CN110427467A - Question and answer processing method, device, computer equipment and storage medium - Google Patents
Question and answer processing method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110427467A CN110427467A CN201910560681.XA CN201910560681A CN110427467A CN 110427467 A CN110427467 A CN 110427467A CN 201910560681 A CN201910560681 A CN 201910560681A CN 110427467 A CN110427467 A CN 110427467A
- Authority
- CN
- China
- Prior art keywords
- user
- question sentence
- entity
- intention
- matching
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Animal Behavior & Ethology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This application involves a kind of question and answer processing method, device, computer equipment and storage mediums.The described method includes: receiving user inquires instruction, inquire that instruction obtains user's question sentence according to user；It identifies user's question sentence, entity and intention is determined from user's question sentence；Using the incidence relation Matching Model matching entities and intention trained, matching result is obtained；Answer corresponding with user's question sentence is searched in default bivariate table knowledge base according to matching result.It can be improved to obtain the accuracy that user's question sentence corresponds to answer using this method.
Description
Technical field
This application involves Internet technical field, more particularly to a kind of question and answer processing method, device, computer equipment and
Storage medium.
Background technique
With the development of internet technology, artificial customer service can not meet the growing demand of enterprise, in order to reduce
Entreprise cost, enterprise begin to use customer service robot to answer the problem of user.Enterprise is by establishing customer service robot
Knowledge mapping, then customer service robot is replied according to searching corresponding answer the problem of user in knowledge mapping.Mesh
Before, customer service robot is usually to obtain the intention of user's question sentence when carrying out voice response, answer is matched according to intention, thus
It is replied according to answer.It is this directly to make to obtain when the intention of user's question sentence occurs more using intention matching answer
Answer accuracy it is low.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of question and answer processing method that can be improved search efficiency,
Device, computer equipment and storage medium.
A kind of question and answer processing method, which comprises
It receives user and inquires instruction, inquire that instruction obtains user's question sentence according to user；
It identifies user's question sentence, entity and intention is determined from user's question sentence；
Using the incidence relation Matching Model matching entities and intention trained, matching result is obtained；
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base according to matching result.
In one of the embodiments, receive user inquire instruction, according to user inquire instruction obtain user's question sentence it
Before, comprising:
Default bivariate table knowledge base is established, presetting includes entity, intention and entity and intention pair in bivariate table knowledge base
The answer answered.
User's question sentence is identified in one of the embodiments, and entity and intention are determined from user's question sentence, comprising:
User's question sentence is identified using the Named Entity Extraction Model trained, determines the entity in user's question sentence；
User's question sentence is identified using the intention assessment model trained, determines the intention in user's question sentence.
User's question sentence is identified using the Named Entity Extraction Model trained in one of the embodiments, determines and uses
Entity in the question sentence of family, comprising:
User's question sentence is segmented, word segmentation result is obtained；
Term vector is obtained according to word segmentation result, term vector is input to the Named Entity Extraction Model identification trained, is obtained
To Entity recognition result vector；
The entity in user's question sentence is determined according to Entity recognition result vector.
User's question sentence is identified using the intention assessment model trained in one of the embodiments, determines that user asks
Intention in sentence, comprising:
User's question sentence is segmented, word segmentation result is obtained；
Term vector is obtained according to word segmentation result, term vector is input to the intention assessment model identification trained, is anticipated
Figure recognition result vector；
The intention in user's question sentence is determined according to intention assessment result vector.
Matching entities and intention in one of the embodiments, obtain matching result, comprising:
Entity and intention are matched, matching result to be determined is obtained；
Matching vector to be determined is obtained according to matching result to be determined, matching vector to be determined is input to the pass trained
It is identified in connection relationship match model, obtains match cognization result vector；
Object matching result is determined according to match cognization result vector.
It is corresponding to search user's question sentence in default bivariate table knowledge base according to matching result in one of the embodiments,
Answer, comprising:
The target entity in matching result is searched in the entities field in default bivariate table knowledge base, and in default two dimension
In intention field in table knowledge library search matching result in the matched target intention of target entity；
When finding target entity and target intention, target entity and target meaning are obtained from default bivariate table knowledge base
Scheme corresponding answer.
User's question sentence correspondence is being searched in default bivariate table knowledge base according to matching result in one of the embodiments,
Answer after, further includes:
The corresponding answer of user's question sentence is returned into terminal, so that terminal display answer.
A kind of question and answer processing unit, described device include:
Question sentence obtains module, inquires instruction for receiving user, inquires that instruction obtains user's question sentence according to user；
Identification module, user's question sentence obtains the corresponding entity of user's question sentence and intention for identification；
Matching module is obtained for using the incidence relation Matching Model trained to match the entity and the intention
Matching result；
Searching module is tied for being matched in default entity and intention bivariate table according to association results according to matching
Fruit obtains the corresponding answer of user's question sentence.
Question and answer processing unit in one of the embodiments, comprising:
Bivariate table establishes module, and for establishing default bivariate table knowledge base, presetting includes entity, meaning in bivariate table knowledge base
Figure and entity and the corresponding answer of intention.
Identification module in one of the embodiments, comprising:
Entity determining module determines user for using the Named Entity Extraction Model trained to identify user's question sentence
Entity in question sentence；
It is intended to determining module, for using the intention assessment model trained to identify user's question sentence, determines user's question sentence
In intention.
Entity determining module in one of the embodiments, comprising:
Word segmentation module obtains word segmentation result for segmenting user's question sentence；
It is real to be input to the name trained for obtaining term vector according to word segmentation result by Entity recognition module for term vector
The identification of body identification model, obtains Entity recognition result vector；The entity in user's question sentence is determined according to Entity recognition result vector.
It is intended to determining module in one of the embodiments, comprising:
Word segmentation module obtains word segmentation result for segmenting user's question sentence；
Term vector is input to the intention trained and known by intention assessment module for obtaining term vector according to word segmentation result
Other model identification, obtains intention assessment result vector；The intention in user's question sentence is determined according to intention assessment result vector.
Matching module in one of the embodiments, comprising:
As a result module is obtained, for matching to entity and intention, obtains matching result to be determined；
As a result identification module, for obtaining matching vector to be determined according to matching result to be determined, by it is to be determined match to
Amount, which is input in the incidence relation Matching Model trained, to be identified, match cognization result vector is obtained；
As a result determining module, for determining object matching result according to match cognization result vector.
Searching module in one of the embodiments, comprising:
Target searching module, for searching the target in matching result in the entities field in default bivariate table knowledge base
Entity, and search in matching result in the intention field in default bivariate table knowledge base and anticipate with the matched target of target entity
Figure；
Answer obtains module, for being obtained from default bivariate table knowledge base when finding target entity and target intention
Take target entity and the corresponding answer of target intention.
Question and answer processing unit in one of the embodiments, further includes:
Answer return module, for the corresponding answer of user's question sentence to be returned to terminal, so that terminal display answer.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device performs the steps of when executing the computer program
It receives user and inquires instruction, inquire that instruction obtains user's question sentence according to user；
It identifies user's question sentence, entity and intention is determined from user's question sentence；
The entity and the intention are matched using the incidence relation Matching Model trained, obtains matching result；
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base according to matching result.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
It receives user and inquires instruction, inquire that instruction obtains user's question sentence according to user；
It identifies user's question sentence, entity and intention is determined from user's question sentence；
The entity and the intention are matched using the incidence relation Matching Model trained, obtains matching result；
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base according to matching result.
Above-mentioned question and answer processing method, device, computer equipment and storage medium, by identification user's question sentence in entity and
It is intended to, then by the incidence relation Matching Model matching entities and intention trained, according to matching result in default bivariate table
Answer corresponding with user's question sentence is searched in knowledge base.When there are multiple entities or it is multiple intention or exist simultaneously multiple
When entity and multiple intentions, the accuracy for obtaining matching result is improved, can accurately obtain associated entity and is intended to.From
And the case where there are when multiple entities and multiple intentions, entity is with associated errors are intended to is reduced, it is then direct using matching result
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base, improves to obtain the accuracy of question sentence answer.
Detailed description of the invention
Fig. 1 is the application scenario diagram of question and answer processing method in one embodiment；
Fig. 2 is the flow diagram of question and answer processing method in one embodiment；
Fig. 3 is the flow diagram that user's question sentence entity and intention are determined in one embodiment；
Fig. 4 is the flow diagram that user's question sentence entity is determined in one embodiment；
Fig. 5 is the flow diagram for determining user's question sentence in one embodiment and being intended to；
Fig. 6 is the flow diagram that object matching result is determined in one embodiment；
Fig. 7 is the flow diagram for searching answer in one embodiment in default bivariate table database；
Fig. 7 a is the application scenario diagram of question and answer processing method in a specific embodiment；
Fig. 8 is the flow diagram of question and answer processing method in a specific embodiment；
Fig. 9 is the structural block diagram of question and answer processing method device in one embodiment；
Figure 10 is the internal structure chart of computer equipment in one embodiment；
Figure 11 is the internal structure chart of computer equipment in another embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Question and answer processing method provided by the present application, can be applied in application environment as shown in Figure 1.Wherein, terminal 102
It is communicated by network with server 104.Server 104 receives the user that terminal 102 is sent and inquires instruction, is ask according to user
Ask that instruction obtains user's question sentence；Server 104 identifies user's question sentence, and entity and intention are determined from user's question sentence；Server 104
Matching entities and intention, obtain matching result；Server 104 is searched according to matching result and is used in default bivariate table knowledge base
The corresponding answer of family question sentence.Then the corresponding answer of user's question sentence can be returned into terminal 102.Wherein, terminal 102 can
With but be not limited to various personal computers, laptop, smart phone, tablet computer and portable wearable device, take
Business device 104 can be realized with the server cluster of the either multiple server compositions of independent server.
In one embodiment, as shown in Fig. 2, providing a kind of question sentence processing method, it is applied in Fig. 1 in this way
It is illustrated for server, comprising the following steps:
S202 receives user and inquires instruction, inquires that instruction obtains user's question sentence according to user.
Specifically, terminal gets the INQUIRE statement of user, the user speech that can be obtained by voice device, by user
Voice is converted to user's question sentence text.User's question sentence text of user's input can also be got by input unit.At this time eventually
End sends user to server according to user file text and inquires instruction, and server receives the user and inquires that instruction, parsing should
User inquires instruction, obtains user's question sentence text.
S204 identifies user's question sentence, and entity and intention are determined from user's question sentence.
Wherein, entity refers to certain types of word in user's question sentence, for example, name, place name, mechanism name and proper noun etc.
Deng.It is intended to refer to the relationship or attribute that user wants inquiry for entity.Such as: what inquiry name is, that is, is inquired
Name entity attributes.
Specifically, server identifies user's question sentence using machine learning algorithm, determines that the user asks from user's question sentence
The corresponding entity of sentence and intention.Single entity and single intention can be determined from user's question sentence.For example, user's question sentence is
" Xiao Ming's lives at which ".It is assured that out that entity is " Xiao Ming " and is intended to " family " from user's question sentence.It can also be from
Determine that single entity and multiple intentions perhaps multiple entities and single are intended to or multiple entities and more in user's question sentence
A intention etc..For example, user's question sentence can be " the Changjiang river security and the stock code of Iflytek what is ".From user's question sentence
In determine two entities " the Changjiang river security " and " Iflytek ", it is single to be intended to " stock code ".It is determined from user's question sentence
Intention can be the word for including in user's question sentence, be also possible to the word being not included in user's question sentence.For example, " the parent of Xiao Ming
Whom relative has ".The entity identified is that the intention that " Xiao Ming " identifies can be " uncle's title ", " aunt's title ", " one's mother's sister
Husband's title " etc..
S206, the incidence relation Matching Model that use has been trained match the entity and the intention, obtain matching result.
Wherein, the incidence relation Matching Model that use has been trained using machine learning algorithm according to existing labeled data into
Row training obtains.Wherein machine learning algorithm can be linear regression algorithm, neural network algorithm etc..For example, according to
Associated entity is labeled as " association " with intention by entity and intention in some question sentences, by not associated entity and is intended to mark
Note is " not associated ", is then trained using the data of mark and neural network algorithm, the incidence relation trained
With model.Matching entities and intention, which refer to, establishes incidence relation for the entity and intention that identify from user's question sentence.
Specifically, the corresponding intention of incidence relation Matching Model matching entities that use has been trained, obtains matching result.It should
Matching result can be entity corresponding one and be intended to as a result, an entity corresponds to multiple intentions or multiple
The corresponding intention of entity or multiple entities correspond to multiple intentions etc..For example, determining two realities from user's question sentence
Body " the Changjiang river security " and " Iflytek ", it is single to be intended to " stock code ".It is matched according to the matching rule pre-set
Obtaining matching result includes " the Changjiang river security " and " stock code " and " Iflytek " and " stock code ", i.e. two entities pair
Should individually it be intended to.When user's question sentence is, " stock code of the Changjiang river security is whom the president of what and Iflytek is " is determined
Entity " the Changjiang river security " and " Iflytek " and intention " stock code " and " president ", the incidence relation matching that use has been trained
Model is matched, and matching result " the Changjiang river security and stock code " " Iflytek and president " is obtained.
S208 searches answer corresponding with user's question sentence according to matching result in default bivariate table knowledge base.
Wherein, the two-dimensional data table that bivariate table knowledge base is the user's question sentence answer pre-set is preset, which is
Refer to entities dimension and is intended to dimension.
Specifically, according to matching result, that is, entity and corresponding intention, in entities dimension in default bivariate table knowledge base
It is middle first to find entity, then the corresponding intention of the entity is found in being intended to dimension.It is corresponding with entity when finding entity
Intention when, obtain the internal information of statement of entity with corresponding intention, the i.e. answer of user's question sentence.
In the above-described embodiments, it by the entity and intention in identification user's question sentence, is then closed by the association trained
It is Matching Model matching entities and intention, is searched in default bivariate table knowledge base according to matching result corresponding with user's question sentence
Answer.When improving to obtain there are multiple entities or multiple intentions or when existing simultaneously multiple entities and multiple intentions
The accuracy of matching result can accurately obtain associated entity and be intended to.To reduce, there are multiple entities and multiple meanings
When figure, entity and the case where be intended to associated errors, then searched directly in default bivariate table knowledge base using matching result and
The corresponding answer of user's question sentence improves to obtain the accuracy of question sentence answer.And according to matching result in default bivariate table
Answer corresponding with user's question sentence is searched in knowledge base.To search question sentence answer without the use of knowledge mapping, that is, do not need
User's question sentence is converted to special search statement to search in knowledge mapping, is directly existed using entity and the matching result of intention
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base, improves the search efficiency of question sentence answer.
In one embodiment, before step S202, that is, user's inquiry instruction is being received, is inquiring that instruction obtains according to user
Before taking family question sentence, comprising steps of
Default bivariate table knowledge base is established, presetting includes entity, intention and entity and intention pair in bivariate table knowledge base
The answer answered.
Specifically, terminal receives the instruction that user establishes bivariate table knowledge base, and establishes bivariate table knowledge base according to this
Instruction get entities dimension information, it is intended that dimensional information and user's answer information.Then terminal to server transmission is built
The instruction of vertical bivariate table knowledge base, server receive the instruction for establishing the bivariate table knowledge base, get reality according to the instruction
Body dimensional information, it is intended that dimensional information and user's answer information.According to the entities dimension information, it is intended that dimensional information and use
Family answer information establishes default bivariate table knowledge base.In this embodiment, by pre-establishing bivariate table knowledge base, facilitate subsequent
Use.In a specific embodiment, partial content is as shown in table 1 in the bivariate table knowledge base of foundation:
Table 1
In one embodiment, as shown in figure 3, step S204, i.e. identification user's question sentence, determine entity from user's question sentence
And intention, comprising steps of
S302, the Named Entity Extraction Model that use has been trained identifies user's question sentence, determines the entity in user's question sentence.
Wherein, the Named Entity Extraction Model entity in user's question sentence for identification, is using name entity identification algorithms
The model of foundation, the name entity identification algorithms can be Keywords matching algorithm, template matching algorithm or sequence labelling and calculate
Method etc..
Specifically, can server using by user's question sentence as Keywords matching algorithm, template matching algorithm or sequence mark
Infuse algorithm input, Entity recognition is named according to user's question sentence, obtain Keywords matching algorithm, template matching algorithm or
The entity of sequence labelling algorithm output is to get having arrived the entity that user's question sentence is included.In one embodiment, sequence can be used
Column dimensioning algorithm is as name entity identification algorithms.Sequence mark is used previously according to existing user's question sentence and the entity marked out
Note algorithm is trained, using drop user's question sentence as the input of sequence labelling algorithm, using the entity marked out as sequence labelling
The label of algorithm is trained, when reaching trained completion condition, the Named Entity Extraction Model trained.The training is complete
It can be trained the number of iterations at condition and be less than preset threshold more than preset threshold or the value of loss function.
S304, the intention assessment model that use has been trained identifies user's question sentence, determines the intention in user's question sentence.
Wherein, it is intended that identification model user identifies the intention in user's question sentence, is according to the training data of mark using machine
Device learning algorithm, search technique scheme or the training of other algorithms obtain.
Specifically, server is using user's question sentence as the input of intention assessment model, it is intended that identification model is answered according to input
User's question sentence carries out intention assessment, obtains the intention of intention assessment model output to get intention included in user's question sentence is arrived.
In one embodiment, it is exported according to existing training data using depth mind network algorithm training intention assessment model, this is existing
Training data include existing user's question sentence and intention that user's question sentence has marked out.Using user's question sentence as depth mind
The intention marked out is trained by the input of network algorithm as label, when reaching preset condition, i.e. depth mind network
When the number of iterations of algorithm is less than preset threshold more than preset threshold or the loss function value of depth mind network algorithm, train
At the intention assessment model trained.
In the above-described embodiments, server passes through the Named Entity Extraction Model trained and intention assessment model and can know
Not Chu entity and intention in user's question sentence, improve the efficiency and accuracy of the identification of user's question sentence.
In one embodiment, as shown in figure 4, step S302, i.e., using the Named Entity Extraction Model trained to
The identification of family question sentence, determines the entity in user's question sentence, comprising steps of
User's question sentence is segmented, obtains word segmentation result by S402.
Wherein, participle is the process that continuous word sequence is reassembled into word sequence according to certain specification.
Specifically, segmenting method of the server based on string matching, the segmenting method based on understanding or based on statistics
Segmenting method user's question sentence is segmented, the word sequence after being segmented.
S404 obtains term vector according to word segmentation result, and term vector is input to the Named Entity Extraction Model trained and is known
Not, Entity recognition result vector is obtained.
Specifically, word sequence is mapped in vector space by server according to the word sequence after participle, obtain word to
Amount.Bag of words can be used, word segmentation result vectorization is obtained into term vector, it (is that a group is used to produce that word2vec, which also can be used,
The correlation model of new word vector.These models are the shallow and double-deck neural network, are used to training with the word of construction linguistics again
Text) word segmentation result vectorization obtains term vector etc. by algorithm.Obtained term vector is input to the life trained by server
It is identified in name entity recognition model, obtains Entity recognition result vector.The Entity recognition result vector is for indicating the use
The entity for including in the question sentence of family.
S406 determines the entity in user's question sentence according to Entity recognition result vector.
Specifically, the corresponding relationship of the Entity recognition result vector and entity that obtain in training is obtained.For example, in entity
Non-physical is expressed as " 0 " in recognition result vector.Entity uses " LOC ", and " PER ", " ORG ", " MISC " etc. respectively indicates position,
Name, organization name and miscellaneous etc..
The entity in the corresponding user's question sentence of Entity recognition result vector obtained at this time according to the corresponding relationship.Than
Such as: such as user's question sentence is " what the work of Xiao Ming is ", is according to the Entity recognition result vector that user's question sentence obtains
(per,0,0,0).Obtaining entity is " Xiao Ming ".
In the above-described embodiments, term vector is obtained by first segmenting user's question sentence, is then being named using term vector
It is identified in entity recognition model identification, obtains recognition result vector, the entity in user's question sentence is obtained according to recognition result vector,
It improves to obtain the accuracy of entity.
In one embodiment, as shown in figure 5, step S304, the intention assessment model that use has been trained is to user's question sentence
Identification, determines the intention in user's question sentence, comprising steps of
User's question sentence is segmented, obtains word segmentation result by S502.
S504 obtains term vector according to word segmentation result, and term vector is input to the intention assessment model identification trained, is obtained
To intention assessment result vector.
Specifically, server, which can be, segments user's question sentence based on various segmentation methods, then obtains participle knot
Fruit, segmentation methods can be the segmentation methods based on string matching, the segmentation methods based on understanding or point based on statistics
Word algorithm.Then according to the word sequence after participle, word sequence is mapped in vector space, obtains term vector.By term vector
It is input in the intention assessment model trained and is identified, the intention assessment result vector exported.The intention assessment knot
Fruit vector is used to indicate the intention for including in user's question sentence.For example, user's question sentence " what work of Xiao Ming is " is segmented,
Obtain word segmentation result " Xiao Ming ", " ", " work ", "Yes" and " what ".It is obtained according to word segmentation result using word2vec algorithm
Term vector.
S506 determines the intention in user's question sentence according to intention assessment result vector.
Specifically, according to the intention assessment result vector being arranged in the intention assessment model being trained and pair of intention
It should be related to obtain the corresponding intention of intention assessment result vector, obtain the intention in user's question sentence.The corresponding relationship refers to use
The word is to be intended to then be indicated with " 1 " after family question sentence participle, is not intended to then be identified with " 0 ".For example, user's question sentence " work of Xiao Ming
What work is ", obtained intention assessment result vector can be (0,1,0,0).Wherein, the word in " 1 " corresponding user's question sentence
It is " work ", then the intention in the user file exactly " works ".
In the above-described embodiments, by the way that according to word segmentation result term vector will be obtained after user's question sentence participle, using word to
It measures in the intention assessment model identification trained, the intention in user's question sentence, energy is finally obtained according to intention assessment result vector
Enough improve obtains the accuracy being intended in user's question sentence.
In one embodiment, as shown in fig. 6, step S206, i.e., matched using the incidence relation Matching Model trained
Entity and intention, obtain matching result, comprising steps of
S602 matches entity and intention, obtains matching result to be determined.
Wherein, matching result to be determined, which refers to, is associated the result obtained later for entity and intention at random.
Specifically, if what is obtained is single entity and single intention, available unique matching result to be determined.
If what is obtained is multiple entities and single intention, multiple entities are associated with the single intention one by one directly, are obtained multiple
Matching result to be determined.If what is obtained is single entity and multiple intentions, directly by multiple intentions and the single entity
It is associated with one by one, it can obtain multiple matching results to be determined.It, at random should when obtained multiple entities and multiple intentions
Multiple entities and multiple intentions are associated one by one, obtain multiple and different matching results to be determined.
S604 obtains matching vector to be determined according to matching result to be determined, matching vector to be determined is input to and has been instructed
It is identified in experienced incidence relation Matching Model, obtains match cognization result vector.
Wherein, the incidence relation Matching Model trained is to be carried out using machine learning algorithm according to existing labeled data
What training obtained.For example, according to existing entity and intention, the random association for carrying out entity and intention obtains matching result.So
Real associated matching result is labeled as " 1 " afterwards.Not associated matching result is labeled as " 0 ", obtains annotation results.It will
Input of the matching result as machine learning algorithm, is trained annotation results as the label of machine learning algorithm, when reaching
To incidence relation Matching Model when training completion condition, trained.The training completion condition can be trained iteration
Number is less than preset threshold value more than the loss function of preset number or machine learning algorithm.The machine learning algorithm can
To be logistic regression algorithm, neural network algorithm etc..
Specifically, matching result vectorization to be determined is obtained matching vector to be determined by server, by it is to be determined match to
Amount is input in the incidence relation Matching Model trained the match cognization result vector for being identified, being exported.
S606 determines object matching result according to match cognization result vector.
Specifically, server obtains corresponding of the match cognization result vector according to the corresponding relationship being arranged in training
With as a result, the matching result as object matching result.For example, matching result to be determined includes " being intended to 1 and entity 1 ", " is intended to
2 and entity 2 " and " be intended to 3 and entity 3 ".Obtained match cognization result vector can be (0,1,0).It is arranged when according to training
Corresponding relationship obtain the corresponding matching result of vector " 1 " " being intended to 2 and entity 2 ", " 2 and entity 2 will be intended to " and be used as target
Matching result, the object matching result are exactly finally to confirm associated entity and intention.
In this embodiment, it when having identified multiple entities and multiple intentions, can be obtained using incidence relation Matching Model
To real associated matching result, using the associated matching result as object matching as a result, being obtained using object matching result
The answer of user's question sentence, the answer that can make are more accurate.
In one embodiment, as shown in fig. 7, step S208, i.e., according to matching result in default bivariate table knowledge base
The corresponding answer of user's question sentence is searched, comprising steps of
S702 searches the target entity in matching result in the entities field in default bivariate table knowledge base, and pre-
If in the intention field in bivariate table knowledge base search matching result in the matched target intention of target entity.
S704, when finding target entity and target intention, from default bivariate table knowledge base obtain target entity and
The corresponding answer of target intention.
Wherein, entities field, which refers to, is arranged in field belonging to entity all in the bivariate table knowledge base, it is intended that word
Field belonging to intention all in the bivariate table knowledge base is arranged in section.
Specifically, when obtaining matching result, server lookup in the entities field in default bivariate table knowledge base
With the target entity in result, and this preset in the intention field in bivariate table knowledge base search matching result in target reality
The matched target intention of body.When can be found in entities field target entity and be intended to field in find target intention
When, illustrate that default bivariate table knowledge base is stored with target entity and the corresponding answer of target intention.Then know from default bivariate table
Know in the table corresponding with target entity and target intention of library and get answer in unit, which is exactly the target entity and target
It is intended to corresponding answer, the i.e. answer of user's question sentence.
When not finding target entity in entities field or not finding target intention in being intended to field, illustrate
It is default that bivariate table knowledge base is not stored target entity and the corresponding answer of target intention.It is answered at this point, user's question sentence does not exist
Case.Server prompts inquiry error.
In the above-described embodiments, by searching the answer of user's question sentence in default bivariate table knowledge base, do not need by
User's question sentence is converted to special query statement, it is only necessary to entity and meaning in default bivariate table knowledge storehouse matching user's question sentence
Figure, so that it may obtain the answer of user's question sentence, can be improved to obtain the efficiency of user's question sentence answer.
In one embodiment, default bivariate table knowledge base translation bit knowledge mapping can be stored.It specifically, will be pre-
If head entity and tail entity of the answer as knowledge mapping triple in entities field and corresponding table in bivariate table knowledge base, will
Corresponding entities field is established knowledge mapping and is stored as the relationship in triple in default bivariate table knowledge base.
In one embodiment, it after step S208, i.e., is being looked into default bivariate table knowledge base according to matching result
It looks for after the corresponding answer of user's question sentence, further comprises the steps of:
The corresponding answer of user's question sentence is returned into terminal, so that terminal display answer.
Wherein, terminal is used to receive the answer of user's question sentence and is shown, and it is corresponding which can be user's question sentence
Terminal, be also possible to other terminals.The terminal be not limited to personal computer, laptop, smart phone, tablet computer and
Portable wearable device.Displaying is not limited by text or image shows, played by voice, by video playing etc.
Deng.For example, the answer received can be shown that terminal can also be in video playing circle by terminal in terminal display interface
Face carries out display broadcasting, can also be converted to voice messaging by voice device and play out.
Specifically, the obtained corresponding answer of user's question sentence can be returned to the corresponding terminal of user's question sentence by server,
The corresponding terminal of user's question sentence is shown after receiving the answer, so that user is obtained the answer information of question sentence, user is facilitated to make
With.The terminal that the corresponding answer of user's question sentence can also be returned to user or server settings, in the terminal display of setting
The corresponding answer of user's question sentence.For example, as shown in Figure 7a, user is received by mobile phone 7A in one specifically embodiment
User's question sentence, server 7B get user's question sentence, find user by question and answer processing method in any of the above-described embodiment and ask
User's question sentence answer is shown by sentence answer back to computer 7C set by user.
In a specific embodiment, as shown in figure 8, the question and answer processing method comprising steps of
Bivariate table knowledge base is completed in S802, building, includes entity, intention and entity and meaning in the bivariate table knowledge base
Scheme corresponding answer.
S804 obtains intention assessment model using the intention training in deep learning algorithm and bivariate table knowledge base.
S806 gets user's question sentence, using user's question sentence as the input of intention assessment model, obtains user's question sentence and is wrapped
The intention I and intention M contained.
S808 obtains the name entity E for including in user's question sentence using user's question sentence as the input of Keywords matching algorithm
With name entity N.
Obtained intention and entity are matched using the incidence relation Matching Model trained, are matched by S810
As a result E*I.
S812 searches corresponding user's question sentence answer according to matching result E*I from bivariate table knowledge base.
The terminal that the answer of user's question sentence returns to user's question sentence is shown by S814.
It should be understood that although each step in the flow chart of Fig. 2-8 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-8
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in figure 9, providing a kind of question and answer processing unit 900, comprising: question sentence obtains module
900, identification module 904, matching module 906 and searching module 908, in which:
Question sentence obtains module 900, inquires instruction for receiving user, inquires that instruction obtains user's question sentence according to user；
Identification module 904, user's question sentence, determines entity and intention from user's question sentence for identification；
Matching module 906 obtains matching knot for using the incidence relation Matching Model matching entities trained and intention
Fruit；
Searching module 908, it is corresponding with user's question sentence for being searched in default bivariate table knowledge base according to matching result
Answer.
In the above-described embodiments, module 900 is obtained by question sentence and obtains user's question sentence, determined and used by identification module 904
Entity and intention in the question sentence of family, matching entities and intention, obtain matching result in matching module 906, finally by lookup
Search answer corresponding with user's question sentence in module 908 in default bivariate table knowledge base according to matching result.By each
The execution of module is not only able to raising and obtains the accuracy of user's question sentence answer, additionally it is possible to quickly find answering for user's question sentence
Case improves the efficiency of inquiry.
In one embodiment, question and answer processing unit 900, comprising:
Bivariate table establishes module, and for establishing default bivariate table knowledge base, presetting includes entity, meaning in bivariate table knowledge base
Figure and entity and the corresponding answer of intention.
In one embodiment, identification module 904, comprising:
Entity determining module determines user for using the Named Entity Extraction Model trained to identify user's question sentence
Entity in question sentence；
It is intended to determining module, for using the intention assessment model trained to identify user's question sentence, determines user's question sentence
In intention.
In one embodiment, entity determining module, comprising:
Word segmentation module obtains word segmentation result for segmenting user's question sentence；
It is real to be input to the name trained for obtaining term vector according to word segmentation result by Entity recognition module for term vector
The identification of body identification model, obtains Entity recognition result vector；The entity in user's question sentence is determined according to Entity recognition result vector.
In one embodiment, it is intended that determining module, comprising:
Word segmentation module obtains word segmentation result for segmenting user's question sentence；
Term vector is input to the intention trained and known by intention assessment module for obtaining term vector according to word segmentation result
Other model identification, obtains intention assessment result vector；The intention in user's question sentence is determined according to intention assessment result vector.
In one embodiment, matching module 906, comprising:
As a result module is obtained, for matching to entity and intention, obtains matching result to be determined；
As a result identification module, for obtaining matching vector to be determined according to matching result to be determined, by it is to be determined match to
Amount, which is input in the incidence relation Matching Model trained, to be identified, match cognization result vector is obtained；
As a result determining module, for determining object matching result according to match cognization result vector.
In one embodiment, searching module 908, comprising:
Target searching module, for searching the target in matching result in the entities field in default bivariate table knowledge base
Entity, and search in matching result in the intention field in default bivariate table knowledge base and anticipate with the matched target of target entity
Figure；
Answer obtains module, for being obtained from default bivariate table knowledge base when finding target entity and target intention
Take target entity and the corresponding answer of target intention.
In one embodiment, question and answer processing unit 900, further includes:
Answer return module, for answer to be returned to the corresponding terminal of user's question sentence, so that the corresponding terminal of user's question sentence
Show answer.
Specific about question and answer processing unit limits the restriction that may refer to above for question and answer processing method, herein not
It repeats again.Modules in above-mentioned question and answer processing unit can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 10.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is for storing bivariate table database.The network interface of the computer equipment is used to pass through with external terminal
Network connection communication.To realize a kind of question and answer processing method when the computer program is executed by processor.
In one embodiment, a kind of computer equipment is provided, which can be terminal, internal structure
Figure can be as shown in figure 11.The computer equipment includes the processor connected by system bus, memory, network interface, shows
Display screen and input unit.Wherein, the processor of the computer equipment is for providing calculating and control ability.The computer equipment
Memory includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system and computer
Program.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The meter
The network interface for calculating machine equipment is used to communicate with external terminal by network connection.When the computer program is executed by processor
To realize a kind of question and answer processing method.The display screen of the computer equipment can be liquid crystal display or electric ink is shown
Screen, the input unit of the computer equipment can be the touch layer covered on display screen, be also possible on computer equipment shell
Key, trace ball or the Trackpad of setting can also be external keyboard, Trackpad or mouse etc..
It will be understood by those skilled in the art that structure shown in Figure 10 or Figure 11, only related to application scheme
Part-structure block diagram, do not constitute the restriction for the computer equipment being applied thereon to application scheme, it is specific to count
Calculating machine equipment may include perhaps combining certain components or with different portions than more or fewer components as shown in the figure
Part arrangement.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, which performs the steps of when executing computer program receives user's inquiry instruction, is inquired according to user
Instruction obtains user's question sentence；It identifies user's question sentence, entity and intention is determined from user's question sentence；Use the incidence relation trained
Matching Model matching entities and intention, obtain matching result；It is searched and is used in default bivariate table knowledge base according to matching result
The corresponding answer of family question sentence.
In one embodiment, foundation default bivariate table is also performed the steps of when processor executes computer program to know
Know library, presetting in bivariate table knowledge base includes entity, intention and entity and the corresponding answer of intention.
In one embodiment, it is also performed the steps of when processor executes computer program using the name trained
Entity recognition model identifies user's question sentence, determines the entity in user's question sentence；Using the intention assessment model trained to
The identification of family question sentence, determines the intention in user's question sentence.
In one embodiment, it is also performed the steps of when processor executes computer program and divides user's question sentence
Word obtains word segmentation result；Term vector is obtained according to word segmentation result, term vector is input to the Named Entity Extraction Model trained
Identification, obtains Entity recognition result vector；The entity in user's question sentence is determined according to Entity recognition result vector.
In one embodiment, it is also performed the steps of when processor executes computer program and segments user's question sentence, obtained
To word segmentation result；Term vector is obtained according to word segmentation result, term vector is input to the intention assessment model identification trained, is obtained
Intention assessment result vector；The intention in user's question sentence is determined according to intention assessment result vector.
In one embodiment, it is also performed the steps of when processor executes computer program to entity and is intended to carry out
Matching, obtains matching result to be determined；Matching vector to be determined is obtained according to matching result to be determined, by matching vector to be determined
It is input in the incidence relation Matching Model trained and is identified, obtain match cognization result vector；According to match cognization knot
Fruit vector determines object matching result.
In one embodiment, it is also performed the steps of when processor executes computer program in default two-dimentional table knowledge
The target entity in matching result is searched in entities field in library, and is looked into the intention field in default bivariate table knowledge base
Look in matching result with the matched target intention of target entity；When finding target entity and target intention, from default two dimension
Target entity and the corresponding answer of target intention are obtained in table knowledge library.
In one embodiment, it is also performed the steps of when processor executes computer program and asks answer return user
The corresponding terminal of sentence, so that the corresponding terminal display answer of user's question sentence.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of when being executed by processor receives user's inquiry instruction, inquires that instruction obtains user and asks according to user
Sentence；It identifies user's question sentence, entity is determined from user's question sentence and is intended to match using the incidence relation Matching Model trained real
Body and intention, obtain matching result；It is searched in default bivariate table knowledge base according to matching result corresponding with user's question sentence
Answer.
In one embodiment, foundation default bivariate table is also performed the steps of when computer program is executed by processor
Knowledge base, presetting in bivariate table knowledge base includes entity, intention and entity and the corresponding answer of intention.
In one embodiment, it also performs the steps of when computer program is executed by processor using the life trained
Name entity recognition model identifies user's question sentence, determines the entity in user's question sentence；Use the intention assessment model pair trained
The identification of user's question sentence, determines the intention in user's question sentence.
In one embodiment, it is also performed the steps of when computer program is executed by processor and carries out user's question sentence
Participle, obtains word segmentation result；Term vector is obtained according to word segmentation result, term vector is input to the name Entity recognition mould trained
Type identification, obtains Entity recognition result vector；The entity in user's question sentence is determined according to Entity recognition result vector.
In one embodiment, it is also performed the steps of when computer program is executed by processor and segments user's question sentence,
Obtain word segmentation result；Term vector is obtained according to word segmentation result, term vector is input to the intention assessment model identification trained, is obtained
To intention assessment result vector；The intention in user's question sentence is determined according to intention assessment result vector.
In one embodiment, also performed the steps of when computer program is executed by processor to entity and be intended into
Row matching, obtains matching result to be determined；Obtain matching vector to be determined according to matching result to be determined, by it is to be determined match to
Amount, which is input in the incidence relation Matching Model trained, to be identified, match cognization result vector is obtained；According to match cognization
Result vector determines object matching result.
In one embodiment, it also performs the steps of when computer program is executed by processor and knows in default bivariate table
Know the target entity searched in matching result in the entities field in library, and in the intention field in default bivariate table knowledge base
Search matching result in the matched target intention of target entity；When finding target entity and target intention, from default two
Target entity and the corresponding answer of target intention are obtained in dimension table knowledge base.
In one embodiment, it is also performed the steps of when computer program is executed by processor and answer is returned into user
The corresponding terminal of question sentence, so that the corresponding terminal display answer of user's question sentence.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (11)
1. a kind of question and answer processing method, which comprises
It receives user and inquires instruction, inquire that instruction obtains user's question sentence according to the user；
It identifies user's question sentence, entity and intention is determined from user's question sentence；
The entity and the intention are matched using the incidence relation Matching Model trained, obtains matching result；
Answer corresponding with user's question sentence is searched in default bivariate table knowledge base according to the matching result.
2. being instructed the method according to claim 1, wherein being inquired in the reception user, inquired according to user
Instruction obtains before user's question sentence, comprising:
The default bivariate table knowledge base is established, includes entity, intention and entity and meaning in the default bivariate table knowledge base
Scheme corresponding answer.
3. the method according to claim 1, wherein described identify user's question sentence, from user's question sentence
Middle determining entity and intention, comprising:
User's question sentence is identified using the Named Entity Extraction Model trained, determines the entity in user's question sentence；
User's question sentence is identified using the intention assessment model trained, determines the intention in user's question sentence.
4. according to the method described in claim 3, it is characterized in that, described use the Named Entity Extraction Model trained to institute
The identification of user's question sentence is stated, determines the entity in user's question sentence, comprising:
User's question sentence is segmented, word segmentation result is obtained；
Term vector is obtained according to the word segmentation result, the term vector is input to the Named Entity Extraction Model trained and is known
Not, Entity recognition result vector is obtained；
The entity in user's question sentence is determined according to the Entity recognition result vector.
5. according to the method described in claim 3, it is characterized in that, described use the intention assessment model trained to the use
The identification of family question sentence, determines the intention in user's question sentence, comprising:
User's question sentence is segmented, word segmentation result is obtained；
Term vector is obtained according to the word segmentation result, the term vector is input to the intention assessment model identification trained, is obtained
To intention assessment result vector；
The intention in user's question sentence is determined according to the intention assessment result vector.
6. according to claim 1 to method described in 5 any one, which is characterized in that described to use the incidence relation trained
Matching Model matches the entity and the intention, obtains matching result, comprising:
The entity and the intention are matched, matching result to be determined is obtained；
Matching vector to be determined is obtained according to the matching result to be determined, the matching vector to be determined is input to and has been trained
Incidence relation Matching Model in identified, obtain match cognization result vector；
Object matching result is determined according to match cognization result vector.
7. according to claim 1 to method described in 5 any one, which is characterized in that it is described according to the matching result pre-
If searching the corresponding answer of user's question sentence in bivariate table knowledge base, comprising:
The target entity in the matching result is searched in the entities field in the default bivariate table knowledge base, and described
Searched in intention field in default bivariate table knowledge base in the matching result with the matched target intention of the target entity；
When finding the target entity and the target intention, the target is obtained from the default bivariate table knowledge base
Entity and the corresponding answer of the target intention.
8. according to claim 1 to method described in 5 any one, which is characterized in that existed described according to the matching result
It presets after searching the corresponding answer of user's question sentence in bivariate table knowledge base, further includes:
The corresponding answer of user's question sentence is returned into terminal, so that answer described in the terminal display.
9. a kind of question and answer processing unit, which is characterized in that described device includes:
Question sentence obtains module, inquires instruction for receiving user, inquires that instruction obtains user's question sentence according to user；
Identification module, user's question sentence obtains the corresponding entity of user's question sentence and intention for identification；
Matching module is matched for using the incidence relation Matching Model trained to match the entity and the intention
As a result；
Searching module is tied for being matched in default entity and intention bivariate table according to the association results according to matching
Fruit obtains the corresponding answer of user's question sentence.
10. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 8 the method when executing the computer program.
11. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any item of the claim 1 to 8 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560681.XA CN110427467B (en) | 2019-06-26 | 2019-06-26 | Question-answer processing method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910560681.XA CN110427467B (en) | 2019-06-26 | 2019-06-26 | Question-answer processing method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110427467A true CN110427467A (en) | 2019-11-08 |
| CN110427467B CN110427467B (en) | 2022-10-11 |
Family
ID=68409532
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910560681.XA Active CN110427467B (en) | 2019-06-26 | 2019-06-26 | Question-answer processing method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110427467B (en) |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110851484A (en) * | 2019-11-13 | 2020-02-28 | 北京香侬慧语科技有限责任公司 | Method and device for obtaining multi-index question answers |
| CN110930989A (en) * | 2019-11-27 | 2020-03-27 | 深圳追一科技有限公司 | Speech intention recognition method and device, computer equipment and storage medium |
| CN111209384A (en) * | 2020-01-08 | 2020-05-29 | 腾讯科技（深圳）有限公司 | Question and answer data processing method and device based on artificial intelligence and electronic equipment |
| CN111309882A (en) * | 2020-02-13 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing intelligent customer service question answering |
| CN111368044A (en) * | 2020-02-21 | 2020-07-03 | 深圳追一科技有限公司 | Intelligent question answering method and device, computer equipment and storage medium |
| CN111414467A (en) * | 2020-03-20 | 2020-07-14 | 中国建设银行股份有限公司 | Question-answer dialogue method and device, electronic equipment and computer readable storage medium |
| CN111444313A (en) * | 2020-03-04 | 2020-07-24 | 深圳追一科技有限公司 | Knowledge graph-based question and answer method and device, computer equipment and storage medium |
| CN111506723A (en) * | 2020-07-01 | 2020-08-07 | 平安国际智慧城市科技股份有限公司 | Question-answer response method, device, equipment and storage medium |
| CN112071300A (en) * | 2020-11-12 | 2020-12-11 | 深圳追一科技有限公司 | Voice conversation method, device, computer equipment and storage medium |
| CN112101035A (en) * | 2020-09-21 | 2020-12-18 | 贝壳技术有限公司 | Named entity identification method and device, readable storage medium and electronic equipment |
| CN112183098A (en) * | 2020-09-30 | 2021-01-05 | 完美世界(北京)软件科技发展有限公司 | Session processing method and device, storage medium and electronic device |
| WO2021174717A1 (en) * | 2020-03-05 | 2021-09-10 | 苏宁易购集团股份有限公司 | Text intent recognition method and apparatus, computer device and storage medium |
| US20210365639A1 (en) * | 2020-05-20 | 2021-11-25 | Beijing Baidu Netcom Science Technology Co., Ltd. | Intent recognition optimization processing method, apparatus, and storage medium |
| CN111858966B (en) * | 2020-08-05 | 2021-12-31 | 龙马智芯（珠海横琴）科技有限公司 | Knowledge graph updating method and device, terminal equipment and readable storage medium |
| CN115982391A (en) * | 2023-03-17 | 2023-04-18 | 恒生电子股份有限公司 | Information processing method and device |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107748757A (en) * | 2017-09-21 | 2018-03-02 | 北京航空航天大学 | A kind of answering method of knowledge based collection of illustrative plates |
| CN107918634A (en) * | 2017-06-27 | 2018-04-17 | 上海壹账通金融科技有限公司 | Intelligent answer method, apparatus and computer-readable recording medium |
| CN107992543A (en) * | 2017-11-27 | 2018-05-04 | 上海智臻智能网络科技股份有限公司 | Question and answer exchange method and device, computer equipment and computer-readable recording medium |
| CN108268580A (en) * | 2017-07-14 | 2018-07-10 | 广东神马搜索科技有限公司 | The answering method and device of knowledge based collection of illustrative plates |
| CN108446286A (en) * | 2017-02-16 | 2018-08-24 | 阿里巴巴集团控股有限公司 | A kind of generation method, device and the server of the answer of natural language question sentence |
| CN108959627A (en) * | 2018-07-23 | 2018-12-07 | 北京光年无限科技有限公司 | Question and answer exchange method and system based on intelligent robot |
| US20190065576A1 (en) * | 2017-08-23 | 2019-02-28 | Rsvp Technologies Inc. | Single-entity-single-relation question answering systems, and methods |
| CN109522393A (en) * | 2018-10-11 | 2019-03-26 | 平安科技（深圳）有限公司 | Intelligent answer method, apparatus, computer equipment and storage medium |
-
2019
- 2019-06-26 CN CN201910560681.XA patent/CN110427467B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108446286A (en) * | 2017-02-16 | 2018-08-24 | 阿里巴巴集团控股有限公司 | A kind of generation method, device and the server of the answer of natural language question sentence |
| CN107918634A (en) * | 2017-06-27 | 2018-04-17 | 上海壹账通金融科技有限公司 | Intelligent answer method, apparatus and computer-readable recording medium |
| CN108268580A (en) * | 2017-07-14 | 2018-07-10 | 广东神马搜索科技有限公司 | The answering method and device of knowledge based collection of illustrative plates |
| US20190065576A1 (en) * | 2017-08-23 | 2019-02-28 | Rsvp Technologies Inc. | Single-entity-single-relation question answering systems, and methods |
| CN107748757A (en) * | 2017-09-21 | 2018-03-02 | 北京航空航天大学 | A kind of answering method of knowledge based collection of illustrative plates |
| CN107992543A (en) * | 2017-11-27 | 2018-05-04 | 上海智臻智能网络科技股份有限公司 | Question and answer exchange method and device, computer equipment and computer-readable recording medium |
| CN108959627A (en) * | 2018-07-23 | 2018-12-07 | 北京光年无限科技有限公司 | Question and answer exchange method and system based on intelligent robot |
| CN109522393A (en) * | 2018-10-11 | 2019-03-26 | 平安科技（深圳）有限公司 | Intelligent answer method, apparatus, computer equipment and storage medium |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110851484A (en) * | 2019-11-13 | 2020-02-28 | 北京香侬慧语科技有限责任公司 | Method and device for obtaining multi-index question answers |
| CN110930989A (en) * | 2019-11-27 | 2020-03-27 | 深圳追一科技有限公司 | Speech intention recognition method and device, computer equipment and storage medium |
| CN110930989B (en) * | 2019-11-27 | 2021-04-06 | 深圳追一科技有限公司 | Speech intention recognition method and device, computer equipment and storage medium |
| CN111209384B (en) * | 2020-01-08 | 2023-08-15 | 腾讯科技（深圳）有限公司 | Question-answer data processing method and device based on artificial intelligence and electronic equipment |
| CN111209384A (en) * | 2020-01-08 | 2020-05-29 | 腾讯科技（深圳）有限公司 | Question and answer data processing method and device based on artificial intelligence and electronic equipment |
| CN111309882A (en) * | 2020-02-13 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing intelligent customer service question answering |
| CN111309882B (en) * | 2020-02-13 | 2023-06-09 | 支付宝(杭州)信息技术有限公司 | Method and device for realizing intelligent customer service question and answer |
| CN111368044A (en) * | 2020-02-21 | 2020-07-03 | 深圳追一科技有限公司 | Intelligent question answering method and device, computer equipment and storage medium |
| CN111444313A (en) * | 2020-03-04 | 2020-07-24 | 深圳追一科技有限公司 | Knowledge graph-based question and answer method and device, computer equipment and storage medium |
| CN111444313B (en) * | 2020-03-04 | 2023-10-24 | 深圳追一科技有限公司 | Knowledge graph-based question and answer method, knowledge graph-based question and answer device, computer equipment and storage medium |
| WO2021174717A1 (en) * | 2020-03-05 | 2021-09-10 | 苏宁易购集团股份有限公司 | Text intent recognition method and apparatus, computer device and storage medium |
| CN111414467A (en) * | 2020-03-20 | 2020-07-14 | 中国建设银行股份有限公司 | Question-answer dialogue method and device, electronic equipment and computer readable storage medium |
| US11972219B2 (en) * | 2020-05-20 | 2024-04-30 | Beijing Baidu Netcom Science Technology Co., Ltd. | Intent recognition optimization processing method, apparatus, and storage medium |
| US20210365639A1 (en) * | 2020-05-20 | 2021-11-25 | Beijing Baidu Netcom Science Technology Co., Ltd. | Intent recognition optimization processing method, apparatus, and storage medium |
| CN111506723A (en) * | 2020-07-01 | 2020-08-07 | 平安国际智慧城市科技股份有限公司 | Question-answer response method, device, equipment and storage medium |
| CN111858966B (en) * | 2020-08-05 | 2021-12-31 | 龙马智芯（珠海横琴）科技有限公司 | Knowledge graph updating method and device, terminal equipment and readable storage medium |
| CN112101035B (en) * | 2020-09-21 | 2022-08-09 | 贝壳技术有限公司 | Named entity identification method and device, readable storage medium and electronic equipment |
| CN112101035A (en) * | 2020-09-21 | 2020-12-18 | 贝壳技术有限公司 | Named entity identification method and device, readable storage medium and electronic equipment |
| CN112183098B (en) * | 2020-09-30 | 2022-05-06 | 完美世界(北京)软件科技发展有限公司 | Session processing method and device, storage medium and electronic device |
| CN112183098A (en) * | 2020-09-30 | 2021-01-05 | 完美世界(北京)软件科技发展有限公司 | Session processing method and device, storage medium and electronic device |
| CN112071300B (en) * | 2020-11-12 | 2021-04-06 | 深圳追一科技有限公司 | Voice conversation method, device, computer equipment and storage medium |
| CN112071300A (en) * | 2020-11-12 | 2020-12-11 | 深圳追一科技有限公司 | Voice conversation method, device, computer equipment and storage medium |
| CN115982391A (en) * | 2023-03-17 | 2023-04-18 | 恒生电子股份有限公司 | Information processing method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110427467B (en) | 2022-10-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110427467A (en) | Question and answer processing method, device, computer equipment and storage medium | |
| CN110765763B (en) | Error correction method and device for voice recognition text, computer equipment and storage medium | |
| CN108563782B (en) | Commodity information format processing method and device, computer equipment and storage medium | |
| CN109522393A (en) | Intelligent answer method, apparatus, computer equipment and storage medium | |
| CN110334179B (en) | Question-answer processing method, device, computer equipment and storage medium | |
| CN111428520B (en) | Text translation method and device | |
| CN111666401B (en) | Document recommendation method, device, computer equipment and medium based on graph structure | |
| CN109670049A (en) | Map path query method, apparatus, computer equipment and storage medium | |
| CN110457431A (en) | Answering method, device, computer equipment and the storage medium of knowledge based map | |
| CN111062215A (en) | Named entity recognition method and device based on semi-supervised learning training | |
| CN110910864A (en) | Training sample selection method and device, computer equipment and storage medium | |
| CN113449489B (en) | Punctuation mark labeling method, punctuation mark labeling device, computer equipment and storage medium | |
| CN109460541A (en) | Lexical relation mask method, device, computer equipment and storage medium | |
| CN113590078A (en) | Virtual image synthesis method and device, computing equipment and storage medium | |
| CN110377618B (en) | Method, device, computer equipment and storage medium for analyzing decision result | |
| CN109033427A (en) | The screening technique and device of stock, computer equipment and readable storage medium storing program for executing | |
| CN111382570B (en) | Text entity recognition method, device, computer equipment and storage medium | |
| WO2021238337A1 (en) | Method and device for entity tagging | |
| CN117521625A (en) | Question answering method, question answering device, electronic equipment and medium | |
| CN108630210B (en) | Speech decoding and recognition method, device, system and machine equipment | |
| CN114861818A (en) | Main data matching method, device, equipment and storage medium based on artificial intelligence | |
| CN110399615B (en) | Transaction risk monitoring method and device | |
| CN113434652A (en) | Intelligent question-answering method, intelligent question-answering device, intelligent question-answering equipment and storage medium | |
| CN109829579A (en) | Minimal path calculation method, device, computer equipment and storage medium | |
| CN110780850A (en) | Requirement case auxiliary generation method and device, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |