CN110704600B - Question-answer dynamic matching method and device and electronic equipment - Google Patents
Question-answer dynamic matching method and device and electronic equipment Download PDFInfo
- Publication number
- CN110704600B CN110704600B CN201910942942.4A CN201910942942A CN110704600B CN 110704600 B CN110704600 B CN 110704600B CN 201910942942 A CN201910942942 A CN 201910942942A CN 110704600 B CN110704600 B CN 110704600B
- Authority
- CN
- China
- Prior art keywords
- entity
- knowledge graph
- vector
- vectors
- entity vector
- 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
- 238000000034 method Methods 0.000 title claims abstract description 63
- 239000013598 vector Substances 0.000 claims abstract description 295
- 230000015654 memory Effects 0.000 claims description 25
- 238000012545 processing Methods 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 14
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 13
- 238000004590 computer program Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000001953 sensory effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides a question-answer dynamic matching method, a question-answer dynamic matching device and electronic equipment, and relates to the field of intelligent search, wherein the method comprises the following steps: acquiring a query statement and a candidate paragraph; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraphs according to the matching degree. The method solves the technical problem that the matching degree between the question and the answer paragraph cannot be accurately calculated in the prior art, so that the determined target paragraph is not accurate enough, improves the accuracy of the target paragraph and meets the use requirements of users by dynamically updating and representing entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and an apparatus for dynamically matching questions and answers in the field of intelligent search, and an electronic device.
Background
Generally, the relevant question-answering tasks such as text similarity, question-answer matching, answer extraction and the like can be processed through a pre-trained model. In the related art, the matching degree of the question and the answer is obtained by directly introducing a pre-trained knowledge vector into a matching model, however, the pre-trained knowledge vector is a knowledge vector trained based on a generic knowledge graph, so that the matching degree of the obtained question and the answer is not consistent with the actual situation, and the accuracy of the target answer is finally influenced.
Disclosure of Invention
The present application is directed to solving, at least to some extent, one of the technical problems in the related art described above.
Therefore, a first objective of the present application is to provide a question-answer dynamic matching method, which solves the technical problem that the matching degree between a question and an answer paragraph cannot be accurately calculated in the prior art, so that a determined target paragraph is not accurate enough, and improves the accuracy of the target paragraph and meets the user requirement by dynamically updating and representing entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors.
A second objective of the present application is to provide a device for dynamically matching questions and answers.
A third object of the present application is to propose a computer device.
A fourth object of the present application is to propose a non-transitory computer-readable storage medium.
To achieve the above object, an embodiment of a first aspect of the present application provides a method for dynamically matching questions and answers, including: acquiring a query statement and a candidate paragraph; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree.
In addition, the question-answer dynamic matching method of the embodiment of the application also has the following additional technical characteristics:
optionally, the constructing a knowledge graph according to the plurality of entity vectors includes: acquiring entity interrelationships among the entity vectors; acquiring attribute information among the entity vectors; and connecting the entity vectors according to the entity correlation and the attribute information to generate the knowledge graph.
Optionally, the updating each entity vector in the knowledge graph to generate an entity vector to be matched includes: s1, determining a first entity vector to be updated from the plurality of entity vectors, and obtaining an adjacent entity list of the first entity vector to be updated; s2, extracting a first adjacent entity vector from the adjacent entity list, and acquiring a local knowledge graph corresponding to the first adjacent entity vector from the knowledge graph; s3, calculating correlation score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similarity score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain a first entity vector to be matched; s4, taking the first entity vector to be matched as the second entity vector to be updated, and repeatedly executing steps S2 and S3 until each adjacent entity vector in the adjacent entity list is taken out to generate an entity vector to be matched; and S5, repeatedly executing the step S1 to the step S4, and acquiring the entity vector to be matched in the knowledge graph.
Optionally, the obtaining a local knowledge-graph corresponding to the first neighboring entity vector from the knowledge-graph includes: and acquiring the local knowledge graph except the first entity vector to be updated in the knowledge graph according to a preset length.
Optionally, the method further includes: according to a preset scoring function, scoring the correlation among the entity vectors in the knowledge graph; and deleting part of entity vectors in the knowledge graph according to the scoring result and a preset threshold value.
To achieve the above object, a second aspect of the present application provides a dynamic question-answer matching device, including: the acquisition module is used for acquiring the query statement and the candidate paragraphs; the identification module is used for carrying out entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors; a construction module for constructing a knowledge graph according to the plurality of entity vectors; the updating module is used for updating each entity vector in the knowledge graph to generate an entity vector to be matched; and the processing module is used for matching the entity vector to be matched to obtain the matching degree of the query statement and the candidate paragraph, and determining a target paragraph according to the matching degree.
In addition, the question answering dynamic matching device of the embodiment of the application also has the following additional technical characteristics:
optionally, the building module includes: a first obtaining unit, configured to obtain entity interrelationships among the entity vectors; a second obtaining unit configured to obtain attribute information between the plurality of entity vectors; and the generating unit is used for connecting the entity vectors according to the entity correlation and the attribute information to generate the knowledge graph.
Optionally, the update module includes: a determining unit, configured to determine a first entity vector to be updated from the multiple entity vectors, and obtain an adjacent entity list of the first entity vector to be updated; a first obtaining unit, configured to take out a first adjacent entity vector from the adjacent entity list, and obtain a local knowledge graph corresponding to the first adjacent entity vector from the knowledge graph; the calculation unit is used for calculating correlation score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similarity score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain a first entity vector to be matched; the updating unit is used for taking the first entity vector to be matched as a second entity vector to be updated, and repeatedly executing the acquiring unit and the calculating unit until each adjacent entity vector in the adjacent entity list is taken out to generate an entity vector to be matched; and the second acquisition unit is used for repeatedly executing the determination unit to the updating unit and acquiring the entity vector to be matched in the knowledge graph.
Optionally, the first obtaining unit is specifically configured to: and acquiring the local knowledge graph except the first entity vector to be updated in the knowledge graph according to a preset length.
Optionally, the apparatus further includes: the scoring unit is used for scoring the correlation among the entity vectors in the knowledge graph according to a preset scoring function; and the deleting unit is used for deleting part of entity vectors in the knowledge graph according to the scoring result and a preset threshold value.
To achieve the above object, a third aspect of the present application provides a computer device, including: a processor and a memory; the processor reads the executable program code stored in the memory to run a program corresponding to the executable program code, so as to implement the question-answer dynamic matching method according to the embodiment of the first aspect.
To achieve the above object, a non-transitory computer-readable storage medium is provided in a fourth embodiment of the present application, and a computer program is stored thereon, where the computer program is executed by a processor to implement the dynamic question-answer matching method according to the first embodiment.
To achieve the above object, an embodiment of a fifth aspect of the present application provides a computer program product, where instructions of the computer program product, when executed by a processor, implement the method for dynamically matching questions and answers as described in the embodiment of the first aspect.
The technical scheme provided by the embodiment of the application can have the following beneficial effects:
acquiring a query statement and a candidate paragraph; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree. The method solves the technical problem that the matching degree between the question and the answer paragraph cannot be accurately calculated in the prior art, so that the determined target paragraph is not accurate enough, improves the accuracy of the target paragraph by dynamically updating and representing the entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors, and meets the use requirements of users.
Additional aspects and advantages of the present application will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the present application.
Drawings
The foregoing and/or additional aspects and advantages of the present application will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow diagram of a method for dynamic question-answer matching according to one embodiment of the present application;
FIG. 2 is a flow diagram of a method for dynamic question-answer matching according to another embodiment of the present application;
FIG. 3 is an exemplary diagram of constructing a knowledge-graph according to one embodiment of the present application;
FIG. 4 is a flow diagram of a method for dynamic question-answer matching according to yet another embodiment of the present application;
5 a-5 b are exemplary diagrams of dynamic update of building knowledge-graph entity vectors according to one embodiment of the present application;
FIG. 6 is an exemplary diagram of a question-answer matching model according to one embodiment of the present application;
FIG. 7 is a schematic diagram of a dynamic question-answer matching device according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of a dynamic question-answer matching device according to another embodiment of the present application;
FIG. 9 is a schematic structural diagram of a dynamic question-answer matching device according to yet another embodiment of the present application;
fig. 10 is a schematic structural diagram of a dynamic question-answer matching device according to yet another embodiment of the present application;
FIG. 11 is a schematic structural diagram of an electronic device according to one embodiment of the present application.
Detailed Description
Reference will now be made in detail to embodiments of the present application, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are exemplary and intended to be used for explaining the present application and should not be construed as limiting the present application.
The method, the apparatus, and the electronic device for question-answer dynamic matching according to the embodiments of the present application are described below with reference to the accompanying drawings.
Aiming at the technical problems mentioned in the background art, the problem and answer matching is carried out by directly introducing a pre-trained knowledge vector into a matching model in the prior art, the method is static, and in fact, the knowledge vector trained based on a knowledge graph spectrum is easily influenced by information of different dimensions of the whole graph, so that the matching degree of the obtained problem and answer is inconsistent with the actual situation, and the accuracy of the target answer is finally influenced.
Aiming at the problems, the application provides a method for dynamically matching questions and answers, and query sentences and candidate paragraphs are obtained; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree. By dynamically updating and representing the entity vectors in the question and answer paragraphs and calculating the matching degree based on the updated entity vectors, the accuracy of the target paragraph is improved, and the use requirements of users are met.
Specifically, fig. 1 is a flowchart of a dynamic question-answer matching method according to an embodiment of the present application, and as shown in fig. 1, the method includes:
Specifically, a user may input a query statement according to actual application needs, and after receiving the query statement, it may be understood that there are many candidate paragraphs and the matching degree between the query statement and the candidate paragraphs needs to be calculated.
Generally, a plurality of entity vectors in a query statement and a candidate paragraph are identified, then the plurality of entity vectors are introduced into a matching model, which is static, so that the calculation of the matching degree between the query statement and the candidate paragraph is not accurate enough.
It should be noted that the larger the knowledge graph applied, the more the knowledge is expressed, and the entity vector representation obtained by the current representation method is easily affected by the general graph, for example, "liu de hua" is both an actor and a singer, and the entity vector representation obtained by the general graph has information of the actor and the singer, but when the method is applied, if a task in the movie and television field is faced, the general graph representation effect is not as good as the domain graph, but for different field tasks, the cost for training knowledge representation alone is too large, and therefore the entity vector obtained by the general graph still needs to be matched.
Specifically, entity recognition can be performed on the query statement and the candidate paragraph through a preset entity recognition model (such as TransE, TransD, etc.) to obtain a plurality of entity vectors.
Further, a knowledge graph is constructed according to the entity vectors, as a possible implementation manner, the entity interrelation among the entity vectors is obtained, the attribute information among the entity vectors is obtained, and the knowledge graph is generated by connecting the entity vectors according to the entity interrelation and the attribute information.
It can be understood that the knowledge graph of the present application is constructed based on query sentences and candidate paragraphs, and is a sub-graph relative to a general knowledge graph constructed based on internet data such as encyclopedia.
And 103, updating each entity vector in the knowledge graph to generate an entity vector to be matched.
Specifically, each entity vector in the knowledge graph may be updated through a preset formula to generate an entity vector to be matched, for example, a plurality of entity vectors in the knowledge graph are A, B and C, respectively, and after the processing, entity vectors a1, B1, and C1 to be matched may be obtained, that is, a is a vector matrix of n × n, and after the updating, the vector matrix is also an n × n vector matrix, but a numerical value in the matrix changes.
As a possible implementation manner, for example, a representation X of a current entity vector to be updated and a list { Y1, Y2, Y3 … Yn } of adjacent entities are obtained, a local knowledge graph of the entity vectors Y1 and Y1 in the knowledge graph except X is cyclically extracted, a preset formula is applied to X for each entity vector of the Y1 knowledge graph to calculate, the obtained results are summed to obtain an updated X entity vector representation X1, the above steps are repeated until the list of adjacent entities is completely extracted, and a final updated X entity vector representation Xn after knowledge graph tuning is obtained as an entity vector to be matched corresponding to X, so that an entity vector to be matched can be obtained by updating each entity vector in the knowledge graph in the above manner.
And 104, inputting the entity vector to be matched into a preset matching model for processing to obtain the matching degree of the query statement and the candidate paragraph, and determining the target paragraph according to the matching degree.
Specifically, the entity vector to be matched, which is the updated entity vector, is input to a preset matching model to be processed to obtain the matching degree between the query statement and the candidate paragraph, and more specifically, the entity vector to be matched corresponding to the query statement and the entity vector to be matched corresponding to the candidate paragraph are respectively input to the preset matching model to be processed to the matching degree.
It can be understood that there are how many candidate paragraphs have how many matching degrees, and the candidate paragraph with the highest matching degree may be selected as the target paragraph according to the need, or the first candidate paragraphs with higher ranking may be selected as the target paragraph for ranking and displaying, and the like, and the display may be adjusted according to the need of the user.
It should be noted that a cosine or bidirectional lstm or other relevant matching model may be used as the preset matching model.
In summary, the question-answer dynamic matching method of the embodiment of the present application obtains a query sentence and a candidate paragraph; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph through a preset formula to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree. The method solves the technical problem that the matching degree between the question and the answer paragraph cannot be accurately calculated in the prior art, so that the determined target paragraph is not accurate enough, improves the accuracy of the target paragraph by dynamically updating and representing the entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors, and meets the use requirements of users.
Based on the above description of the embodiments, it can be understood that there are many ways to construct a knowledge graph from multiple entity vectors, and the following description is specifically made with reference to fig. 2:
fig. 2 is a flowchart of a method for dynamically matching questions and answers according to another embodiment of the present application, and as shown in fig. 2, step 102 includes:
And step 203, connecting the entity vectors according to the entity interrelation and the attribute information to generate a knowledge graph.
Specifically, if the entity interrelations among the multiple entity vectors are entities that appear together, for example, if the entity interrelations are entities that appear together, the entities that appear together may be edge-connected based on a sliding window, and more specifically, the entity vectors that appear together near a term window of the entity vectors after coreference resolution (entity coreference resolution is different identifiers that identify the same entity in different associated data sources) may be edge-connected.
In one embodiment of the application, multi-hop edge connection processing between the entity vector and other entity vectors in the article (considering the relationship between entity vectors and distant context entities) can also be added.
For example, as shown in fig. 3, the query statement Question is "the first movie of liu de hua evolution", and the candidate Passage is "1981, the movie of liu de hua evolution" was "colorful cloud song", a student in a music class was played in a film, in 1982, the ancestor name of the army translation officer was played in the drama film "ran rage sea", and the best new actor nomination of the 2 nd hong Kong movie golden image prize was obtained by means of the film. After the entity recognition is carried out, a plurality of entity vectors of Liu De Hua, movie, in 1981, the Cai Yun Qu and the Ping Peng Ben Hai can be obtained, and attribute information between the entity vectors, such as the attribute between Liu De Hua and the Cai Yun Qu, is the evolution through the analysis of the entity interrelations, so that the connection processing can be carried out between the two entity vectors, and the knowledge graph in the window of FIG. 3 can be obtained.
Specifically, edges of attributes of the entity vectors and the like can be added, as a possible implementation manner, in order to further improve the relevance of the entity vector relationship in the knowledge graph, the relevance between the entity vectors in the knowledge graph is scored according to a preset scoring function, and part of the entity vectors in the knowledge graph are deleted according to a scoring result and a preset threshold, that is, some edges are discarded by using some scoring manners, for example, Top is taken according to the relevance, the scale of the knowledge graph is reduced, and the calculation efficiency is improved.
Based on the description of the embodiment of the first aspect, it can be known that there are many ways to update each entity vector in the knowledge graph through a preset formula to generate an entity vector to be matched, which is described in detail below with reference to fig. 4.
Fig. 4 is a flowchart of a question-answer dynamic matching method according to another embodiment of the present application, as shown in fig. 4, the method includes:
s1, determining a first entity vector to be updated from the plurality of entity vectors, and obtaining a neighboring entity list of the first entity vector to be updated.
Specifically, each entity vector of the plurality of entity vectors may be used as a first entity vector to be updated for update, and each entity vector of the plurality of entity vectors may be sequentially updated or each entity vector of the plurality of entity vectors may be randomly selected for update until all entity vectors are updated. The definition of the adjacent entity may be selected according to needs, for example, the entity associated with the edge with the defined length of 1 is the adjacent entity.
Further, after the first entity vector to be updated is determined, a neighboring entity list of the first entity vector to be updated, such as the entity vector liudeb in fig. 3, is obtained, where the neighboring entity list is "unworkable sea of projection" and "colorful cloud song".
S2, the first adjacent entity vector is taken out from the adjacent entity list, and the local knowledge graph corresponding to the first adjacent entity vector is obtained from the knowledge graph.
And S3, calculating correlation score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similarity score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain the first entity vector to be matched.
Specifically, adjacent entity vectors are sequentially taken out from an adjacent entity list to perform calculation updating, a first adjacent entity vector is taken out from the adjacent entity list, and a local knowledge graph corresponding to the first adjacent entity vector is obtained from the knowledge graph, wherein the local knowledge graph can be adjusted according to actual application requirements, as a possible implementation manner, the local knowledge graph excluding the first entity vector to be updated in the knowledge graph is obtained according to a preset length, wherein the preset length can be selected according to requirements, for example, each entity vector is only based on the local knowledge graph with the surrounding length of 1.
And further, calculating related scores between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similar score as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain the first entity vector to be matched. Wherein the entity vector correlation score value may be a cosine calculated between two entity vectors. Such as the formula:
Thus, a knowledge-graph is constructed based on all entity vectors of the current context (query statement + candidate paragraph), dynamically updated based on each entity vector in the knowledge-graph (context knowledge-graph is one, but the surrounding entities for updating each entity vector may be different (or the entire knowledge-graph may also be shared), e.g., each entity vector updates the representation based only on the surrounding length-1 entity-vector knowledge-graph, i.e., the local knowledge-graph for updating is a subset (or the full set) of the context knowledge-graph.
And S4, taking the first entity vector to be matched as the second entity vector to be updated, and repeatedly executing the step S2 and the step S3 until each adjacent entity vector in the adjacent entity list is taken out to generate the entity vector to be matched.
And S5, repeatedly executing the step S1 to the step S4, and acquiring the entity vector to be matched in the knowledge graph.
Specifically, the steps are repeated until the adjacent entity list is completely taken out, and the finally updated entity vector is obtained and serves as the entity vector to be matched, so that the entity vector to be matched can be obtained by updating each entity vector in the knowledge graph in the manner.
For the sake of clarity of the above process, the detailed description is made with reference to specific examples, as shown in fig. 5a and 5b, one query statement, two candidate paragraphs passage are taken as examples; the similarity of the single-view text, passB is more relevant to query statements, and Passage A does not even mention the United states. Thus, the present application establishes a relationship between "usa" and "white house", "iraq", "weapons of mass destruction", "military operations" and "iraq war", constructs a knowledge graph (grey part) based on context, updates the static representation to make "usa" and "white house" more closely related in the current context, if represented by a two-dimensional graph, it can be seen that on the original static representation, white house and usa are related, but because of the representation of knowledge graph training, the space vector of usa is affected by other entities (such as harvard university, liberty statue, etc.), and is not very close in presentation; however, after the knowledge graph based on the current context is updated, the spatial distance between the white palace and the United states is closer, the corresponding relation is tighter, a tighter representation is introduced into the matching model, and the calculated matching degree score is higher.
Specifically, for example, in fig. 6B, the entity vectors obtained for the query statement and the candidate paragraphs may be directly input to C for calculation, but are statically introduced, and the calculation result of the matching degree is poor, so that the calculation of C input after dynamically updating each entity vector provided in the embodiment of the present application may be performed through the knowledge graph in the knowledge graph of the part a, thereby improving the accuracy of the calculation of the matching degree. It should be noted that the matching degree score between the query statement and the candidate paragraph may be output according to the task target, and may also be applied to the candidate paragraph extraction task, the answer start position location, and the like. Therefore, the relation between the entities in the current context is deeply modeled by locally constructing the knowledge graph, a method for dynamically updating entity representation based on the local knowledge graph is provided for the general graph, the gap between the general graph and the domain graph is overcome, and the method has wider application value.
Therefore, each entity vector in the knowledge graph is updated through a preset formula to generate an entity vector to be matched, a preset matching model is introduced, the calculated matching degree score is higher, the accuracy of the target paragraph is improved, and the use requirements of users are met.
In order to implement the above embodiments, the present application further provides a device for dynamically matching questions and answers. Fig. 7 is a schematic structural diagram of a dynamic question-answer matching device according to an embodiment of the present application, and as shown in fig. 7, the dynamic question-answer matching device includes: an acquisition module 701, an identification module 702, a construction module 703, an update module 704 and a processing module 705, wherein,
an obtaining module 701, configured to obtain a query statement and a candidate paragraph.
An identifying module 702, configured to perform entity identification on the query statement and the candidate paragraph to obtain a plurality of entity vectors.
A building module 703, configured to build a knowledge graph according to the plurality of entity vectors.
An updating module 704, configured to perform an updating process on each entity vector in the knowledge graph to generate an entity vector to be matched.
The processing module 705 is configured to input the entity vector to be matched to a preset matching model for matching processing, so as to obtain a matching degree between the query statement and the candidate paragraph, and determine a target paragraph according to the matching degree.
In an embodiment of the present application, as shown in fig. 8, on the basis of fig. 7, a module 703 is constructed, including: a first acquiring unit 7031, a second acquiring unit 7032, and a generating unit 7033.
A first obtaining unit 7031, configured to obtain an entity correlation between the multiple entity vectors.
A second obtaining unit 7032, configured to obtain attribute information between the multiple entity vectors.
A generating unit 7033, configured to perform connection processing on the multiple entity vectors according to the entity interrelations and the attribute information to generate the knowledge graph.
In one embodiment of the present application, as shown in fig. 9, based on the method shown in fig. 7, the updating module 704 includes: determining unit 7041, first obtaining unit 7042, calculating unit 7043, updating unit 7044, and second obtaining unit 7045.
A determining unit 7041 is configured to determine a first entity vector to be updated from the multiple entity vectors, and obtain an adjacent entity list of the first entity vector to be updated.
A first obtaining unit 7042, configured to take out a first neighboring entity vector from the neighboring entity list, and obtain a local knowledge graph corresponding to the first neighboring entity vector from the knowledge graph.
A calculating unit 7043, configured to calculate related score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and perform weighted summation on each entity vector in the local knowledge graph by using each similar score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain a first entity vector to be matched.
An updating unit 7044, configured to take the first entity vector to be matched as a second entity vector to be updated, and repeatedly execute the obtaining unit and the calculating unit until each adjacent entity vector in the adjacent entity list is taken out to generate an entity vector to be matched.
A second obtaining unit 7045, configured to repeatedly execute the determining unit to the updating unit, and obtain an entity vector to be matched in the knowledge graph.
In an embodiment of the present application, first obtaining unit 7042 is specifically configured to: and acquiring the local knowledge graph except the first entity vector to be updated in the knowledge graph according to a preset length.
In an embodiment of the present application, as shown in fig. 10, on the basis of fig. 8, the method further includes: a scoring element 7034 and a deletion element 7035.
A scoring unit 7034, configured to score correlations between entity vectors in the knowledge graph according to a preset scoring function.
A deleting unit 7035, configured to delete part of the entity vectors in the knowledge graph according to the scoring result and a preset threshold.
It should be noted that the explanation of the above-mentioned embodiment of the dynamic question-answer matching method is also applicable to the dynamic question-answer matching device of this embodiment, and is not described herein again.
In summary, the question-answer dynamic matching device of the embodiment of the present application obtains the query sentence and the candidate paragraph; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree. The method solves the technical problem that the matching degree between the question and the answer paragraph cannot be accurately calculated in the prior art, so that the determined target paragraph is not accurate enough, improves the accuracy of the target paragraph by dynamically updating and representing the entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors, and meets the use requirements of users.
According to an embodiment of the present application, an electronic device and a readable storage medium are also provided.
Fig. 11 is a block diagram of an electronic device according to an embodiment of the present application. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be examples only, and are not meant to limit implementations of the present application that are described and/or claimed herein.
As shown in fig. 11, the electronic apparatus includes: one or more processors 1101, a memory 1102, and interfaces for connecting the various components, including a high speed interface and a low speed interface. The various components are interconnected using different buses and may be mounted on a common motherboard or in other manners as desired. The processor may process instructions for execution within the electronic device, including instructions stored in or on the memory to display graphical information of a GUI on an external input/output apparatus (such as a display device coupled to the interface). In other embodiments, multiple processors and/or multiple buses may be used, along with multiple memories and multiple memories, as desired. Also, multiple electronic devices may be connected, with each device providing portions of the necessary operations (e.g., as a server array, a group of blade servers, or a multi-processor system). In fig. 11, a processor 1101 is taken as an example.
The memory 1102 is a non-transitory computer readable storage medium as provided herein. The memory stores instructions executable by at least one processor to cause the at least one processor to perform the method for dynamically matching questions and answers provided herein. The non-transitory computer readable storage medium of the present application stores computer instructions for causing a computer to perform the method of question-answer dynamic matching provided by the present application.
The memory 1102 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created from use of the electronic device dynamically matched in question and answer, and the like. Further, the memory 1102 may include high speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 1102 may optionally include memory located remotely from processor 1101, which may be connected to a dynamic question and answer matching electronic device via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The electronic device of the method for dynamically matching questions and answers may further include: an input device 1103 and an output device 1104. The processor 1101, the memory 1102, the input device 1103 and the output device 1104 may be connected by a bus or other means, and are exemplified by being connected by a bus in fig. 11.
The input device 1103 may receive input numeric or character information and generate key signal inputs related to user settings and function controls of the electronic device that are dynamically matched to the question and answer, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, a pointer, one or more mouse buttons, a track ball, a joystick, or other input device. The output devices 1104 may include a display device, auxiliary lighting devices (e.g., LEDs), tactile feedback devices (e.g., vibrating motors), and the like. The display device may include, but is not limited to, a Liquid Crystal Display (LCD), a Light Emitting Diode (LED) display, and a plasma display. In some implementations, the display device can be a touch screen.
Various implementations of the systems and techniques described here can be realized in digital electronic circuitry, integrated circuitry, application specific ASICs (application specific integrated circuits), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs that are executable and/or interpretable on a programmable system including at least one programmable processor, which may be special or general purpose, receiving data and instructions from, and transmitting data and instructions to, a storage system, at least one input device, and at least one output device.
These computer programs (also known as programs, software applications, or code) include machine instructions for a programmable processor, and may be implemented using high-level procedural and/or object-oriented programming languages, and/or assembly/machine languages. As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, apparatus, and/or device (e.g., magnetic discs, optical disks, memory, Programmable Logic Devices (PLDs)) used to provide machine instructions and/or data to a programmable processor, including a machine-readable medium that receives machine instructions as a machine-readable signal. The term "machine-readable signal" refers to any signal used to provide machine instructions and/or data to a programmable processor.
To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic, speech, or tactile input.
The systems and techniques described here can be implemented in a computing system that includes a back-end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
In summary, the electronic device according to the embodiment of the present application obtains the query statement and the candidate paragraphs; performing entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and constructing a knowledge graph according to the entity vectors; updating each entity vector in the knowledge graph to generate an entity vector to be matched; and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree. The method solves the technical problem that the matching degree between the question and the answer paragraph cannot be accurately calculated in the prior art, so that the determined target paragraph is not accurate enough, improves the accuracy of the target paragraph by dynamically updating and representing the entity vectors in the question and the answer paragraph and calculating the matching degree based on the updated entity vectors, and meets the use requirements of users.
In the description herein, reference to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the application. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Moreover, various embodiments or examples and features of various embodiments or examples described in this specification can be combined and combined by one skilled in the art without being mutually inconsistent.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present application, "plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present application in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present application.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present application may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the various steps or methods may be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are well known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present application may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. Although embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present application, and that variations, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present application.
Claims (12)
1. A question-answer dynamic matching method is characterized by comprising the following steps:
acquiring a query statement and a candidate paragraph;
forming a knowledge graph based on the query statement and the candidate paragraphs, wherein entity recognition is performed on the query statement and the candidate paragraphs to obtain a plurality of entity vectors, and the knowledge graph is constructed according to the entity vectors;
updating each entity vector in the knowledge graph to generate an entity vector to be matched;
and matching the entity vectors to be matched to obtain the matching degree of the query statement and the candidate paragraphs, and determining the target paragraph according to the matching degree.
2. The method of claim 1, wherein said constructing a knowledge-graph from said plurality of entity vectors comprises:
acquiring entity interrelationships among the entity vectors;
acquiring attribute information among the entity vectors;
and connecting the entity vectors according to the entity correlation and the attribute information to generate the knowledge graph.
3. The method of claim 1, wherein the updating each entity vector in the knowledge-graph to generate an entity vector to be matched comprises:
s1, determining a first entity vector to be updated from the plurality of entity vectors, and obtaining an adjacent entity list of the first entity vector to be updated;
s2, extracting a first adjacent entity vector from the adjacent entity list, and acquiring a local knowledge graph corresponding to the first adjacent entity vector from the knowledge graph;
s3, calculating correlation score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similarity score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain a first entity vector to be matched;
s4, taking the first entity vector to be matched as the second entity vector to be updated, and repeatedly executing steps S2 and S3 until each adjacent entity vector in the adjacent entity list is taken out to generate an entity vector to be matched;
and S5, repeatedly executing the step S1 to the step S4, and acquiring the entity vector to be matched in the knowledge graph.
4. The method of claim 3, wherein the obtaining the local knowledge-graph corresponding to the first neighboring entity vector from the knowledge-graph comprises:
and acquiring the local knowledge graph except the first entity vector to be updated in the knowledge graph according to a preset length.
5. The method of claim 2, further comprising:
according to a preset scoring function, scoring the correlation among the entity vectors in the knowledge graph;
and deleting part of entity vectors in the knowledge graph according to the scoring result and a preset threshold value.
6. A dynamic question-answer matching device, comprising:
the acquisition module is used for acquiring a query sentence and a candidate paragraph;
the identification module is used for carrying out entity identification on the query statement and the candidate paragraphs to obtain a plurality of entity vectors;
a construction module for constructing a knowledge graph based on the query statement and the candidate paragraphs, wherein the knowledge graph is constructed from the plurality of entity vectors;
the updating module is used for updating each entity vector in the knowledge graph to generate an entity vector to be matched;
and the processing module is used for matching the entity vector to be matched to obtain the matching degree of the query statement and the candidate paragraph, and determining a target paragraph according to the matching degree.
7. The apparatus of claim 6, wherein the build module comprises:
a first obtaining unit, configured to obtain entity interrelationships among the entity vectors;
a second obtaining unit configured to obtain attribute information between the plurality of entity vectors;
and the generating unit is used for connecting the entity vectors according to the entity correlation and the attribute information to generate the knowledge graph.
8. The apparatus of claim 6, wherein the update module comprises:
a determining unit, configured to determine a first entity vector to be updated from the multiple entity vectors, and obtain an adjacent entity list of the first entity vector to be updated;
a first obtaining unit, configured to take out a first adjacent entity vector from the adjacent entity list, and obtain a local knowledge graph corresponding to the first adjacent entity vector from the knowledge graph;
the calculation unit is used for calculating correlation score values between the first entity vector to be updated and each entity vector in the local knowledge graph, and performing weighted summation processing on each entity vector in the local knowledge graph by taking each similarity score value as a weight coefficient corresponding to each entity vector in the local knowledge graph to obtain a first entity vector to be matched;
the updating unit is used for taking the first entity vector to be matched as a second entity vector to be updated, and repeatedly executing the acquiring unit and the calculating unit until each adjacent entity vector in the adjacent entity list is taken out to generate an entity vector to be matched;
and the second acquisition unit is used for repeatedly executing the determination unit to the updating unit and acquiring the entity vector to be matched in the knowledge graph.
9. The apparatus of claim 8, wherein the first obtaining unit is specifically configured to:
and acquiring the local knowledge graph except the first entity vector to be updated in the knowledge graph according to a preset length.
10. The apparatus of claim 7, further comprising:
the scoring unit is used for scoring the correlation among the entity vectors in the knowledge graph according to a preset scoring function;
and the deleting unit is used for deleting part of entity vectors in the knowledge graph according to the scoring result and a preset threshold value.
11. An electronic device, comprising:
at least one processor; and
a memory communicatively coupled to the at least one processor; wherein,
the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-5.
12. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910942942.4A CN110704600B (en) | 2019-09-30 | 2019-09-30 | Question-answer dynamic matching method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910942942.4A CN110704600B (en) | 2019-09-30 | 2019-09-30 | Question-answer dynamic matching method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110704600A CN110704600A (en) | 2020-01-17 |
| CN110704600B true CN110704600B (en) | 2022-05-13 |
Family
ID=69197474
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910942942.4A Active CN110704600B (en) | 2019-09-30 | 2019-09-30 | Question-answer dynamic matching method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110704600B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111414765B (en) * | 2020-03-20 | 2023-07-25 | 北京百度网讯科技有限公司 | Sentence consistency determination method and device, electronic equipment and readable storage medium |
| CN111488441B (en) * | 2020-04-08 | 2023-08-01 | 北京百度网讯科技有限公司 | Question analysis method and device, knowledge graph question answering system and electronic equipment |
| CN111666969B (en) * | 2020-04-22 | 2021-11-23 | 北京百度网讯科技有限公司 | Method and device for calculating image-text similarity, electronic equipment and readable storage medium |
| CN111737430B (en) * | 2020-06-16 | 2024-04-05 | 北京百度网讯科技有限公司 | Entity linking method, device, equipment and storage medium |
| CN112434149B (en) * | 2020-06-24 | 2023-09-19 | 北京金山数字娱乐科技有限公司 | Information extraction method, information extraction device, information extraction equipment and storage medium |
| CN113887225A (en) * | 2020-07-02 | 2022-01-04 | 北京金山数字娱乐科技有限公司 | Word processing method and device based on multitask model |
| CN112784058B (en) * | 2021-01-11 | 2022-04-22 | 北京欧拉认知智能科技有限公司 | Entity correlation obtaining method based on dynamic map |
| CN112883735B (en) * | 2021-02-10 | 2024-01-12 | 卡奥斯数字科技(上海)有限公司 | Method, device, equipment and storage medium for structured processing of form image |
| CN113010663A (en) * | 2021-04-26 | 2021-06-22 | 东华大学 | Adaptive reasoning question-answering method and system based on industrial cognitive map |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647233A (en) * | 2018-04-02 | 2018-10-12 | 北京大学深圳研究生院 | A kind of answer sort method for question answering system |
| CN110059160A (en) * | 2019-04-17 | 2019-07-26 | 东南大学 | A kind of knowledge base answering method and device based on context end to end |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10832145B2 (en) * | 2015-10-05 | 2020-11-10 | International Business Machines Corporation | Techniques for resolving entities in received questions |
| US10248689B2 (en) * | 2015-10-13 | 2019-04-02 | International Business Machines Corporation | Supplementing candidate answers |
-
2019
- 2019-09-30 CN CN201910942942.4A patent/CN110704600B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108647233A (en) * | 2018-04-02 | 2018-10-12 | 北京大学深圳研究生院 | A kind of answer sort method for question answering system |
| CN110059160A (en) * | 2019-04-17 | 2019-07-26 | 东南大学 | A kind of knowledge base answering method and device based on context end to end |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110704600A (en) | 2020-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110704600B (en) | Question-answer dynamic matching method and device and electronic equipment | |
| JP7262571B2 (en) | Knowledge graph vector representation generation method, apparatus and electronic equipment | |
| JP7316453B2 (en) | Object recommendation method and device, computer equipment and medium | |
| CN111259671B (en) | Semantic description processing method, device and equipment for text entity | |
| US20210356290A1 (en) | Method and apparatus for recommending point of interest, device, and medium | |
| CN111582477B (en) | Training method and device for neural network model | |
| CN113094550A (en) | Video retrieval method, device, equipment and medium | |
| CN111274407B (en) | Method and device for calculating triplet confidence in knowledge graph | |
| EP3882791A1 (en) | Method, apparatus for content recommendation, electronic device and storage medium | |
| CN110543558B (en) | Question matching method, device, equipment and medium | |
| CN110114784A (en) | Recurrent neural network | |
| CN111353299B (en) | Dialog scene determining method based on artificial intelligence and related device | |
| CN111582479A (en) | Distillation method and device of neural network model | |
| CN112632403A (en) | Recommendation model training method, recommendation device, recommendation equipment and recommendation medium | |
| JP7289337B2 (en) | Building extraction method, device, equipment and storage medium | |
| CN112541362B (en) | Generalization processing method, device, equipment and computer storage medium | |
| CN111079945A (en) | End-to-end model training method and device | |
| CN111324715A (en) | Method and device for generating question-answering robot | |
| CN111241242B (en) | Method, device, equipment and computer readable storage medium for determining target content | |
| CN111680600A (en) | Face recognition model processing method, device, equipment and storage medium | |
| CN111753761A (en) | Model generation method and device, electronic equipment and storage medium | |
| CN112241716A (en) | Training sample generation method and device | |
| JP2022166215A (en) | Method for training text positioning model and method for text positioning | |
| CN114111813A (en) | High-precision map element updating method and device, electronic equipment and storage medium | |
| CN112328896B (en) | Method, apparatus, electronic device, and medium for outputting information |
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 |