CN110502608A - The interactive method and human-computer dialogue device of knowledge based map - Google Patents
The interactive method and human-computer dialogue device of knowledge based map Download PDFInfo
- Publication number
- CN110502608A CN110502608A CN201910606169.4A CN201910606169A CN110502608A CN 110502608 A CN110502608 A CN 110502608A CN 201910606169 A CN201910606169 A CN 201910606169A CN 110502608 A CN110502608 A CN 110502608A
- Authority
- CN
- China
- Prior art keywords
- word
- input
- slot position
- neural networks
- sentence
- 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
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
- G06F16/367—Ontology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Animal Behavior & Ethology (AREA)
- Human Computer Interaction (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Machine Translation (AREA)
Abstract
The embodiment of the present invention provides a kind of interactive method of knowledge based map, comprising: obtains the sentence of user's input；Term vector processing is carried out to the sentence；Processing result is input in convolutional neural networks model；Using the convolutional neural networks model, the intention of the sentence is identified；Slot position prediction is carried out to the word in the sentence；The intention and slot position prediction result are input in knowledge mapping；According to the corresponding relationship between entity in the knowledge mapping, the corresponding relationship of the word is determined；And the search knowledge mapping, and content corresponding with the relationship in the knowledge mapping is exported, to realize human-computer dialogue.Through the embodiment of the present invention, the problem of common sense and Opening field, can be answered, and can actively initiates to talk with.
Description
Technical field
The present embodiments relate to big data field more particularly to a kind of interactive methods of knowledge based map, people
Machine Interface, computer equipment and readable storage medium storing program for executing.
Background technique
For intelligent Answer System in the form of question-response, accurate position puts question to knowledge required for user, by with user
It interacts, personalized information service is provided for user.
Traditional interactive system is based primarily upon sequence to sequence (Sequence to Sequence, seq2seq) mould
Type.However, this class model is trained according to history corpus data to make immediate reaction according to dialogue, but it can not be user
Relevant knowledge answer is provided.This programme is attempted to pass through the combination of knowledge mapping and Task dialog model, realizes human-computer dialogue system
System.
Summary of the invention
In view of this, it is necessary to provide a kind of interactive methods of knowledge based map, human-computer dialogue device, computer
Equipment and computer readable storage medium can answer the problem of common sense and Opening field, and can actively initiate to talk with.
To achieve the above object, described the embodiment of the invention provides a kind of interactive method of knowledge based map
Method includes:
Obtain the sentence of user's input；
Term vector processing is carried out to the sentence；
Processing result is input in convolutional neural networks model；
Using the convolutional neural networks model, the intention of the sentence is identified；
Slot position prediction is carried out to the word in the sentence；
The intention and slot position prediction result are input in knowledge mapping；
According to the corresponding relationship between entity in the knowledge mapping, the corresponding relationship of the word is determined；And
The knowledge mapping is searched for, and content corresponding with the relationship in the knowledge mapping is exported, to realize people
Machine dialogue.
Optionally, it before step processing result being input in convolutional neural networks model, further comprises the steps of:
Two layers of convolutional layer is established, and it is respectively that 1,3 and 5 convolution kernel carries out the extraction of feature that size, which is arranged,；
On the basis of two layers of convolutional layer, pond layer and articulamentum are constructed；And
By normalization index (softmax) function output, to realize intent classifier.
Optionally, described to establish two layers of convolutional layer, and the convolution kernel that size is respectively 1,3 and 5 is set and carries out mentioning for feature
After the step of taking, further comprise the steps of:
On the basis of two layers of convolutional layer, convolutional layer is increased newly, wherein the newly-increased convolutional layer and described two layers volume
Lamination constitutes full convolutional network, and to realize that the slot position to word is predicted, the result of the slot position prediction is the attribute of the word,
To complete the foundation of the convolutional neural networks model.
Optionally, described on the basis of two layers of convolutional layer, increase newly convolutional layer the step of after, further include step
It is rapid:
Obtain corpus data；
Word segmentation processing is carried out to obtain corresponding word to the corpus data；
The word is input in term vector (word2vector) model, and generates corresponding term vector；
Slot position mark is carried out to the word, the mark includes the slot position mark of the word and the intention of the word
Mark；And
Label result is input in the convolutional neural networks model, so that the convolutional neural networks model is according to institute
State intention and the attribute training convolutional neural networks model.
Optionally, described by the intention and before slot position prediction result is input to the step in knowledge mapping, further include
Step:
The user is obtained according to the intention of question and answer data, to the pass between the entity and the entity in the question and answer data
The annotation results of system；And
The knowledge mapping is established according to the entity and the annotation results, to establish the corresponding relationship.
To achieve the above object, the embodiment of the invention also provides a kind of human-computer dialogue devices, comprising:
Module is obtained, for obtaining the sentence of user's input；
Processing module, for carrying out term vector processing to the sentence；
Input module, for processing result to be input in convolutional neural networks model；
Identification module identifies the intention of the sentence for utilizing the convolutional neural networks model；
Prediction module, for carrying out slot position prediction to the word in the sentence；
The input module is also used to for the intention and slot position prediction result being input in knowledge mapping；
Determining module, for determining that the corresponding of the word is closed according to the corresponding relationship between entity in the knowledge mapping
System；And
Session module, for searching for the knowledge mapping, and by content corresponding with the relationship in the knowledge mapping
Output, to realize human-computer dialogue.
Optionally, the human-computer dialogue device, further includes establishing module, for establishing two layers of convolutional layer, and size is arranged
The convolution kernel of respectively 1,3 and 5 carries out the extraction of feature；
On the basis of two layers of convolutional layer, pond layer and articulamentum are constructed；And
It is exported by softmax function, to realize intent classifier.
Optionally, described to establish module, it is also used on the basis of two layers of convolutional layer, increases convolutional layer newly, wherein
The newly-increased convolutional layer and two layers of convolutional layer constitute full convolutional network, to realize that the slot position to word is predicted, the slot position
The result of prediction is the attribute of the word, to complete the foundation of the convolutional neural networks model.
Optionally, the human-computer dialogue device, further includes training module, is used for:
Obtain corpus data；
Word segmentation processing is carried out to obtain corresponding word to the corpus data；
The word is input in word2vector model, and generates corresponding term vector；
Slot position mark is carried out to the word, the mark includes the slot position mark of the word and the intention of the word
Mark；And
Label result is input in the convolutional neural networks model, so that the convolutional neural networks model is according to institute
State intention and the attribute training convolutional neural networks model.
Optionally, described to establish module, it is also used to:
User is obtained according to the intention of question and answer data, to the mark of relationship between the entity and the entity in the question and answer data
Infuse result；And
The corresponding relationship of the entity is established according to the entity and the annotation results.
To achieve the above object, the embodiment of the invention also provides a kind of computer equipment, the computer equipment storages
Device, processor and it is stored in the computer program that can be run on the memory and on the processor, the computer journey
The step of interactive method of knowledge based map as described above is realized when sequence is executed by processor.
To achieve the above object, the embodiment of the invention also provides a kind of computer readable storage medium, the computers
Computer program is stored in readable storage medium storing program for executing, the computer program can be performed by least one processor, so that institute
State the step of at least one processor executes the interactive method of knowledge based map as described above.
Interactive method, the human-computer dialogue device, computer equipment of knowledge based map provided in an embodiment of the present invention
And computer readable storage medium, by establishing convolutional neural networks model and by the corpus data of acquisition to the convolution mind
It is trained through network model, then the sentence by user's input carries out term vector processing, and term vector processing result is input to
To obtain the intention of the sentence and carry out slot position to the word in the sentence in the convolutional neural networks model that training is completed
Prediction will be intended to and slot position prediction result be input in the knowledge mapping of foundation, with the corresponding relationship of the determination word, and will
Content output corresponding with the relationship in the knowledge mapping, to realize human-computer dialogue.The embodiment of the present invention can be to common sense
And the problem of Opening field, is answered, and can actively initiate to talk with.
Detailed description of the invention
Fig. 1 is the step flow chart of the interactive method of the knowledge based map of the embodiment of the present invention one.
Fig. 2 is the hardware structure schematic diagram of the human-computer dialogue device of the embodiment of the present invention two.
Fig. 3 is the program module schematic diagram of the interactive system of the embodiment of the present invention three.
Appended drawing reference:
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.Based on the embodiments of the present invention, those of ordinary skill in the art are not before making creative work
Every other embodiment obtained is put, shall fall within the protection scope of the present invention.
It should be noted that the description for being related to " first ", " second " etc. in the present invention is used for description purposes only, and cannot
It is interpreted as its relative importance of indication or suggestion or implicitly indicates the quantity of indicated technical characteristic.Define as a result, " the
One ", the feature of " second " can explicitly or implicitly include at least one of the features.In addition, the skill between each embodiment
Art scheme can be combined with each other, but must be based on can be realized by those of ordinary skill in the art, when technical solution
Will be understood that the combination of this technical solution is not present in conjunction with there is conflicting or cannot achieve when, also not the present invention claims
Protection scope within.
Embodiment one
Refering to fig. 1, the step flow chart of the interactive method of the knowledge based map of the embodiment of the present invention one is shown.
The sequence for executing step is defined it is appreciated that the flow chart in this method embodiment is not used in.It should be noted that this reality
It is that executing subject carries out exemplary description that example, which is applied, with human-computer dialogue device 2.It is specific as follows:
Step S100 obtains the sentence of user's input.
Step S102 carries out term vector processing to the sentence.
Processing result is input in convolutional neural networks model by step S104.
In specific embodiment, when getting the sentence of user's input, the sentence is subjected to word segmentation processing, then will be divided
Word processing result is input in word2vector model, generates corresponding term vector, and the term vector is then input to convolution
In neural network model.
In a preferred embodiment, it before the sentence for obtaining user's input, also needs to establish the convolutional neural networks mould
Type.Two layers of convolutional layer is initially set up, and it is respectively that 1,3 and 5 convolution kernel carries out the extraction of feature that size, which is arranged,；At described two layers
On the basis of convolutional layer, pond layer and articulamentum are constructed, and export by softmax function, to realize intent classifier.Example
Such as: if the problem of user inputs is " how much is life-insurance insurance premium ", by intent classifier, identifying and be intended to " insurance premium
With inquiry ".
Establishing two layers of convolutional layer, and convolution kernel length be set be be not 1,3 and 5 convolution kernel carry out feature extraction it
Afterwards, further includes: on the basis of two layers of convolutional layer, increase convolutional layer newly, wherein the newly-increased convolutional layer and described two layers
Convolutional layer constitutes full convolutional network, to realize that the slot position to word is predicted, wherein the result of slot position prediction is the word
Attribute.Such as: for example: including " A insurance products " then to be predicted by slot position in the problem of user, result is " insurance products ".
It should be noted that slot filling refers to extracts key message relevant to task from user session.Such as: in face
Into the conversational system of restaurant recommendation, conversational system needs the information such as the place, the price that provide based on user to push away to carry out restaurant
It recommends, the information that the conversational systems such as place, price need is exactly slot information, and slot filling task is exactly to extract these from user session
Slot information.Slot filling is usually modeled as sequence labelling task, and many sequence labelling methods are all used for slot filling.The property of slot filling
There can be very important influence to the quality of conversational system.When obtaining user session information, slot is carried out to the dialog information
Position prediction, to be labeled to the word in the dialog information.The present invention passes through the convolutional neural networks model realization slot
Position identification, recognition rate are higher compared to series model, and can be realized approximation quality on sequence labelling.
After establishing the convolutional neural networks model, corpus data is obtained first, the corpus data is divided
The word is input in word2vector model, and generate corresponding term vector by word processing with obtaining corresponding word；
Slot position mark is carried out to the word, the mark includes the slot position mark of the word and the intention mark of the word；It will
Label result is input in the convolutional neural networks model, so that the convolutional neural networks model is according to the intention and institute
State the attribute training convolutional neural networks model.
It should be noted that corresponding loss function is set after the foundation of the convolutional neural networks model is completed,
Wherein, the loss function is cross entropy, and is trained by ADAM optimization algorithm to the convolutional neural networks model,
In, the foundation and training of the convolutional neural networks model are realized by the library tensorflow in Python.
Step S106 identifies the intention of the sentence using the convolutional neural networks model.
In a preferred embodiment, intentional type is preset in the human-computer dialogue device 2, if the convolutional neural networks
The sentence that model gets user is " expense of A insurance products is how many ", and the human-computer dialogue device 2 passes through described problem
Intent classifier compares described problem and the intention type, and comparing result is described problem and is intended in type
" insurance premium inquiry " is intended to unanimously, then identify that the user's is intended to " insurance premium inquiry ".
Step S108 carries out slot position prediction to the word in the sentence.
In a preferred embodiment, word table corresponding with label is preset in the human-computer dialogue device 2, if the convolution
The sentence of neural network model is " expense of A insurance products is how many ", and " A insurance products " are corresponding with label with the word
Table is matched, and matches " insurance products " label, " A insurance products " are then stamped to the label of " insurance products ", then " A is protected
The slot position prediction result of dangerous product " is " insurance products ".Wherein, the matching includes that part matches and exactly matches.
The intention and slot position prediction result are input in knowledge mapping by step S110.
Optionally, in a preferred embodiment, the intention and slot position prediction result are being input to it in knowledge mapping
Before, user is also obtained according to the intention of question and answer data, to the mark of relationship between the entity and the entity in the question and answer data
As a result, and knowledge mapping is established according to the entity and the annotation results, to establish the corresponding relationship of the entity.For example,
User annotation entity " A insurance products " and entity " 20,000 ", and its corresponding relationship is " expense ", then establishes entity " A insurance products "
Corresponding relationship with entity " 20,000 " is " expense ".
Step S112 determines the corresponding relationship of the word according to the corresponding relationship between entity in the knowledge mapping.
Illustratively, if user is intended to " insurance premium inquiry ", slot position prediction result according to what question and answer data identified
For " insurance products ", the entity of " expense " corresponding relationship of entity " A insurance products " is " 20,000 " in knowledge mapping, it is determined that language
Sentence is that the word in " expense of A insurance products is how many " with " A insurance products " with corresponding relationship " expense " is " 20,000 ".
Step S114 searches for the knowledge mapping, and content corresponding with the relationship in the knowledge mapping is exported,
To realize human-computer dialogue.
Illustratively, when identifying being intended to " insurance premium inquiry " for user, and " insurance products " slot position is " A insurance
When product ", the knowledge mapping is stored with the subgraph structure of " the A insurance products ", the subgraph where " the A insurance products "
Middle corresponding relationship is that the entity of " expense " is " 20,000 ", then by 20,000 outputs to realize human-computer dialogue.
In a further preferred embodiment, if identifying the intention of user and obtaining slot position prediction result, according to knowledge mapping,
It analyzes and also lacks a slot position and just can determine that corresponding as a result, the server sends the problem corresponding with the slot position extremely at this time
User terminal, and obtain the corresponding entity of the slot position from the user terminal, by searching for the knowledge mapping with will be corresponding
Content output.For example, that identifies user's sentence " expense of A insurance products is how many " is intended to " insurance premium inquiry ", " A
Insurance products " slot position prediction result is " insurance products ", due to having " expense " with entity " A insurance products " in knowledge mapping
The entity of corresponding relationship includes " 20,000 " " B insurance company ", then server output " which insurance company, family you need to buy "
With determination " insurance company's slot position ", and according to the answer of user " B insurance company ", " 20,000 " are exported.
Through the embodiment of the present invention, the problem of common sense and Opening field, can be answered, and can actively initiates to talk with.
Embodiment two
Referring to Fig. 2, showing the hardware structure schematic diagram of the human-computer dialogue device of the embodiment of the present invention two.Human-computer dialogue
Device 2 is include but are not limited to, and connection memory 21, processing 22 and network interface 23 can be in communication with each other by system bus,
Fig. 2 illustrates only the human-computer dialogue device 2 with component 21-23, it should be understood that being not required for implementing all show
Component, the implementation that can be substituted is more or less component.
The memory 21 include at least a type of readable storage medium storing program for executing, the readable storage medium storing program for executing include flash memory,
Hard disk, multimedia card, card-type memory (for example, SD or DX memory etc.), random access storage device (RAM), static random are visited
It asks memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), may be programmed read-only deposit
Reservoir (PROM), magnetic storage, disk, CD etc..In some embodiments, the memory 21 can be described man-machine right
Talk about the internal storage unit of device 2, such as the hard disk or memory of the man-machine Interface 2.In further embodiments, described to deposit
Reservoir is also possible to the plug-in type being equipped on the External memory equipment of the human-computer dialogue device 2, such as the man-machine Interface 2
Hard disk, intelligent memory card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card
(Flash Card) etc..Certainly, the memory 21 can also both including the human-computer dialogue device 2 internal storage unit or
Including its External memory equipment.In the present embodiment, the memory 21 is installed on the human-computer dialogue device 2 commonly used in storage
Operating system and types of applications software, such as the program code of interactive system 20 etc..In addition, the memory 21 may be used also
For temporarily storing the Various types of data that has exported or will export.
The processor 22 can be in some embodiments central processing unit (Central Processing Unit,
CPU), controller, microcontroller, microprocessor or other data processing chips.The processor 22 is commonly used in the control people
The overall operation of machine Interface 2.In the present embodiment, the processor 22 is for running the program stored in the memory 21
Code or processing data, such as run the interactive system 20 etc..
The network interface 23 may include radio network interface or wired network interface, which is commonly used in
Communication connection is established between the human-computer dialogue device 2 and other electronic equipments.For example, the network interface 23 is for passing through net
The human-computer dialogue device 2 is connected by network with exterior terminal, establishes number between the human-computer dialogue device 2 and exterior terminal
According to transmission channel and communication connection etc..The network can be intranet (Intranet), internet (Internet), complete
Ball mobile communcations system (Global System of Mobile communication, GSM), wideband code division multiple access
(Wideband Code Division Multiple Access, WCDMA), 4G network, 5G network, bluetooth (Bluetooth),
The wirelessly or non-wirelessly network such as Wi-Fi.
Embodiment three
Referring to Fig. 3, showing the program module schematic diagram of the interactive system of the embodiment of the present invention three.In this implementation
In example, interactive system 20 may include or be divided into one or more program modules, one or more program module
It is stored in storage medium, and as performed by one or more processors, to complete the present invention, and can realize above-mentioned based on knowing
Know the interactive method of map.The so-called program module of the embodiment of the present invention is a series of meters for referring to complete specific function
Calculation machine program instruction section, the implementation procedure than program itself more suitable for description interactive system 20 in storage medium.With
The function of each program module of the present embodiment will be specifically introduced in lower description:
Module 201 is obtained, for obtaining the sentence of user's input.
Processing module 202, for carrying out term vector processing to the sentence.
Input module 203, for processing result to be input in convolutional neural networks model.
In specific embodiment, when the acquisition module 201 gets the sentence of user's input, the processing module 202
The sentence is subjected to word segmentation processing, then word segmentation processing result is input to word2vector model by the input module 203
In, corresponding term vector is generated, the term vector is input in convolutional neural networks model by the last input module 203.
In a preferred embodiment, before the acquisition module 201 obtains the sentence of user's input, module 208 is established
It also needs to establish the convolutional neural networks model.Two layers of convolutional layer is initially set up, and the convolution that size is respectively 1,3 and 5 is set
Core carries out the extraction of feature, then, on the basis of two layers of convolutional layer, constructs pond layer and articulamentum, and pass through
Softmax function output, to realize intent classifier.Such as: if the problem of user inputs is " life-insurance insurance premium is how much ",
By intent classifier, identifies and be intended to " insurance premium inquiry ".
The module 208 of establishing establishes two layers of convolutional layer, and convolution kernel length be set be be not 1,3 and 5 convolution kernel into
After the extraction of row feature, the module 208 of establishing increases convolutional layer newly, wherein institute on the basis of two layers of convolutional layer
It states newly-increased convolutional layer and two layers of convolutional layer constitutes full convolutional network, to realize that the slot position to word is predicted, wherein the slot
The result of position prediction is the attribute of the word.Such as: include in the problem of user " A insurance products ", then by slot position prediction,
As a result it is " insurance products ".
It should be noted that slot filling refers to extracts key message relevant to task from user session.Such as: in face
Into the conversational system of restaurant recommendation, conversational system needs the information such as the place, the price that provide based on user to push away to carry out restaurant
It recommends, the information that the conversational systems such as place, price need is exactly slot information, and slot filling task is exactly to extract these from user session
Slot information.Slot filling is usually modeled as sequence labelling task, and many sequence labelling methods are all used for slot filling.The property of slot filling
There can be very important influence to the quality of conversational system.When obtaining user session information, slot is carried out to the dialog information
Position prediction, to be labeled to the word in the dialog information.The present invention passes through the convolutional neural networks model realization slot
Position identification, recognition rate are higher compared to series model, and can be realized approximation quality on sequence labelling.
It is established after module 208 establishes the convolutional neural networks model described, training module 209 is to the convolution mind
It is trained through network model.Firstly, obtaining corpus data, word segmentation processing is carried out to obtain corresponding word to the corpus data
The word is input in word2vector model by language, and generates corresponding term vector.Then, slot is carried out to the word
Position mark, the mark include the slot position mark of the word and the intention mark of the word.Finally, will label result input
Extremely in the convolutional neural networks model, so that the convolutional neural networks model is according to the intention and attribute training institute
State convolutional neural networks model.
It should be noted that corresponding loss function is set after the foundation of the convolutional neural networks model is completed,
Wherein, the loss function is cross entropy, and is trained by ADAM optimization algorithm to the convolutional neural networks model,
In, the foundation and training of the convolutional neural networks model are realized by the library tensorflow in Python.
Identification module 204 identifies the intention of the sentence for utilizing the convolutional neural networks model.
In a preferred embodiment, intentional type is preset in the human-computer dialogue device 2, if the convolutional neural networks
The sentence that model gets user is " expense of A insurance products is how many ", and the identification module 204 is by described problem by anticipating
Figure classification, described problem and the intention type are compared, and comparing result is " protecting in described problem and intention type
Dangerous inquiry into expenses " is intended to unanimously, then identify that the user's is intended to " insurance premium inquiry ".
Prediction module 205, for carrying out slot position prediction to the word in the sentence.
In a preferred embodiment, word table corresponding with label is preset in the human-computer dialogue device 2, if the convolution
The sentence of neural network model is " expense of A insurance products is how many ", and the prediction module 205 is by " A insurance products " and institute
Predicate language table corresponding with label is matched, and " insurance products " label is matched, and " A insurance products " are then stamped to " insurance production
The label of product ", then the slot position prediction result of " A insurance products " is " insurance products ".Wherein, it is described matching include part matching and
Exact matching.
The input module 203 is also used to for the intention and slot position prediction result being input in knowledge mapping.
Optionally, in a preferred embodiment, the intention and slot position prediction result are inputted in the input module 203
Before into knowledge mapping, the module 208 of establishing also obtains user according to the intention of question and answer data, in the question and answer data
Entity and the entity between relationship annotation results, and knowledge mapping is established according to the entity and the annotation results, with
Establish the corresponding relationship of the entity.For example, user annotation entity " A insurance products " and entity " 20,000 ", and its corresponding relationship is
" expense ", then the module 208 of establishing establishes the corresponding relationship of entity " A insurance products " and entity " 20,000 " as " expense ".
Determining module 206, for determining the correspondence of the word according to the corresponding relationship between entity in the knowledge mapping
Relationship.
Illustratively, if user is intended to " insurance premium inquiry ", slot position prediction result according to what question and answer data identified
For " insurance products ", the entity of " expense " corresponding relationship of entity " A insurance products " is " 20,000 " in knowledge mapping, it is determined that mould
Block 206 determines that sentence is the word in " expense of A insurance products is how many " with " A insurance products " with corresponding relationship " expense "
For " 20,000 ".
Session module 207, for searching for the knowledge mapping, and will be corresponding with the relationship interior in the knowledge mapping
Hold output, to realize human-computer dialogue.
Illustratively, when the identification module 204 identifies being intended to " insurance premium inquiry " for user, and " insurance production
When product " slot position is " A insurance products ", the knowledge mapping is stored with the subgraph structure of " the A insurance products ", " the A insurance
It is " 20,000 " that corresponding relationship, which is the entity of " expense ", in subgraph where product ", then the session module 207 is by 20,000 outputs with reality
Existing human-computer dialogue.
In a further preferred embodiment, if identifying the intention of user and obtaining slot position prediction result, according to knowledge mapping,
Analyze also lack a slot position just can determine that it is corresponding as a result, at this time the interactive system 20 send it is corresponding with the slot position
The problem of to user terminal, and obtain the corresponding entity of the slot position from the user terminal, by searching for the knowledge mapping with
Corresponding content is exported.For example, that identifies user's sentence " expense of A insurance products is how many " is intended to " insurance premium
Inquiry ", " A insurance products " slot position prediction result is " insurance products ", due to having in knowledge mapping with entity " A insurance products "
The entity of the corresponding relationship of " expense " includes " 20,000 " " B insurance company ", then " which house insurance you need to buy for server output
Company " with determination " insurance company's slot position ", and according to the answer of user " B insurance company ", " 20,000 " are exported.
Through the embodiment of the present invention, the problem of common sense and Opening field, can be answered, and can actively initiates to talk with.
The present invention also provides a kind of computer equipments, can such as execute smart phone, tablet computer, the notebook electricity of program
Brain, desktop computer, rack-mount server, blade server, tower server or Cabinet-type server (including independent clothes
Server cluster composed by business device or multiple servers) etc..The computer equipment of the present embodiment includes at least but unlimited
In: memory, the processor etc. of connection can be in communication with each other by system bus.
The present embodiment also provides a kind of computer readable storage medium, such as flash memory, hard disk, multimedia card, card-type memory
(for example, SD or DX memory etc.), random access storage device (RAM), static random-access memory (SRAM), read-only memory
(ROM), electrically erasable programmable read-only memory (EEPROM), programmable read only memory (PROM), magnetic storage, magnetic
Disk, CD, server, App are stored thereon with computer program, phase are realized when program is executed by processor using store etc.
Answer function.The computer readable storage medium of the present embodiment is for storing interactive system 20, realization when being executed by processor
The interactive method of the knowledge based map of embodiment one.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (10)
1. a kind of interactive method of knowledge based map characterized by comprising
Obtain the sentence of user's input；
Term vector processing is carried out to the sentence；
Processing result is input in convolutional neural networks model；
Using the convolutional neural networks model, the intention of the sentence is identified；
Slot position prediction is carried out to the word in the sentence；
The intention and slot position prediction result are input in knowledge mapping；
According to the corresponding relationship between entity in the knowledge mapping, the corresponding relationship of the word is determined；And
The knowledge mapping is searched for, and content corresponding with the relationship in the knowledge mapping is exported, it is man-machine right to realize
Words.
2. the interactive method of knowledge based map as described in claim 1, which is characterized in that described that processing result is defeated
Before entering the step into convolutional neural networks model, further comprise the steps of:
Two layers of convolutional layer is established, and it is respectively that 1,3 and 5 convolution kernel carries out the extraction of feature that size, which is arranged,；
On the basis of two layers of convolutional layer, pond layer and articulamentum are constructed；And
It is exported by softmax function, to realize intent classifier.
3. the interactive method of knowledge based map as claimed in claim 2, which is characterized in that described to establish two layers of convolution
Layer, and after the step of extraction for the convolution kernel progress feature that size is respectively 1,3 and 5 is set, it further comprises the steps of:
On the basis of two layers of convolutional layer, convolutional layer is increased newly, wherein the newly-increased convolutional layer and two layers of convolutional layer
Full convolutional network is constituted, to realize that the slot position to word is predicted, the result of the slot position prediction is the attribute of the word, with complete
At the foundation of the convolutional neural networks model.
4. the interactive method of knowledge based map as claimed in claim 3, which is characterized in that described at described two layers volume
On the basis of lamination, increase newly convolutional layer the step of after, further comprise the steps of:
Obtain corpus data；
Word segmentation processing is carried out to obtain corresponding word to the corpus data；
The word is input in word2vector model, and generates corresponding term vector；
Slot position mark is carried out to the word, the mark includes the slot position mark of the word and the intention mark of the word
Note；And
Label result is input in the convolutional neural networks model, so that the convolutional neural networks model is according to the meaning
Figure and the attribute training convolutional neural networks model.
5. the interactive method of knowledge based map as described in claim 1, which is characterized in that it is described by it is described intention and
Slot position prediction result is input to before the step in knowledge mapping, is further comprised the steps of:
The user is obtained according to the intention of question and answer data, to the relationship between the entity and the entity in the question and answer data
Annotation results；And
The knowledge mapping is established according to the entity and the annotation results, to establish the corresponding relationship.
6. a kind of human-computer dialogue device characterized by comprising
Module is obtained, for obtaining the sentence of user's input；
Processing module, for carrying out term vector processing to the sentence；
Input module, for processing result to be input in convolutional neural networks model；
Identification module identifies the intention of the sentence for utilizing the convolutional neural networks model；
Prediction module, for carrying out slot position prediction to the word in the sentence；
The input module is also used to for the intention and slot position prediction result being input in knowledge mapping；
Determining module, for determining the corresponding relationship of the word according to the corresponding relationship between entity in the knowledge mapping；And
Session module is exported for searching for the knowledge mapping, and by content corresponding with the relationship in the knowledge mapping,
To realize human-computer dialogue.
7. human-computer dialogue device as claimed in claim 6, which is characterized in that further include establishing module, be used for:
Two layers of convolutional layer is established, and it is respectively that 1,3 and 5 convolution kernel carries out the extraction of feature that size, which is arranged,；
On the basis of two layers of convolutional layer, pond layer and articulamentum are constructed；And
It is exported by softmax function, to realize intent classifier.
8. human-computer dialogue device as claimed in claim 7, which is characterized in that the module of establishing is also used to:
On the basis of two layers of convolutional layer, convolutional layer is increased newly, wherein the newly-increased convolutional layer and two layers of convolutional layer
Full convolutional network is constituted, to realize that the slot position to word is predicted, the result of the slot position prediction is the attribute of the word, with complete
At the foundation of the convolutional neural networks model.
9. a kind of computer equipment, which is characterized in that the computer equipment memory, processor and be stored in the memory
Computer program that is upper and can running on the processor, is realized when the computer program is executed by processor as right is wanted
The step of seeking the interactive method of knowledge based map described in any one of 1-5.
10. a kind of computer readable storage medium, which is characterized in that be stored with computer in the computer readable storage medium
Program, the computer program can be performed by least one processors, so that at least one described processor executes such as right
It is required that described in any one of 1-5 the step of the interactive method of knowledge based map.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910606169.4A CN110502608B (en) | 2019-07-05 | 2019-07-05 | Man-machine conversation method and man-machine conversation device based on knowledge graph |
| PCT/CN2019/103243 WO2021003819A1 (en) | 2019-07-05 | 2019-08-29 | Man-machine dialog method and man-machine dialog apparatus based on knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910606169.4A CN110502608B (en) | 2019-07-05 | 2019-07-05 | Man-machine conversation method and man-machine conversation device based on knowledge graph |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110502608A true CN110502608A (en) | 2019-11-26 |
| CN110502608B CN110502608B (en) | 2023-10-10 |
Family
ID=68585963
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910606169.4A Active CN110502608B (en) | 2019-07-05 | 2019-07-05 | Man-machine conversation method and man-machine conversation device based on knowledge graph |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110502608B (en) |
| WO (1) | WO2021003819A1 (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111026842A (en) * | 2019-11-29 | 2020-04-17 | 微民保险代理有限公司 | Natural language processing method, natural language processing device and intelligent question-answering system |
| CN111125309A (en) * | 2019-12-23 | 2020-05-08 | 中电云脑(天津)科技有限公司 | Natural language processing method and device, computing equipment and storage medium |
| CN111143561A (en) * | 2019-12-26 | 2020-05-12 | 北京百度网讯科技有限公司 | Intention recognition model training method and device and electronic equipment |
| CN111191000A (en) * | 2019-12-20 | 2020-05-22 | 北京淇瑀信息科技有限公司 | Dialog management method, device and system of intelligent voice robot |
| CN111341308A (en) * | 2020-02-12 | 2020-06-26 | 百度在线网络技术（北京）有限公司 | Method and apparatus for outputting information |
| CN111339770A (en) * | 2020-02-18 | 2020-06-26 | 百度在线网络技术（北京）有限公司 | Method and apparatus for outputting information |
| CN111460124A (en) * | 2020-04-21 | 2020-07-28 | 文思海辉智科科技有限公司 | Intelligent interaction method and device and robot |
| CN111651988A (en) * | 2020-06-03 | 2020-09-11 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for training a model |
| CN111737990A (en) * | 2020-06-24 | 2020-10-02 | 深圳前海微众银行股份有限公司 | Word slot filling method, device, equipment and storage medium |
| CN111738016A (en) * | 2020-06-28 | 2020-10-02 | 中国平安财产保险股份有限公司 | Multi-intention recognition method and related equipment |
| CN111737432A (en) * | 2020-06-22 | 2020-10-02 | 四川长虹电器股份有限公司 | Automatic dialogue method and system based on joint training model |
| CN111949787A (en) * | 2020-08-21 | 2020-11-17 | 平安国际智慧城市科技股份有限公司 | Automatic question-answering method, device, equipment and storage medium based on knowledge graph |
| CN112328808A (en) * | 2020-11-03 | 2021-02-05 | 四川长虹电器股份有限公司 | Knowledge graph-based question and answer method and device, electronic equipment and storage medium |
| CN113157939A (en) * | 2021-03-26 | 2021-07-23 | 联想(北京)有限公司 | Information processing method and device |
| CN113326367A (en) * | 2021-06-30 | 2021-08-31 | 四川启睿克科技有限公司 | Task type dialogue method and system based on end-to-end text generation |
| CN113742480A (en) * | 2020-06-18 | 2021-12-03 | 北京汇钧科技有限公司 | Customer service response method and device |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801492B (en) * | 2021-01-22 | 2023-07-25 | 中国平安人寿保险股份有限公司 | Knowledge-hierarchy-based data quality inspection method and device and computer equipment |
| CN112905884B (en) * | 2021-02-10 | 2024-05-31 | 北京百度网讯科技有限公司 | Method, apparatus, medium and program product for generating sequence annotation model |
| CN113010780B (en) * | 2021-03-11 | 2022-07-15 | 北京三快在线科技有限公司 | Model training and click rate estimation method and device |
| CN112989046B (en) * | 2021-03-18 | 2023-07-21 | 中国平安人寿保险股份有限公司 | Real-time speech prejudging method, device, computer equipment and storage medium |
| CN112965603A (en) * | 2021-03-26 | 2021-06-15 | 南京阿凡达机器人科技有限公司 | Method and system for realizing man-machine interaction |
| CN113010660A (en) * | 2021-04-22 | 2021-06-22 | 国网信息通信产业集团有限公司 | Intelligent question and answer method and device based on knowledge graph |
| CN113190668A (en) * | 2021-05-19 | 2021-07-30 | 平安国际智慧城市科技股份有限公司 | Man-machine interaction method, device and equipment based on multi-turn conversation and storage medium |
| CN113515611B (en) * | 2021-06-22 | 2022-04-26 | 镁佳(北京)科技有限公司 | Intention recognition method and recognition system for task type multi-intention conversation |
| CN113468307B (en) * | 2021-06-30 | 2023-06-30 | 网易（杭州）网络有限公司 | Text processing method, device, electronic equipment and storage medium |
| CN113641807B (en) * | 2021-07-28 | 2024-05-24 | 北京百度网讯科技有限公司 | Training method, device, equipment and storage medium of dialogue recommendation model |
| CN113609827B (en) * | 2021-08-09 | 2023-05-26 | 海南大学 | Content processing method and system based on intent-driven DIKW |
| CN113836320B (en) * | 2021-11-26 | 2022-03-25 | 北京世纪好未来教育科技有限公司 | Exercise recommendation method and device, storage medium and electronic equipment |
| CN116774569B (en) * | 2023-07-25 | 2024-04-05 | 博纯材料股份有限公司 | Artificial intelligence-based method and system for updating operation system of oxygen-argon separation equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180025008A1 (en) * | 2016-07-20 | 2018-01-25 | Baidu Usa Llc | Systems and methods for homogeneous entity grouping |
| WO2018149326A1 (en) * | 2017-02-16 | 2018-08-23 | 阿里巴巴集团控股有限公司 | Natural language question answering method and apparatus, and server |
| CN109033305A (en) * | 2018-07-16 | 2018-12-18 | 深圳前海微众银行股份有限公司 | Question answering method, equipment and computer readable storage medium |
| CN109063035A (en) * | 2018-07-16 | 2018-12-21 | 哈尔滨工业大学 | A kind of man-machine more wheel dialogue methods towards trip field |
| CN109063221A (en) * | 2018-11-02 | 2018-12-21 | 北京百度网讯科技有限公司 | Query intention recognition methods and device based on mixed strategy |
| CN109241524A (en) * | 2018-08-13 | 2019-01-18 | 腾讯科技（深圳）有限公司 | Semantic analysis method and device, computer readable storage medium, electronic equipment |
| WO2019024162A1 (en) * | 2017-08-04 | 2019-02-07 | 平安科技（深圳）有限公司 | Intention obtaining method, electronic device, and computer-readable storage medium |
| CN109947916A (en) * | 2019-03-01 | 2019-06-28 | 河北尚云信息科技有限公司 | Question answering system device and answering method based on meteorological field knowledge mapping |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10586156B2 (en) * | 2015-06-25 | 2020-03-10 | International Business Machines Corporation | Knowledge canvassing using a knowledge graph and a question and answer system |
| CN109635117B (en) * | 2018-12-26 | 2021-05-14 | 零犀(北京)科技有限公司 | Method and device for recognizing user intention based on knowledge graph |
-
2019
- 2019-07-05 CN CN201910606169.4A patent/CN110502608B/en active Active
- 2019-08-29 WO PCT/CN2019/103243 patent/WO2021003819A1/en active Application Filing
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180025008A1 (en) * | 2016-07-20 | 2018-01-25 | Baidu Usa Llc | Systems and methods for homogeneous entity grouping |
| WO2018149326A1 (en) * | 2017-02-16 | 2018-08-23 | 阿里巴巴集团控股有限公司 | Natural language question answering method and apparatus, and server |
| WO2019024162A1 (en) * | 2017-08-04 | 2019-02-07 | 平安科技（深圳）有限公司 | Intention obtaining method, electronic device, and computer-readable storage medium |
| CN109033305A (en) * | 2018-07-16 | 2018-12-18 | 深圳前海微众银行股份有限公司 | Question answering method, equipment and computer readable storage medium |
| CN109063035A (en) * | 2018-07-16 | 2018-12-21 | 哈尔滨工业大学 | A kind of man-machine more wheel dialogue methods towards trip field |
| CN109241524A (en) * | 2018-08-13 | 2019-01-18 | 腾讯科技（深圳）有限公司 | Semantic analysis method and device, computer readable storage medium, electronic equipment |
| CN109063221A (en) * | 2018-11-02 | 2018-12-21 | 北京百度网讯科技有限公司 | Query intention recognition methods and device based on mixed strategy |
| CN109947916A (en) * | 2019-03-01 | 2019-06-28 | 河北尚云信息科技有限公司 | Question answering system device and answering method based on meteorological field knowledge mapping |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111026842B (en) * | 2019-11-29 | 2023-09-12 | 微民保险代理有限公司 | Natural language processing method, natural language processing device and intelligent question-answering system |
| CN111026842A (en) * | 2019-11-29 | 2020-04-17 | 微民保险代理有限公司 | Natural language processing method, natural language processing device and intelligent question-answering system |
| CN111191000A (en) * | 2019-12-20 | 2020-05-22 | 北京淇瑀信息科技有限公司 | Dialog management method, device and system of intelligent voice robot |
| CN111191000B (en) * | 2019-12-20 | 2023-08-18 | 北京淇瑀信息科技有限公司 | Dialogue management method, device and system of intelligent voice robot |
| CN111125309A (en) * | 2019-12-23 | 2020-05-08 | 中电云脑(天津)科技有限公司 | Natural language processing method and device, computing equipment and storage medium |
| CN111143561A (en) * | 2019-12-26 | 2020-05-12 | 北京百度网讯科技有限公司 | Intention recognition model training method and device and electronic equipment |
| CN111143561B (en) * | 2019-12-26 | 2023-04-07 | 北京百度网讯科技有限公司 | Intention recognition model training method and device and electronic equipment |
| CN111341308A (en) * | 2020-02-12 | 2020-06-26 | 百度在线网络技术（北京）有限公司 | Method and apparatus for outputting information |
| CN111341308B (en) * | 2020-02-12 | 2023-07-21 | 百度在线网络技术（北京）有限公司 | Method and device for outputting information |
| CN111339770B (en) * | 2020-02-18 | 2023-07-21 | 百度在线网络技术（北京）有限公司 | Method and device for outputting information |
| CN111339770A (en) * | 2020-02-18 | 2020-06-26 | 百度在线网络技术（北京）有限公司 | Method and apparatus for outputting information |
| CN111460124A (en) * | 2020-04-21 | 2020-07-28 | 文思海辉智科科技有限公司 | Intelligent interaction method and device and robot |
| CN111651988A (en) * | 2020-06-03 | 2020-09-11 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for training a model |
| CN111651988B (en) * | 2020-06-03 | 2023-05-19 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for training model |
| CN113742480A (en) * | 2020-06-18 | 2021-12-03 | 北京汇钧科技有限公司 | Customer service response method and device |
| CN111737432A (en) * | 2020-06-22 | 2020-10-02 | 四川长虹电器股份有限公司 | Automatic dialogue method and system based on joint training model |
| CN111737990B (en) * | 2020-06-24 | 2023-05-23 | 深圳前海微众银行股份有限公司 | Word slot filling method, device, equipment and storage medium |
| CN111737990A (en) * | 2020-06-24 | 2020-10-02 | 深圳前海微众银行股份有限公司 | Word slot filling method, device, equipment and storage medium |
| CN111738016A (en) * | 2020-06-28 | 2020-10-02 | 中国平安财产保险股份有限公司 | Multi-intention recognition method and related equipment |
| CN111738016B (en) * | 2020-06-28 | 2023-09-05 | 中国平安财产保险股份有限公司 | Multi-intention recognition method and related equipment |
| CN111949787A (en) * | 2020-08-21 | 2020-11-17 | 平安国际智慧城市科技股份有限公司 | Automatic question-answering method, device, equipment and storage medium based on knowledge graph |
| CN112328808A (en) * | 2020-11-03 | 2021-02-05 | 四川长虹电器股份有限公司 | Knowledge graph-based question and answer method and device, electronic equipment and storage medium |
| CN113157939A (en) * | 2021-03-26 | 2021-07-23 | 联想(北京)有限公司 | Information processing method and device |
| CN113326367A (en) * | 2021-06-30 | 2021-08-31 | 四川启睿克科技有限公司 | Task type dialogue method and system based on end-to-end text generation |
| CN113326367B (en) * | 2021-06-30 | 2023-06-16 | 四川启睿克科技有限公司 | Task type dialogue method and system based on end-to-end text generation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110502608B (en) | 2023-10-10 |
| WO2021003819A1 (en) | 2021-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110502608A (en) | The interactive method and human-computer dialogue device of knowledge based map | |
| CN107688614B (en) | It is intended to acquisition methods, electronic device and computer readable storage medium | |
| CN109634698B (en) | Menu display method and device, computer equipment and storage medium | |
| US20190108242A1 (en) | Search method and processing device | |
| US10438133B2 (en) | Spend data enrichment and classification | |
| CN110322349B (en) | Data processing method, device and equipment | |
| CN107832338B (en) | Method and system for recognizing core product words | |
| US10956469B2 (en) | System and method for metadata correlation using natural language processing | |
| CN110086658A (en) | Interface switching method, device and computer readable storage medium | |
| CN109815381A (en) | User's portrait construction method, system, computer equipment and storage medium | |
| CN109902250A (en) | Sharing method, sharing means, computer equipment and the storage medium of questionnaire survey | |
| CN111538794A (en) | Data fusion method, device and equipment | |
| CN114780746A (en) | Knowledge graph-based document retrieval method and related equipment thereof | |
| CN112132238A (en) | Method, device, equipment and readable medium for identifying private data | |
| CN111259167B (en) | User request risk identification method and device | |
| CN109446515A (en) | Group information analysis method, electronic device and computer readable storage medium | |
| CN106934006B (en) | Page recommendation method and device based on multi-branch tree model | |
| CN110532229B (en) | Evidence file retrieval method, device, computer equipment and storage medium | |
| CN111782946A (en) | Book friend recommendation method, calculation device and computer storage medium | |
| CN113569118B (en) | Self-media pushing method, device, computer equipment and storage medium | |
| CN114461761A (en) | Searching method, system, computer device and storage medium based on label matching | |
| CN112363814A (en) | Task scheduling method and device, computer equipment and storage medium | |
| CN112559866A (en) | College book reading recommendation method, device, equipment and storage medium | |
| CN115952800A (en) | Named entity recognition method and device, computer equipment and readable storage medium | |
| CN109710874A (en) | Processing method and processing device, storage medium, the computer equipment of page data |
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 |