CN108197105A - Natural language processing method, apparatus, storage medium and electronic equipment - Google Patents
Natural language processing method, apparatus, storage medium and electronic equipment Download PDFInfo
- Publication number
- CN108197105A CN108197105A CN201711466311.7A CN201711466311A CN108197105A CN 108197105 A CN108197105 A CN 108197105A CN 201711466311 A CN201711466311 A CN 201711466311A CN 108197105 A CN108197105 A CN 108197105A
- Authority
- CN
- China
- Prior art keywords
- information
- natural language
- semantic
- semantic parsing
- parsing
- 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
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
This application discloses a kind of natural language processing method, apparatus, storage medium and electronic equipments.This method includes:Obtain natural language information input by user and relevant information associated with the natural language information;Multiple semantic parsing information are obtained according to the natural language information;By the multiple semantic parsing information and the relevant information input prediction model, the prediction model obtains multiple probability values of corresponding the multiple semantic parsing information according to the relevant information;Determine that the maximum semantic parsing information of probability value parses information for target semanteme from the multiple probability value.Parsing identification is no longer carried out in itself only for natural language information, in prediction model, natural language information is assisted in identifying by relevant information, improves the accuracy to natural language information parsing identification.
Description
Technical field
The application belongs to a kind of field of communication technology more particularly to natural language processing method, apparatus, storage medium and electricity
Sub- equipment.
Background technology
Man-machine interaction mode between people and electronic equipment mainly include mouse click Menu, input through keyboard order,
The modes such as touch screen control.But these man-machine interaction modes are required for user to carry out specific control operation, such as click specific
Icon, the specific order of input, interaction are inconvenient.
With the development of artificial intelligence technology, human-computer interaction is carried out by the natural language of user, it can be easily and fast
Human-computer interaction is carried out, is not limited by specifically ordering or icon could carry out human-computer interaction.Natural language can allow user side
Just the wish of oneself is fast and accurately expressed, can it is most heavy to be increasingly becoming with the intention of true representation user for intelligent Service field
The man-machine interaction mode wanted.
But because of the characteristics such as opening, randomness of natural language, cause to carry out natural language semantic parsing, identification
During its real meaning, it be easy to cause ambiguity.
Invention content
The application provides a kind of natural language processing method, apparatus, storage medium and electronic equipment, can be promoted to nature
The recognition accuracy of language.
In a first aspect, the embodiment of the present application provides a kind of natural language processing method, applied to electronic equipment, the method
Including:
Obtain natural language information input by user and relevant information associated with the natural language information;
Multiple semantic parsing information are obtained according to the natural language information;
By the multiple semantic parsing information and the relevant information input prediction model, the prediction model is according to
Relevant information obtains multiple probability values of corresponding the multiple semantic parsing information;
Determine that the maximum semantic parsing information of probability value parses information for target semanteme from the multiple probability value.
Second aspect, the embodiment of the present application provides a kind of natural language processing device, applied to electronic equipment, described device
Including:
First acquisition module, for obtain natural language information input by user and with the natural language information phase
Associated relevant information;
Second acquisition module, for obtaining multiple semantic parsing information according to the natural language information;
Probability value acquisition module, for the multiple semanteme to be parsed information and the relevant information input prediction model,
The prediction model obtains multiple probability values of corresponding the multiple semantic parsing information according to the relevant information;
Determining module, for determining that the maximum semantic parsing information of probability value is semantic for target from the multiple probability value
Parse information.
The third aspect, the embodiment of the present application provide a kind of storage medium, computer program are stored thereon with, when the calculating
When machine program is run on computers so that the computer performs above-mentioned natural language processing method.
Fourth aspect, the embodiment of the present application provide a kind of electronic equipment, and including processor and memory, the memory has
Computer program, the processor is by calling the computer program, for performing above-mentioned natural language processing method.
Natural language processing method, apparatus provided by the embodiments of the present application, storage medium and electronic equipment pass through to obtain and use
The natural language information of family input and relevant information associated with natural language information;It is obtained according to natural language information
Multiple semantic parsing information;Multiple semantic parsing information and relevant information input prediction model, prediction model are believed according to correlation
Breath obtains multiple probability values of corresponding multiple semantic parsing information;The maximum semantic parsing of probability value is determined from multiple probability values
Information parses information for target semanteme.No longer parsing identification is carried out in itself only for natural language information, in prediction model,
Natural language information is assisted in identifying by relevant information, improves the accuracy to natural language information parsing identification,
The final goal semanteme parsing information made is more in line with the true idea of user.
Description of the drawings
In order to illustrate more clearly of the technical solution in the embodiment of the present application, make required in being described below to embodiment
Attached drawing is briefly described.It should be evident that the accompanying drawings in the following description is only some embodiments of the present application, for
For those skilled in the art, without creative efforts, it can also be obtained according to these attached drawings other attached
Figure.
Fig. 1 is the application scenarios schematic diagram of natural language processing device provided by the embodiments of the present application.
Fig. 2 is the first flow diagram of natural language processing method provided by the embodiments of the present application.
Fig. 3 is second of flow diagram of natural language processing method provided by the embodiments of the present application.
Fig. 4 is the third flow diagram of natural language processing method provided by the embodiments of the present application.
Fig. 5 is the 4th kind of flow diagram of natural language processing method provided by the embodiments of the present application.
Fig. 6 is the 5th kind of flow diagram of natural language processing method provided by the embodiments of the present application.
Fig. 7 is the first structure diagram of natural language processing device provided by the embodiments of the present application.
Fig. 8 is second of structure diagram of natural language processing device provided by the embodiments of the present application.
Fig. 9 is the third structure diagram of natural language processing device provided by the embodiments of the present application.
Figure 10 is the 4th kind of structure diagram of natural language processing device provided by the embodiments of the present application.
Figure 11 is the structure diagram of electronic equipment provided by the embodiments of the present application.
Figure 12 is another structure diagram of electronic equipment provided by the embodiments of the present application.
Specific embodiment
Schema is please referred to, wherein identical element numbers represent identical component, the principle of the application is to implement one
It is illustrated in appropriate computing environment.The following description be based on illustrated the application specific embodiment, should not be by
It is considered as limitation the application other specific embodiments not detailed herein.
In the following description, the specific embodiment of the application will be with reference to as the step performed by one or multi-section computer
And symbol illustrates, unless otherwise stating clearly.Therefore, these steps and operation will have to mention for several times is performed by computer, this paper institutes
The computer execution of finger includes by representing with the computer processing unit of the electronic signal of the data in a structuring pattern
Operation.This operation is converted at the data or the position being maintained in the memory system of the computer, reconfigurable
Or in addition change the running of the computer in a manner of known to the tester of this field.The data structure that the data are maintained
For the provider location of the memory, there is the specific feature as defined in the data format.But the application principle is with above-mentioned text
Word illustrates that be not represented as a kind of limitation, this field tester will appreciate that plurality of step as described below and behaviour
Also it may be implemented in hardware.
Term as used herein " module " can regard the software object to be performed in the arithmetic system as.It is as described herein
Different components, module, engine and service can be regarded as the objective for implementation in the arithmetic system.And device as described herein and side
Method can be implemented in a manner of software, can also be implemented on hardware certainly, within the application protection domain.
Term " first ", " second " and " third " in the application etc. is for distinguishing different objects rather than for retouching
State particular order.In addition, term " comprising " and " having " and their any deformations, it is intended that cover non-exclusive include.
Such as contain the step of process, method, system, product or the equipment of series of steps or module is not limited to list or
Module, but some embodiments further include the step of not listing or module or some embodiments further include for these processes,
Method, product or equipment intrinsic other steps or module.
Referenced herein " embodiment " is it is meant that a particular feature, structure, or characteristic described can wrap in conjunction with the embodiments
It is contained at least one embodiment of the application.Each position in the description occur the phrase might not each mean it is identical
Embodiment, nor the independent or alternative embodiment with other embodiments mutual exclusion.Those skilled in the art explicitly and
Implicitly understand, embodiment described herein can be combined with other embodiments.
Referring to Fig. 1, Fig. 1 is the application scenarios schematic diagram of natural language processing device provided by the embodiments of the present application.Example
Such as, natural language processing device first obtains natural language information input by user and phase associated with natural language information
Close information;Multiple semantic parsing information are obtained according to natural language information, multiple semantic parsing information include semantic parsing information
A, semantic parsing information b and semantic parsing information c;By multiple semantic parsing information and relevant information input prediction model, prediction
Model obtains multiple probability values of corresponding multiple semantic parsing information according to relevant information;Probability value is determined from multiple probability values
Maximum semantic parsing information a parses information for target semanteme.Later again by the corresponding displaying information of target semanteme parsing information
Displaying is on an electronic device.
The embodiment of the present application provides a kind of natural language processing method, and the executive agent of the natural language processing method can be with
It is natural language processing device provided by the embodiments of the present application or is integrated with the electronic equipment of the natural language processing device,
Wherein hardware may be used in the natural language processing device or the mode of software is realized.
The embodiment of the present application will be described from the angle of natural language processing device, and the natural language processing device is specific
It can integrate in the electronic device.The natural language processing method includes:Obtain natural language information input by user, Yi Jiyu
The associated relevant information of natural language information;Multiple semantic parsing information are obtained according to natural language information;By multiple semantemes
Information and relevant information input prediction model are parsed, prediction model obtains corresponding multiple semantic parsing information according to relevant information
Multiple probability values;Determine that the maximum semantic parsing information of probability value parses information for target semanteme from multiple probability values.
It is understood that the executive agent of the embodiment of the present application can be the end of smart mobile phone or tablet computer etc.
End equipment.
Referring to Fig. 2, Fig. 2 is the first flow diagram of natural language processing method provided by the embodiments of the present application.
Natural language processing method provided by the embodiments of the present application is applied to electronic equipment, and idiographic flow can be as follows:
Step 101, obtain natural language information input by user and to natural language information it is associated it is related letter
Breath.
Natural language information can refer to people's language message used in everyday.It can be appreciated that the day of user is commonly used
The information of language or similar works and expressions for everyday use building form.
It can be the text information obtained by input through keyboard to obtain natural language information input by user, input through keyboard packet
The keyboard of entity and virtual keyboard are included, virtual keyboard can include the keyboard that touch screen is shown or display screen display
Then the keyboard shown inputs information by modes such as mouse clicks.It obtains natural language information input by user and can also be logical
Phonetic entry is crossed, input voice information is such as obtained by the microphone of electronic equipment, then identification parses the voice messaging, by it
It is converted to text information.The composing method of the text information is the building form of natural language.Such as natural language information can be with
For " I wants to watch movie ".
When obtaining natural language information, relevant information associated with natural language information is also obtained.
Referring to Fig. 3, Fig. 3 is second of flow diagram of natural language processing method provided by the embodiments of the present application.
The step of obtaining relevant information associated with natural language information, idiographic flow can be as follows:
Step 1011, label time point is obtained while obtaining natural language information.
When getting natural language information, the time point recorded at this time is label time point.Marking time point can be
The present system time of electronic equipment.The system time of electronic equipment can be the time that Network Synchronization is shared or set
The time calculated after fixing time according to electronic equipment itself component.
Step 1012, acquisition marks the history display information of preset time period before time point.
After obtaining label time point, the history display information of preset time period before label time point is obtained.It is wherein default
Time can be the time of electronic apparatus system setting, such as 3 minutes, 5 minutes, 10 minutes.The preset time may be use
The time of family setting.
History display information is included in the information occurred in the display screen of electronic equipment.It is defeated that history display information includes history
Enter at least one in information, dialog history information, historical search information and historical viewings information.
History input information can be the information that preset time period is such as inputted in 3 minutes before time point mark, can be with
The input information of user is saved, such as storage in memory, caching it is medium.Input information be included in chat interface,
The information of the medium place input of search interface, or the information inputted in interface of input method.Even it is included in input method circle
The information of face input, but the information deleted below.The information such as inputted in chat interface, behind the letter that does not send out and delete
Breath.
Dialog history information can be the dialog information occurred in chat interface, including information input by user and other
Information input by user.
Historical search information can be the search information that user opens in the impact of search interface midpoint.As user inputs a name
Word pops up multiple search information according to the title, and then user clicks one of search information and enters the search information pair
The webpage answered, the noun and the corresponding information of webpage are all historical search information.Such as user inputs " Beijing Shenzhen ", pop-up north
Capital to the air ticket in Shenzhen, the train in Beijing to Shenzhen, the different adventures in daily life in Beijing Shenzhen, Beijing Shenzhen the retrievals such as weather letter
Breath, then the weather in wherein Beijing Shenzhen is clicked, then the weather in Beijing Shenzhen and Beijing Shenzhen all remains history retrieval
Information.
Historical viewings information can be information of the user in web page browsing, as user does not input any letter in a browser
Breath, but browsing is provided in the link provided in browser.
Step 1013, it is screened, obtained related to natural language information from history display information according to natural language information
The relevant information of connection.
It obtains after history display information, needing to be screened according to natural language information in preset time period, it will not phase
The information deletion of pass retains relevant information associated with natural language information.As setting includes in natural language information one
Or the information of multiple words is relevant information, the classification of natural language information is identified, by the letter generic with natural language information
Breath is set as relevant information, such as recognizes natural language information as movies category, then is all set as related letter about the information of film
Breath, including movie name information, performer's name information, plot information etc..
Referring to Fig. 4, Fig. 4 is the third flow diagram of natural language processing method provided by the embodiments of the present application.
It is screened from history display information according to natural language information, obtains the step of relevant information associated with natural language information
Suddenly, idiographic flow can be as follows:
Step 1014, history display information and natural language information are inputted into screening model.
Step 1015, screening model is screened according to natural language information from history display information, is obtained and natural language
The associated relevant information of information.
According to natural language information, screening model goes out phase associated with natural language information from history display information sifting
Close information.More accurately, conveniently.First the corresponding text information of natural language information can be carried out disassembling identification, such as " I wants to see
Film " can disassemble into " I ", " thinking ", " seeing ", " film ", and then be screened respectively according to the word after disassembling, can be with root
It is further parsed according to the word after disassembling, " film " is main screening mode, and the screening for carrying out the first round obtains all
It is relevant information to be obtained and see, such as video class with the relevant information of film, the screening taken turns later according to " seeing " progress second
Information.It is screened further according to " I ", such as other people can deleting of mentioning in chat record, only retains local electronic equipment
Relevant information in history display information.It should be noted that " I ", " thinking " can not also be screened.By natural language
After information disassembles identification, the importance of the sub-information after each disassemble can be set, is screened successively according to importance, is screened
Obtained information can then stop screening after being less than certain amount, then determine last time or screen what is obtained second from the bottom time
Information is relevant information associated with natural language information.
Step 102, multiple semantic parsing information are obtained according to natural language information.
Multiple semantic parsing information are obtained after identification natural language information.One or more default fields can first be preset
Then scape substitutes into the natural language information in one or more default scenes, obtain multiple semantic parsing information.
Referring to Fig. 5, Fig. 5 is the 4th kind of flow diagram of natural language processing method provided by the embodiments of the present application.
Step 1021, one or more keywords are extracted from natural language information.
The keyword of natural language information can be first extracted, if natural language information is " I wants to see Swordman ", first will
It splits, and can obtain " I ", " wanting to see ", " Swordman " several sub-informations, then be sorted according to importance, before selection
Several is keyword.
Step 1022, a variety of default semantic scenes are obtained according to one or more keywords.
Then its corresponding default scene is obtained according to keyword, can be such as obtained according to keyword " Swordman " default
Scene can include film scene, novel scene etc..
Step 1023, the corresponding multiple semantic parsing information of natural language information are obtained according to a variety of default semantic scenes.
And then obtain the semantic parsing information of corresponding film and the semantic parsing information of corresponding novel.
Step 103, by multiple semantic parsing information and relevant information input prediction model, prediction model is according to relevant information
Obtain multiple probability values of corresponding multiple semantic parsing information.
Prediction model can be convolutional neural networks model, Recognition with Recurrent Neural Network model or bayesian algorithm model etc..
By multiple semantic parsing information and relevant information input prediction model, prediction model can be to multiple semantic parsing letters
Breath is predicted, and obtains the probability of corresponding each semantic parsing information.It can also be in conjunction with relevant information to multiple semantic solutions
Analysis information is predicted, obtains the probability of corresponding each semantic parsing information.Information source is more rich, prediction it is also more accurate.It needs
It is noted that the screening model in above-described embodiment can be the different piece of same model or screening with prediction model
Model is a part for prediction model.It, can according to keyword Swordman if natural language information is " I wants to see Swordman "
It can include film scene, novel scene etc. to obtain default scene, and then obtain the semantic parsing information of corresponding film and right
Answer the semantic parsing information of novel.Relevant information includes the information of film, then can improve the semantic parsing letter of corresponding film
The probability value of breath makes its probability value be significantly greater than other semanteme parsing information.In another example language message is " Beijing " freely,
Can obtain default scene according to keyword Beijing can include Beijing essential information, tourism of Beijing sight spot, Beijing weather etc., into
And obtain the semantic parsing letter of semantic parsing information, corresponding tourism of Beijing sight spot and Beijing weather of corresponding Beijing essential information etc.
Breath.Relevant information includes the information of weather, specifically can be chat record in, before several days included elsewhere
Gas information can then improve the probability value of the semantic parsing information of corresponding Beijing weather, its probability value is made to be significantly greater than other languages
Justice parsing information.
Step 104, determine that the maximum semantic parsing information of probability value parses information for target semanteme from multiple probability values.
After obtaining the probability value of corresponding multiple semantic parsing information, the maximum semantic parsing information of probability value is therefrom selected,
And determine that semanteme parsing information parses information for target.Later, then can information be parsed according to the target and shows corresponding letter
Breath.Such as show the movie link of Swordman, the movie reviews of Swordman.
Referring to Fig. 6, Fig. 6 is the 5th kind of flow diagram of natural language processing method provided by the embodiments of the present application.
Step 201, obtain natural language information input by user and to natural language information it is associated it is related letter
Breath.
Natural language information can refer to people's language message used in everyday.It can be appreciated that the day of user is commonly used
The information of language or similar works and expressions for everyday use building form.When obtaining natural language information, also obtain associated with natural language information
Relevant information.
Step 202, a variety of default semantic scenes are obtained according to natural language information and relevant information and correspondence is more respectively
Multiple first weighted values of the default semantic scene of kind.
History display information is included in the information occurred in the display screen of electronic equipment.It is defeated that history display information includes history
Enter at least one in information, dialog history information, historical search information and historical viewings information.
A variety of default semantic scenes are obtained according to natural language information, then according to history display information to a variety of default languages
Adopted scene carries out weight assignment, i.e., presetting semantic scene to each assigns a weighted value, its corresponding probability value.Likewise,
Can natural language information and history display information be inputted a Weight prediction model to predict, obtained corresponding a variety of default
The weighted value of semantic scene.It should be noted that default semantic scene of the weighted value less than weight threshold can be excluded.
Step 203, according to a variety of default semantic scenes and multiple first weighted values, it is corresponding more to obtain natural language information
A semanteme parses information the second weighted value corresponding with its.
Natural language information can parse to obtain one or more semantic parsing information under a kind of default semantic scene,
The weighted value of semantic parsing information that parsing obtains under same default semantic scene can be its corresponding first weighted value,
It can be the second weighted value obtained after being adjusted.
Step 204, by multiple semantic parsing information and relevant information input prediction model, prediction model is according to relevant information
Obtain multiple probability values of corresponding multiple semantic parsing information.
By multiple semantic parsing information and relevant information input prediction model, prediction model can be to multiple semantic parsing letters
Breath is predicted, and obtains the probability of corresponding each semantic parsing information.It can also be in conjunction with relevant information to multiple semantic solutions
Analysis information is predicted, obtains the probability of corresponding each semantic parsing information.Information source is more rich, prediction it is also more accurate.
Step 205, the corresponding probability value of semanteme parsing information and the second weighted value are multiplied to obtain multiple weight probability values,
The maximum semantic parsing information of probability value is chosen from multiple weight probability values and parses information as target semanteme.
The corresponding probability value of semanteme parsing information and the second weighted value are multiplied to obtain multiple weight probability values, semanteme parsing
Information is multiplied by the probability value that prediction model obtains with the second weighted value obtained before, more close to default semantic scene.
The maximum semantic parsing information of probability value is chosen from multiple weight probability values and parses information as target semanteme.Then according to mesh
Poster justice parsing information shows corresponding information.
From the foregoing, it will be observed that natural language processing method provided by the embodiments of the present application, by obtaining natural language input by user
Say information and relevant information associated with natural language information;Multiple semantic parsing letters are obtained according to natural language information
Breath;By multiple semantic parsing information and relevant information input prediction model, prediction model obtains corresponding multiple according to relevant information
Multiple probability values of semanteme parsing information;Determine that the maximum semantic parsing information of probability value is semantic for target from multiple probability values
Parse information.Parsing identification is no longer carried out in itself only for natural language information, in prediction model, passes through relevant information pair
Natural language information is assisted in identifying, improve to natural language information parsing identification accuracy, so as to get final mesh
Poster justice parsing information is more in line with the true idea of user.
Referring to Fig. 7, Fig. 7 is the first structure diagram of natural language processing device provided by the embodiments of the present application.
Wherein the natural language processing device 300 is applied to electronic equipment, which includes the first acquisition module
301st, the second acquisition module 302, probability value acquisition module 303 and determining module 304.Wherein:
First acquisition module 301, for obtaining natural language information input by user and related to natural language information
The relevant information of connection.
Natural language information can refer to people's language message used in everyday.It can be appreciated that the day of user is commonly used
The information of language or similar works and expressions for everyday use building form.
It can be the text information obtained by input through keyboard to obtain natural language information input by user, input through keyboard packet
The keyboard of entity and virtual keyboard are included, virtual keyboard can include the keyboard that touch screen is shown or display screen display
Then the keyboard shown inputs information by modes such as mouse clicks.It obtains natural language information input by user and can also be logical
Phonetic entry is crossed, input voice information is such as obtained by the microphone of electronic equipment, then identification parses the voice messaging, by it
It is converted to text information.The composing method of the text information is the building form of natural language.Such as natural language information can be with
For " I wants to watch movie ".
When obtaining natural language information, relevant information associated with natural language information is also obtained.
Referring to Fig. 8, Fig. 8 is second of structure diagram of natural language processing device provided by the embodiments of the present application.
In the embodiment, the first acquisition module 301 includes label time point acquisition submodule 3011, history display acquisition of information submodule
3012 and relevant information acquisition submodule 3013.Wherein:
Time point acquisition submodule 3011 is marked, for obtaining natural language information while obtains label time point.
When getting natural language information, the time point recorded at this time is label time point.Marking time point can be
The present system time of electronic equipment.The system time of electronic equipment can be the time that Network Synchronization is shared or set
The time calculated after fixing time according to electronic equipment itself component.
History display acquisition of information submodule 3012 marks the history display letter of preset time period before time point for acquisition
Breath.
After obtaining label time point, the history display information of preset time period before label time point is obtained.It is wherein default
Time can be the time of electronic apparatus system setting, such as 3 minutes, 5 minutes, 10 minutes.The preset time may be use
The time of family setting.
History display information is included in the information occurred in the display screen of electronic equipment.It is defeated that history display information includes history
Enter at least one in information, dialog history information, historical search information and historical viewings information.
History input information can be the information that preset time period is such as inputted in 3 minutes before time point mark, can be with
The input information of user is saved, such as storage in memory, caching it is medium.Input information be included in chat interface,
The information of the medium place input of search interface, or the information inputted in interface of input method.Even it is included in input method circle
The information of face input, but the information deleted below.The information such as inputted in chat interface, behind the letter that does not send out and delete
Breath.
Dialog history information can be the dialog information occurred in chat interface, including information input by user and other
Information input by user.
Historical search information can be the search information that user opens in the impact of search interface midpoint.As user inputs a name
Word pops up multiple search information according to the title, and then user clicks one of search information and enters the search information pair
The webpage answered, the noun and the corresponding information of webpage are all historical search information.Such as user inputs " Beijing Shenzhen ", pop-up north
Capital to the air ticket in Shenzhen, the train in Beijing to Shenzhen, the different adventures in daily life in Beijing Shenzhen, Beijing Shenzhen the retrievals such as weather letter
Breath, then the weather in wherein Beijing Shenzhen is clicked, then the weather in Beijing Shenzhen and Beijing Shenzhen all remains history retrieval
Information.
Historical viewings information can be information of the user in web page browsing, as user does not input any letter in a browser
Breath, but browsing is provided in the link provided in browser.
Relevant information acquisition submodule 3013 for being screened from history display information according to natural language information, obtains
Relevant information associated with natural language information.
It obtains after history display information, needing to be screened according to natural language information in preset time period, it will not phase
The information deletion of pass retains relevant information associated with natural language information.As setting includes in natural language information one
Or the information of multiple words is relevant information, the classification of natural language information is identified, by the letter generic with natural language information
Breath is set as relevant information, such as recognizes natural language information as movies category, then is all set as related letter about the information of film
Breath, including movie name information, performer's name information, plot information etc..
In some embodiments, history display information includes history input information, historical search information and historical viewings letter
Breath;
Relevant information acquisition submodule is additionally operable to history display information and natural language information input screening model, sieve
Modeling type is screened according to natural language information from history display information, is obtained and the associated related letter of natural language information
Breath.
Second acquisition module 302, for obtaining multiple semantic parsing information according to natural language information.
Multiple semantic parsing information are obtained after identification natural language information.One or more default fields can first be preset
Then scape substitutes into the natural language information in one or more default scenes, obtain multiple semantic parsing information.
Referring to Fig. 9, Fig. 9 is the third structure diagram of natural language processing device provided by the embodiments of the present application.
In the embodiment, the second acquisition module 302 includes keyword extraction submodule 3021, semantic scene acquisition submodule 3022 and language
Justice parsing acquisition of information submodule 3023.Wherein:
Keyword extraction submodule 3021, for extracting one or more keywords from natural language information.
The keyword of natural language information can be first extracted, if natural language information is " I wants to see Swordman ", first will
It splits, and can obtain " I ", " wanting to see ", " Swordman " several sub-informations, then be sorted according to importance, before selection
Several is keyword.
Semantic scene acquisition submodule 3022, for obtaining a variety of default semantic scenes according to one or more keywords.
Then its corresponding default scene is obtained according to keyword, can be such as obtained according to keyword " Swordman " default
Scene can include film scene, novel scene etc..
Semanteme parsing acquisition of information submodule 3023, for obtaining natural language information pair according to a variety of default semantic scenes
The multiple semantic parsing information answered.
And then obtain the semantic parsing information of corresponding film and the semantic parsing information of corresponding novel.
Probability value acquisition module 303 for multiple semantemes to be parsed information and relevant information input prediction model, predicts mould
Type obtains multiple probability values of corresponding multiple semantic parsing information according to relevant information.
Prediction model can be convolutional neural networks model, Recognition with Recurrent Neural Network model or bayesian algorithm model etc..
By multiple semantic parsing information and relevant information input prediction model, prediction model can be to multiple semantic parsing letters
Breath is predicted, and obtains the probability of corresponding each semantic parsing information.It can also be in conjunction with relevant information to multiple semantic solutions
Analysis information is predicted, obtains the probability of corresponding each semantic parsing information.Information source is more rich, prediction it is also more accurate.It needs
It is noted that the screening model in above-described embodiment can be the different piece of same model or screening with prediction model
Model is a part for prediction model.It, can according to keyword Swordman if natural language information is " I wants to see Swordman "
It can include film scene, novel scene etc. to obtain default scene, and then obtain the semantic parsing information of corresponding film and right
Answer the semantic parsing information of novel.Relevant information includes the information of film, then can improve the semantic parsing letter of corresponding film
The probability value of breath makes its probability value be significantly greater than other semanteme parsing information.In another example language message is " Beijing " freely,
Can obtain default scene according to keyword Beijing can include Beijing essential information, tourism of Beijing sight spot, Beijing weather etc., into
And obtain the semantic parsing letter of semantic parsing information, corresponding tourism of Beijing sight spot and Beijing weather of corresponding Beijing essential information etc.
Breath.Relevant information includes the information of weather, specifically can be chat record in, before several days included elsewhere
Gas information can then improve the probability value of the semantic parsing information of corresponding Beijing weather, its probability value is made to be significantly greater than other languages
Justice parsing information.
Determining module 304, for determining that the maximum semantic parsing information of probability value is semantic for target from multiple probability values
Parse information.
After obtaining the probability value of corresponding multiple semantic parsing information, the maximum semantic parsing information of probability value is therefrom selected,
And determine that semanteme parsing information parses information for target.Later, then can information be parsed according to the target and shows corresponding letter
Breath.Such as show the movie link of Swordman, the movie reviews of Swordman.
Referring to Fig. 10, Figure 10 is the 4th kind of structural representation of natural language processing device provided by the embodiments of the present application
Figure.The natural language processing device 300 includes the first acquisition module 301, the second acquisition module 302, probability value acquisition module 303
With determining module 304.Wherein:
First acquisition module 301, for obtaining natural language information input by user and related to natural language information
The relevant information of connection.
Natural language information can refer to people's language message used in everyday.It can be appreciated that the day of user is commonly used
The information of language or similar works and expressions for everyday use building form.When obtaining natural language information, also obtain associated with natural language information
Relevant information.
Second acquisition module 302, for obtaining multiple semantic parsing information according to natural language information.Specifically, second
Acquisition module 302 includes the first acquisition submodule 3024 and the second acquisition submodule 3025.
First acquisition submodule 3024, for obtaining a variety of default semantic fields according to natural language information and relevant information
Scape and multiple first weighted values for corresponding to a variety of default semantic scenes respectively.
History display information is included in the information occurred in the display screen of electronic equipment.It is defeated that history display information includes history
Enter at least one in information, dialog history information, historical search information and historical viewings information.
A variety of default semantic scenes are obtained according to natural language information, then according to history display information to a variety of default languages
Adopted scene carries out weight assignment, i.e., presetting semantic scene to each assigns a weighted value, its corresponding probability value.Likewise,
Can natural language information and history display information be inputted a Weight prediction model to predict, obtained corresponding a variety of default
The weighted value of semantic scene.It should be noted that default semantic scene of the weighted value less than weight threshold can be excluded.
Second acquisition submodule 3025, for according to a variety of default semantic scenes and multiple first weighted values, obtaining nature
The corresponding multiple semantemes of language message parse information the second weighted value corresponding with its.
Natural language information can parse to obtain one or more semantic parsing information under a kind of default semantic scene,
The weighted value of semantic parsing information that parsing obtains under same default semantic scene can be its corresponding first weighted value,
It can be the second weighted value obtained after being adjusted.
Probability value acquisition module 303 for multiple semantemes to be parsed information and relevant information input prediction model, predicts mould
Type obtains multiple probability values of corresponding multiple semantic parsing information according to relevant information.
By multiple semantic parsing information and relevant information input prediction model, prediction model can be to multiple semantic parsing letters
Breath is predicted, and obtains the probability of corresponding each semantic parsing information.It can also be in conjunction with relevant information to multiple semantic solutions
Analysis information is predicted, obtains the probability of corresponding each semantic parsing information.Information source is more rich, prediction it is also more accurate.
Determining module 304 is additionally operable to be multiplied to obtain by the corresponding probability value of semanteme parsing information and the second weighted value multiple
Weight probability value is chosen the maximum semantic parsing information of probability value from multiple weight probability values and is parsed as target semanteme and believed
Breath.
The corresponding probability value of semanteme parsing information and the second weighted value are multiplied to obtain multiple weight probability values, semanteme parsing
Information is multiplied by the probability value that prediction model obtains with the second weighted value obtained before, more close to default semantic scene.
The maximum semantic parsing information of probability value is chosen from multiple weight probability values and parses information as target semanteme.Then according to mesh
Poster justice parsing information shows corresponding information.
From the foregoing, it will be observed that natural language processing device provided by the embodiments of the present application, by obtaining natural language input by user
Say information and relevant information associated with natural language information;Multiple semantic parsing letters are obtained according to natural language information
Breath;By multiple semantic parsing information and relevant information input prediction model, prediction model obtains corresponding multiple according to relevant information
Multiple probability values of semanteme parsing information;Determine that the maximum semantic parsing information of probability value is semantic for target from multiple probability values
Parse information.Parsing identification is no longer carried out in itself only for natural language information, in prediction model, passes through relevant information pair
Natural language information is assisted in identifying, improve to natural language information parsing identification accuracy, so as to get final mesh
Poster justice parsing information is more in line with the true idea of user.
When it is implemented, Yi Shang modules can be independent entity to realize, arbitrary combination can also be carried out, is made
It is realized for same or several entities, the specific implementation of more than modules can be found in the embodiment of the method for front, herein not
It repeats again.
In the embodiment of the present application, natural language processing device belongs to same with the natural language processing method in foregoing embodiments
One design, can run the either method provided in natural language processing embodiment of the method on natural language processing device,
Specific implementation process refers to the embodiment of natural language processing method, and details are not described herein again.
The embodiment of the present application also provides a kind of electronic equipment.Please refer to Fig.1 1, electronic equipment 600 include processor 601 with
And memory 602.Wherein, processor 601 is electrically connected with memory 602.
Processor 600 is the control centre of electronic equipment 600, utilizes various interfaces and the entire electronic equipment of connection
Various pieces computer program in memory 602 and are called by operation or load store and are stored in memory 602
Interior data perform the various functions of electronic equipment 600 and handle data, so as to carry out integral monitoring to electronic equipment 600.
Memory 602 can be used for storage software program and unit, and processor 601 is stored in memory 602 by operation
Computer program and unit, so as to perform various functions application and data processing.Memory 602 can mainly include storage
Program area and storage data field, wherein, storing program area can storage program area, the computer program needed at least one function
(such as sound-playing function, image player function etc.) etc.;Storage data field can be stored to be created according to using for electronic equipment
Data etc..In addition, memory 602 can include high-speed random access memory, nonvolatile memory, example can also be included
Such as at least one disk memory, flush memory device or other volatile solid-state parts.Correspondingly, memory 602 may be used also
To include Memory Controller, to provide access of the processor 601 to memory 602.
In the embodiment of the present application, the processor 601 in electronic equipment 600 can be according to the steps, by one or one
The corresponding instruction of process of a above computer program is loaded into memory 602, and be stored in by the operation of processor 601
Computer program in reservoir 602, it is as follows so as to fulfill various functions:
Obtain natural language information input by user and relevant information associated with the natural language information;
Multiple semantic parsing information are obtained according to the natural language information;
By the multiple semantic parsing information and the relevant information input prediction model, the prediction model is according to
Relevant information obtains multiple probability values of corresponding the multiple semantic parsing information;
Determine that the maximum semantic parsing information of probability value parses information for target semanteme from the multiple probability value.
In some embodiments, processor 601 is additionally operable to perform following steps:
One or more keywords are extracted from the natural language information;
A variety of default semantic scenes are obtained according to one or more of keywords;
The corresponding multiple semantic parsing information of the natural language information are obtained according to a variety of default semantic scenes.
In some embodiments, processor 601 is additionally operable to perform following steps:
Label time point is obtained while obtaining the natural language information;
Obtain it is described label time point before preset time period history display information;
It is screened, obtained and the natural language information phase from the history display information according to the natural language information
Associated relevant information.
In some embodiments, the history display information is searched including history input information, dialog history information, history
At least one of in rope information and historical viewings information;
Processor 601 is additionally operable to perform following steps:
By the history display information and natural language information input screening model;
The screening model is screened according to the natural language information from the history display information, obtain with it is described from
The associated relevant information of right language message.
In some embodiments, processor 601 is additionally operable to perform following steps:
A variety of default semantic scenes are obtained according to the natural language information and the relevant information and correspondence is more respectively
Multiple first weighted values of the default semantic scene of kind;
According to a variety of default semantic scenes and the multiple first weighted value, obtain the natural language information and correspond to
Multiple semantic parsing information the second weighted value corresponding with its;
By it is described it is semantic parse the corresponding probability value of information be multiplied with second weighted value to obtain multiple weights it is general
Rate value chooses the maximum semantic parsing information of probability value from the multiple weight probability value and parses information as target semanteme.
It can be seen from the above, electronic equipment provided by the embodiments of the present application, by obtaining natural language information input by user,
And relevant information associated with natural language information;Multiple semantic parsing information are obtained according to natural language information;To be more
A semantic parsing information and relevant information input prediction model, prediction model obtain corresponding multiple semantic parsings according to relevant information
Multiple probability values of information;Determine the maximum semantic parsing information of probability value for target semanteme parsing letter from multiple probability values
Breath.Parsing identification is no longer carried out in itself only for natural language information, in prediction model, by relevant information to natural language
Speech information is assisted in identifying, and improves the accuracy to natural language information parsing identification, so as to get final goal it is semantic
Parsing information is more in line with the true idea of user.
Also referring to Figure 12, in some embodiments, electronic equipment 600 can also include:Display 603, radio frequency
Circuit 604, voicefrequency circuit 605 and power supply 606.Wherein, wherein, display 603, radio circuit 604, voicefrequency circuit 605 with
And power supply 606 is electrically connected respectively with processor 601.
Display 603 is displayed for by information input by user or is supplied to the information of user and various figures to use
Family interface, these graphical user interface can be made of figure, text, icon, video and its arbitrary combination.Display 603
Can include display panel, in some embodiments, may be used liquid crystal display (Liquid Crystal Display,
LCD) or display surface is configured in the forms such as Organic Light Emitting Diode (Organic Light-Emitting Diode, OLED)
Plate.
Radio circuit 604 can be used for transceiving radio frequency signal, to be set by radio communication with the network equipment or other electronics
It is standby to establish wireless telecommunications, the receiving and transmitting signal between the network equipment or other electronic equipments.
Voicefrequency circuit 605 can be used for providing the audio interface between user and electronic equipment by loud speaker, microphone.
Power supply 606 is used to all parts power supply of electronic equipment 600.In some embodiments, power supply 606 can
With logically contiguous by power-supply management system and processor 601, thus charged, discharged by power-supply management system realization management,
And the functions such as power managed.
Although being not shown in Figure 12, electronic equipment 600 can also include camera, bluetooth unit etc., and details are not described herein.
It is understood that the electronic equipment of the embodiment of the present application can be the end of smart mobile phone or tablet computer etc.
End equipment.
The embodiment of the present application also provides a kind of storage medium, and storage medium is stored with computer program, works as computer program
When running on computers so that computer performs the natural language processing method in any of the above-described embodiment, such as:By obtaining
Take natural language information input by user and relevant information associated with natural language information;According to natural language information
Obtain multiple semantic parsing information;By multiple semantic parsing information and relevant information input prediction model, prediction model is according to phase
It closes information and obtains multiple probability values of corresponding multiple semantic parsing information;The semanteme of probability value maximum is determined from multiple probability values
It parses information and parses information for target semanteme.
In the embodiment of the present application, storage medium can be magnetic disc, CD, read-only memory (Read Only Memory,
) or random access memory (Random Access Memory, RAM) etc. ROM.
In the above-described embodiments, it all emphasizes particularly on different fields to the description of each embodiment, there is no the portion being described in detail in some embodiment
Point, it may refer to the associated description of other embodiment.
It should be noted that for the natural language processing method of the embodiment of the present application, this field common test personnel
It is appreciated that all or part of flow of realization the embodiment of the present application natural language processing method, is that can pass through computer program
It is completed to control relevant hardware, computer program can be stored in a computer read/write memory medium, be such as stored in electricity
It in the memory of sub- equipment, and is performed, may include in the process of implementation as certainly by least one processor in the electronic equipment
The flow of the embodiment of right language processing method.Wherein, storage medium can be magnetic disc, CD, read-only memory, arbitrary access
Memory body etc..
For the natural language processing device of the embodiment of the present application, each functional unit can be integrated in a processing core
In piece or each unit is individually physically present, can also two or more units integrate in a unit.On
The form realization that hardware had both may be used in integrated unit is stated, can also be realized in the form of SFU software functional unit.Integrated
If unit is realized in the form of SFU software functional unit and is independent product sale or in use, can also be stored in one
In computer read/write memory medium, storage medium is for example read-only memory, disk or CD etc..
A kind of natural language processing method, apparatus, storage medium and the electronics provided above the embodiment of the present application is set
Standby to be described in detail, the principle and implementation of this application are described for specific case used herein, more than
The explanation of embodiment is merely used to help understand the present processes and its core concept;Meanwhile for those skilled in the art
Member, according to the thought of the application, there will be changes in specific embodiments and applications, in conclusion this explanation
Book content should not be construed as the limitation to the application.
Claims (12)
1. a kind of natural language processing method, applied to electronic equipment, which is characterized in that the method includes:
Obtain natural language information input by user and relevant information associated with the natural language information;
Multiple semantic parsing information are obtained according to the natural language information;
By the multiple semantic parsing information and the relevant information input prediction model, the prediction model is according to the correlation
Information obtains multiple probability values of corresponding the multiple semantic parsing information;
Determine that the maximum semantic parsing information of probability value parses information for target semanteme from the multiple probability value.
2. natural language processing method according to claim 1, which is characterized in that obtained according to the natural language information
The step of multiple semantic parsing information, including;
One or more keywords are extracted from the natural language information;
A variety of default semantic scenes are obtained according to one or more of keywords;
The corresponding multiple semantic parsing information of the natural language information are obtained according to a variety of default semantic scenes.
3. natural language processing method according to claim 1, which is characterized in that obtain and the natural language information phase
The step of associated relevant information, including:
Label time point is obtained while obtaining the natural language information;
Obtain it is described label time point before preset time period history display information;
It is screened, obtained associated with the natural language information from the history display information according to the natural language information
Relevant information.
4. natural language processing method according to claim 3, which is characterized in that the history display information includes history
Input at least one in information, dialog history information, historical search information and historical viewings information;
It is screened, obtained associated with the natural language information from the history display information according to the natural language information
Relevant information the step of, including:
By the history display information and natural language information input screening model;
The screening model is screened according to the natural language information from the history display information, is obtained and the natural language
Say the associated relevant information of information.
5. natural language processing method according to claim 3, which is characterized in that obtained according to the natural language information
The step of multiple semantic parsing information, including:
A variety of default semantic scenes are obtained according to the natural language information and the relevant information and are corresponded to respectively a variety of pre-
If multiple first weighted values of semantic scene;
According to a variety of default semantic scenes and the multiple first weighted value, it is corresponding more to obtain the natural language information
A semanteme parses information the second weighted value corresponding with its;
The step of maximum semantic parsing information of probability value parses information as target semanteme is chosen from the multiple probability value,
Including:
The semantic corresponding probability value of information that parses is multiplied with second weighted value to obtain multiple weight probability values,
The maximum semantic parsing information of probability value is chosen from the multiple weight probability value and parses information as target semanteme.
6. a kind of natural language processing device, applied to electronic equipment, which is characterized in that described device includes:
First acquisition module, for obtaining natural language information input by user and associated with the natural language information
Relevant information;
Second acquisition module, for obtaining multiple semantic parsing information according to the natural language information;
Probability value acquisition module, it is described for the multiple semanteme to be parsed information and the relevant information input prediction model
Prediction model obtains multiple probability values of corresponding the multiple semantic parsing information according to the relevant information;
Determining module, for determining that the maximum semantic parsing information of probability value is parsed for target semanteme from the multiple probability value
Information.
7. natural language processing device according to claim 6, which is characterized in that second acquisition module includes:
Keyword extraction submodule, for extracting one or more keywords from the natural language information;
Semantic scene acquisition submodule, for obtaining a variety of default semantic scenes according to one or more of keywords;
Semanteme parsing acquisition of information submodule, for obtaining the natural language information pair according to a variety of default semantic scenes
The multiple semantic parsing information answered.
8. natural language processing device according to claim 6, which is characterized in that first acquisition module includes:
Time point acquisition submodule is marked, for obtaining natural language information while obtains label time point;
History display acquisition of information submodule, for obtain it is described label time point before preset time period history display information;
Relevant information acquisition submodule for being screened from the history display information according to the natural language information, obtains
Relevant information associated with the natural language information.
9. natural language processing device according to claim 8, which is characterized in that the history display information includes history
Input information, historical search information and historical viewings information;
The relevant information acquisition submodule is additionally operable to the history display information and natural language information input screening
Model, the screening model are screened from the history display information according to the natural language information, are obtained and the nature
The associated relevant information of language message.
10. natural language processing device according to claim 8, which is characterized in that second acquisition module includes:
First acquisition submodule, for obtaining a variety of default semantic fields according to the natural language information and the relevant information
Scape and multiple first weighted values for corresponding to a variety of default semantic scenes respectively;
Second acquisition submodule, for according to a variety of default semantic scenes and the multiple first weighted value, described in acquisition
The corresponding multiple semantemes of natural language information parse information the second weighted value corresponding with its;
The determining module is additionally operable to the semantic corresponding probability value of information that parses being multiplied with second weighted value
Multiple weight probability values are obtained, the maximum semantic parsing information of probability value is chosen from the multiple weight probability value as target
Semanteme parsing information.
11. a kind of storage medium, is stored thereon with computer program, which is characterized in that when the computer program is in computer
During upper operation so that the computer performs such as natural language processing method described in any one of claim 1 to 5.
12. a kind of electronic equipment, including processor and memory, the memory has computer program, which is characterized in that described
Processor is by calling the computer program, for performing such as natural language processing side described in any one of claim 1 to 5
Method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711466311.7A CN108197105B (en) | 2017-12-28 | 2017-12-28 | Natural language processing method, device, storage medium and electronic equipment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711466311.7A CN108197105B (en) | 2017-12-28 | 2017-12-28 | Natural language processing method, device, storage medium and electronic equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108197105A true CN108197105A (en) | 2018-06-22 |
CN108197105B CN108197105B (en) | 2021-08-24 |
Family
ID=62585809
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711466311.7A Active CN108197105B (en) | 2017-12-28 | 2017-12-28 | Natural language processing method, device, storage medium and electronic equipment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108197105B (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110941631A (en) * | 2019-10-30 | 2020-03-31 | 联想(北京)有限公司 | Information processing method and electronic equipment |
CN111784087A (en) * | 2019-04-03 | 2020-10-16 | 长鑫存储技术有限公司 | Production information processing method, device, system, equipment and readable storage medium |
CN111797631A (en) * | 2019-04-04 | 2020-10-20 | 北京猎户星空科技有限公司 | Information processing method and device and electronic equipment |
CN113704408A (en) * | 2021-08-31 | 2021-11-26 | 工银科技有限公司 | Retrieval method, retrieval apparatus, electronic device, storage medium, and program product |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103268313A (en) * | 2013-05-21 | 2013-08-28 | 北京云知声信息技术有限公司 | Method and device for semantic analysis of natural language |
CN104199810A (en) * | 2014-08-29 | 2014-12-10 | 科大讯飞股份有限公司 | Intelligent service method and system based on natural language interaction |
CN104360994A (en) * | 2014-12-04 | 2015-02-18 | 科大讯飞股份有限公司 | Natural language understanding method and natural language understanding system |
CN105786793A (en) * | 2015-12-23 | 2016-07-20 | 百度在线网络技术(北京)有限公司 | Method and device for analyzing semanteme of spoken language text information |
CN106057205A (en) * | 2016-05-06 | 2016-10-26 | 北京云迹科技有限公司 | Intelligent robot automatic voice interaction method |
-
2017
- 2017-12-28 CN CN201711466311.7A patent/CN108197105B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103268313A (en) * | 2013-05-21 | 2013-08-28 | 北京云知声信息技术有限公司 | Method and device for semantic analysis of natural language |
CN104199810A (en) * | 2014-08-29 | 2014-12-10 | 科大讯飞股份有限公司 | Intelligent service method and system based on natural language interaction |
CN104360994A (en) * | 2014-12-04 | 2015-02-18 | 科大讯飞股份有限公司 | Natural language understanding method and natural language understanding system |
CN105786793A (en) * | 2015-12-23 | 2016-07-20 | 百度在线网络技术(北京)有限公司 | Method and device for analyzing semanteme of spoken language text information |
CN106057205A (en) * | 2016-05-06 | 2016-10-26 | 北京云迹科技有限公司 | Intelligent robot automatic voice interaction method |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111784087A (en) * | 2019-04-03 | 2020-10-16 | 长鑫存储技术有限公司 | Production information processing method, device, system, equipment and readable storage medium |
CN111797631A (en) * | 2019-04-04 | 2020-10-20 | 北京猎户星空科技有限公司 | Information processing method and device and electronic equipment |
CN110941631A (en) * | 2019-10-30 | 2020-03-31 | 联想(北京)有限公司 | Information processing method and electronic equipment |
CN110941631B (en) * | 2019-10-30 | 2024-01-23 | 联想(北京)有限公司 | Information processing method and electronic equipment |
CN113704408A (en) * | 2021-08-31 | 2021-11-26 | 工银科技有限公司 | Retrieval method, retrieval apparatus, electronic device, storage medium, and program product |
Also Published As
Publication number | Publication date |
---|---|
CN108197105B (en) | 2021-08-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20230015606A1 (en) | Named entity recognition method and apparatus, device, and storage medium | |
CN109145303A (en) | Name entity recognition method, device, medium and equipment | |
US10845950B2 (en) | Web browser extension | |
CN108363792A (en) | Link generation method, device, storage medium and the electronic equipment of five application page | |
CN107958042B (en) | Target topic pushing method and mobile terminal | |
WO2019124647A1 (en) | Method and computer apparatus for automatically building or updating hierarchical conversation flow management model for interactive ai agent system, and computer-readable recording medium | |
CN108197105A (en) | Natural language processing method, apparatus, storage medium and electronic equipment | |
CN108334608A (en) | Link generation method, device, storage medium and the electronic equipment of five application page | |
CN107527619A (en) | The localization method and device of Voice command business | |
CN113095080B (en) | Theme-based semantic recognition method and device, electronic equipment and storage medium | |
CN101782833B (en) | Intelligent operating system and method | |
CN108549568A (en) | Using entrance processing method, apparatus, storage medium and electronic equipment | |
CN111539212A (en) | Text information processing method and device, storage medium and electronic equipment | |
CN110989847B (en) | Information recommendation method, device, terminal equipment and storage medium | |
CN111597804B (en) | Method and related device for training entity recognition model | |
WO2021017238A1 (en) | Text generation method and apparatus | |
CN108287918A (en) | Method for playing music, device, storage medium based on five application page and electronic equipment | |
CN109634436A (en) | Association method, device, equipment and the readable storage medium storing program for executing of input method | |
CN109710732A (en) | Information query method, device, storage medium and electronic equipment | |
CN107155121A (en) | The display methods and device of Voice command text | |
CN111612178A (en) | Model diagnosis method and related equipment | |
CN109101505A (en) | A kind of recommended method, recommendation apparatus and the device for recommendation | |
CN109271624A (en) | A kind of target word determines method, apparatus and storage medium | |
CN114117056B (en) | Training data processing method and device and storage medium | |
CN110196833A (en) | Searching method, device, terminal and the storage medium of application program |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: Changan town in Guangdong province Dongguan 523860 usha Beach Road No. 18 Applicant after: OPPO Guangdong Mobile Communications Co.,Ltd. Address before: Changan town in Guangdong province Dongguan 523860 usha Beach Road No. 18 Applicant before: GUANGDONG OPPO MOBILE TELECOMMUNICATIONS Corp.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |