into a structured newly added entity, and sending the newly added entity to the knowledge graph updating module;
the knowledge graph updating module is used for clustering the newly added entities to form a core entity and determining an entity relationship with the similarity of the entity relationship with the core entity larger than a set threshold value as an associated entity relationship; selecting an entity containing the relationship of the associated entities from the original knowledge graph as the associated entities; adopting a bidirectional random walk model, taking the associated entity relationship as a semantic focus, and taking the associated entity as a starting point to divide and form a subspace; fusing the newly added entity with the subspace to form an update subspace; performing knowledge graph completion on the updated subspace to form an updated knowledge graph; sending the updated knowledge graph to the knowledge graph storage module;
the knowledge graph storage module is used for storing a knowledge graph and returning a corresponding answer according to the query statement; and combining the updated knowledge graph with the original knowledge graph to form a new knowledge graph.
5. The system of claim 4, wherein the user interaction module is further configured to score the answers and form a high-score answer sheet according to the scores.
6. The system according to claim 5, further comprising a prediction recommendation module, wherein the prediction recommendation module is configured to construct a compound query statement according to an entity relationship included in the high-ranking answer table, and send the compound query statement to the knowledge graph storage module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010495779.4A CN111767376B (en) | 2020-06-03 | 2020-06-03 | Question-answering system and method based on dynamic knowledge graph |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010495779.4A CN111767376B (en) | 2020-06-03 | 2020-06-03 | Question-answering system and method based on dynamic knowledge graph |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111767376A CN111767376A (en) | 2020-10-13 |
CN111767376B true CN111767376B (en) | 2023-04-07 |
Family
ID=72719963
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010495779.4A Expired - Fee Related CN111767376B (en) | 2020-06-03 | 2020-06-03 | Question-answering system and method based on dynamic knowledge graph |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111767376B (en) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112966084B (en) * | 2021-03-11 | 2022-04-05 | 北京三快在线科技有限公司 | Knowledge graph-based answer query method, device, equipment and storage medium |
CN112905806B (en) * | 2021-03-25 | 2022-11-01 | 哈尔滨工业大学 | Knowledge graph materialized view generator based on reinforcement learning and generation method |
CN113760996A (en) * | 2021-09-09 | 2021-12-07 | 上海明略人工智能(集团)有限公司 | Data integration method, system, equipment and storage medium |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2019103727A1 (en) * | 2017-11-21 | 2019-05-31 | Google Llc | Improved onboarding of entity data |
CN110750649A (en) * | 2018-07-06 | 2020-02-04 | 中兴通讯股份有限公司 | Knowledge graph construction and intelligent response method, device, equipment and storage medium |
CN109308321A (en) * | 2018-11-27 | 2019-02-05 | 烟台中科网络技术研究所 | A kind of knowledge question answering method, knowledge Q-A system and computer readable storage medium |
CN110059196A (en) * | 2019-04-12 | 2019-07-26 | 张晓红 | A kind of Relation extraction method and system of medical health domain knowledge map |
CN110275966B (en) * | 2019-07-01 | 2021-10-01 | 科大讯飞(苏州)科技有限公司 | Knowledge extraction method and device |
CN110532399A (en) * | 2019-08-07 | 2019-12-03 | 广州多益网络股份有限公司 | Knowledge mapping update method, system and the device of object game question answering system |
CN110851614A (en) * | 2019-09-09 | 2020-02-28 | 中国电子科技集团公司电子科学研究院 | Relation prediction deduction method of knowledge graph and dynamic updating method of knowledge graph |
CN111209410B (en) * | 2019-12-27 | 2023-04-18 | 中国地质大学(武汉) | Anchor point-based dynamic knowledge graph representation learning method and system |
-
2020
- 2020-06-03 CN CN202010495779.4A patent/CN111767376B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN111767376A (en) | 2020-10-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111753060B (en) | Information retrieval method, apparatus, device and computer readable storage medium | |
CN108563653B (en) | Method and system for constructing knowledge acquisition model in knowledge graph | |
CN111767376B (en) | Question-answering system and method based on dynamic knowledge graph | |
WO2021093755A1 (en) | Matching method and apparatus for questions, and reply method and apparatus for questions | |
KR102054514B1 (en) | The System and the method of offering the Optimized answers to legal experts utilizing a Deep learning training module and a Prioritization framework module based on Artificial intelligence and providing an Online legal dictionary utilizing a character Strings Dictionary Module that converts legal information into significant vector | |
CN111475623A (en) | Case information semantic retrieval method and device based on knowledge graph | |
CN111950285A (en) | Intelligent automatic construction system and method of medical knowledge map based on multi-modal data fusion | |
CN102663129A (en) | Medical field deep question and answer method and medical retrieval system | |
CN113673943B (en) | Personnel exemption aided decision making method and system based on historical big data | |
CN117290489A (en) | Method and system for quickly constructing industry question-answer knowledge base | |
CN114238653B (en) | Method for constructing programming education knowledge graph, completing and intelligently asking and answering | |
CN115599899B (en) | Intelligent question-answering method, system, equipment and medium based on aircraft knowledge graph | |
CN112597285A (en) | Man-machine interaction method and system based on knowledge graph | |
CN115269899A (en) | Remote sensing image overall planning system based on remote sensing knowledge map | |
CN116340530A (en) | Intelligent design method based on mechanical knowledge graph | |
CN117909466A (en) | Domain question-answering system, construction method, electronic device and storage medium | |
CN115033706A (en) | Method for automatically complementing and updating knowledge graph | |
CN118132719A (en) | Intelligent dialogue method and system based on natural language processing | |
CN116991976A (en) | Model training method, device, electronic equipment and readable storage medium | |
JP2024006944A (en) | Semantic retrieval model training method, apparatus, electronic device, and storage medium | |
CN116049376A (en) | Method, device and system for retrieving and replying information and creating knowledge | |
CN113779981A (en) | Recommendation method and device based on pointer network and knowledge graph | |
CN117235237B (en) | Text generation method and related device | |
RU2707917C1 (en) | Method of searching for methods of resolving technical contradictions and a system based on a trained neural network for its implementation | |
CN112883281B (en) | User clustering search system based on deep learning |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20230407 |