CN110020429A - Method for recognizing semantics and equipment - Google Patents
Method for recognizing semantics and equipment Download PDFInfo
- Publication number
- CN110020429A CN110020429A CN201910147719.0A CN201910147719A CN110020429A CN 110020429 A CN110020429 A CN 110020429A CN 201910147719 A CN201910147719 A CN 201910147719A CN 110020429 A CN110020429 A CN 110020429A
- Authority
- CN
- China
- Prior art keywords
- nlp
- result
- keyword
- answer statement
- nlp result
- 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
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS OR SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
Abstract
The embodiment of the present invention provides a kind of method for recognizing semantics and equipment, and this method includes handling by NLP technology curent interrogation information, the first NLP result of acquisition；Wherein the first NLP result includes type, keyword and movement；Judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance；The answer statement being of practical significance if it exists obtains the 3rd NLP result then according to the first NLP result and the 2nd NLP as a result, be replaced operation or padding operation；Corresponding answer statement is obtained according to the 3rd NLP result.The recognition result that the embodiment of the present invention is generated based on context semanteme, it can more accurately explain the semanteme of user, adapt to habit of speaking (such as the use of pronoun in person-to-person communication, the use etc. of breviary sentence), in order to during human-computer dialogue, correct feedback is provided a user, user experience is improved.
Description
Technical field
The present embodiments relate to field of artificial intelligence more particularly to a kind of method for recognizing semantics and equipment.
Background technique
Semantics recognition is a branch of artificial intelligence, semantics recognition technology can analyze webpage, file, mail, audio,
Light data in forum, social media, application field is extensive, both can directly apply and the industries such as medical treatment, education, finance.
Can also by technical interface be applied to all intelligent sound interaction scenarios, as smart home, vehicle-mounted voice, wearable device,
VR, robot etc. can also be divided into from interactive mode: true question and answer, knowledge retrieval, classification problem etc..Intelligent sound is handed over
Mutually it is seen as application scenarios most worthy of expecting in the following artificial intelligence technology.
In the existing semantics recognition scheme based on context, the fixation clause of identification weather inquiry, if subsequent when occurring
Between or city name, then generate the time or the city corresponding Weather information.
However, scene is single in above scheme and the fixation of question sentence mode lacks flexibility, the habit of speaking of people is not adapted to
(such as use of pronoun, the use etc. of breviary sentence).
Summary of the invention
The embodiment of the present invention provides a kind of method for recognizing semantics and equipment, to improve the flexibility of semantics recognition, adapts to people
Habit of speaking.
In a first aspect, the embodiment of the present invention provides a kind of method for recognizing semantics, comprising:
Curent interrogation information is handled by NLP technology, obtains the first NLP result；Wherein the first NLP result
Including type, keyword and movement；
Judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance；
The answer statement being of practical significance if it exists, then according to the first NLP result and the 2nd NLP as a result, into
Row replacement operation or padding operation obtain the 3rd NLP result；
Corresponding answer statement is obtained according to the 3rd NLP result.
In a kind of possible design, it is described according to the first NLP result with the 2nd NLP as a result, being replaced
Operation or padding operation, comprising:
According to the first NLP as a result, judging in curent interrogation information with the presence or absence of directive property pronoun or there is only nouns；
If there is only noun, according to the first NLP result and the 2nd NLP as a result, being replaced operation；
Directive property pronoun if it exists, then according to the first NLP as a result, judge type in the first NLP result with
It whether defined acts；If defined, according to the first NLP result and the 2nd NLP as a result, carrying out cover behaviour
Make.
In a kind of possible design, it is described according to the first NLP as a result, judging whether only deposit in curent interrogation information
Before noun, further includes:
Judge keyword in the type Yu the 2nd NLP result of keyword in the first NLP result type whether
It is identical；
It is described according to the first NLP as a result, judging whether there is only nouns in curent interrogation information, comprising:
If the first NLP result is identical as the type of keyword in the 2nd NLP result, according to described first
NLP is as a result, judge whether there is only nouns in curent interrogation information.
In a kind of possible design, it is described according to the first NLP result with the 2nd NLP as a result, being replaced
Operation, comprising:
Keyword in the 2nd NLP result is replaced with to the keyword in the first NLP result.
In a kind of possible design, it is described according to the first NLP as a result, judging to whether there is in curent interrogation information
Directive property pronoun, comprising:
Judge whether the keyword in the first NLP result is directive property pronoun, if so, determining curent interrogation information
In there are directive property pronouns.
In a kind of possible design, it is described according to the first NLP result and the 2nd NLP as a result, carrying out cover
Operation, comprising:
According to the corresponding answer statement of the 2nd NLP result, obtain corresponding with the keyword in the first NLP result
Content；
Delete the keyword in the first NLP result, and by the content cover to the key of the first NLP result
At word.
In a kind of possible design, which is characterized in that whether the 2nd NLP result for judging previous inquiry message deposits
After the answer statement being of practical significance, further includes:
The answer statement being of practical significance if it does not exist then obtains corresponding answer statement according to the first NLP result.
Second aspect, the embodiment of the present invention provide a kind of semantics recognition equipment, comprising:
Processing module obtains the first NLP result for handling by NLP technology curent interrogation information；Wherein institute
Stating the first NLP result includes type, keyword and movement；
Judgment module, for judging the 2nd NLP result of previous inquiry message, whether there are the answer languages of practical significance
Sentence；
Operation module, the answer statement for being of practical significance if it exists, then according to the first NLP result and described the
Two NLP obtain the 3rd NLP result as a result, be replaced operation or padding operation；
Feedback module, for obtaining corresponding answer statement according to the 3rd NLP result.
In a kind of possible design, the operation module includes:
Judging unit, the answer statement for being of practical significance if it exists, according to the first NLP as a result, judgement is current
It whether there is directive property pronoun in inquiry message or there is only nouns；
Replacement operation unit, if for there is only noun, according to the first NLP result and the 2nd NLP as a result,
It is replaced operation；
Padding operation unit, for directive property pronoun if it exists, then according to the first NLP as a result, judging described first
Whether the type and movement in NLP result are defined；If defined, according to the first NLP result and described second
NLP is as a result, carry out padding operation.
In a kind of possible design, the judging unit is specifically used for:
Judge keyword in the type Yu the 2nd NLP result of keyword in the first NLP result type whether
It is identical；If they are the same, then judge whether there is only nouns for curent interrogation information.
In a kind of possible design, the replacement operation unit is specifically used for:
Keyword in the 2nd NLP result is replaced with into the keyword in the first NLP result, obtains described
Three NLP are as a result, and obtain corresponding answer statement according to the 3rd NLP result.
In a kind of possible design, the judging unit is specifically used for:
Judge whether the keyword in the first NLP result is directive property pronoun, if so, determining curent interrogation information
In there are directive property pronouns.
In a kind of possible design, the padding operation unit is specifically used for:
According to the corresponding answer statement of the 2nd NLP result, obtain corresponding with the keyword in the first NLP result
Content；
Delete the keyword in the first NLP result, and by the content cover to the key of the first NLP result
At word.
In a kind of possible design, if the feedback module is also used to the 2nd NLP result, there is no have practical meaning
The answer statement of justice then obtains corresponding answer statement according to the first NLP result.
The third aspect, the embodiment of the present invention provide a kind of semantics recognition equipment, comprising: at least one processor and storage
Device；
The memory stores computer executed instructions；
At least one described processor executes the computer executed instructions of memory storage so that it is described at least one
Processor executes method described in the various possible designs of first aspect and first aspect as above.
Fourth aspect, the embodiment of the present invention provide a kind of computer readable storage medium, the computer-readable storage medium
Computer executed instructions are stored in matter, when processor execute the computer executed instructions when, realize first aspect as above with
And method described in the various possible designs of first aspect.
Method for recognizing semantics provided in this embodiment and equipment, this method is by using NLP technology to curent interrogation information
It is handled, and generates the first NLP result；Wherein the first NLP result includes type, keyword and movement；Judge previous
Whether there are the answer statements of practical significance for the corresponding 2nd NLP result of inquiry message；The answer being of practical significance if it exists
Sentence is then generated according to the first NLP result and the 2nd NLP result by replacement operation or padding operation described current
The final recognition result of inquiry message.The recognition result that the present embodiment is generated based on context semanteme can be solved more accurately
The semanteme of user is released, the habit of speaking (such as use of pronoun, the use etc. of breviary sentence) in person-to-person communication is adapted to, in order to
During human-computer dialogue, correct feedback is provided a user, improves user experience.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is this hair
Bright some embodiments for those of ordinary skill in the art without any creative labor, can be with
It obtains other drawings based on these drawings.
Fig. 1 is the configuration diagram for the man-machine voice interaction system that one embodiment of the invention provides；
Fig. 2 is the structural schematic diagram for the semantics recognition equipment that further embodiment of this invention provides；
Fig. 3 is the flow diagram for the method for recognizing semantics that yet another embodiment of the invention provides；
Fig. 4 be another embodiment of the present invention provides semantics recognition equipment structural schematic diagram；
Fig. 5 is the flow diagram for the method for recognizing semantics that yet another embodiment of the invention provides；
Fig. 6 be another embodiment of the present invention provides semantics recognition equipment structural schematic diagram；
Fig. 7 is the structural schematic diagram for the semantics recognition equipment that yet another embodiment of the invention provides；
Fig. 8 is the hardware structural diagram for the semantics recognition equipment that one embodiment of the invention provides.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
Fig. 1 is the configuration diagram for the man-machine voice interaction system that one embodiment of the invention provides.As shown in Figure 1, this reality
The system for applying example offer includes terminal 101 and server 102.Wherein, terminal 101 can for children-story machine, mobile phone, plate,
Car-mounted terminal etc..The present embodiment is not particularly limited the implementation of terminal 101, if the terminal 101 can with user into
Row interactive voice.
Interactive voice (Speech Interaction) is based on speech recognition, natural language processing (Natural
Language Processing, NLP), the technologies such as speech synthesis, under multiple practical applications scene, assigning terminal " can listen, meeting
Say, understand you " experience of the intelligent human-machine interaction of formula.Suitable for multiple application scenarios, including intelligent answer, intelligence play, intelligence
The scenes such as lookup.
Terminal 101 receives the voice messaging of user's input, and obtains answer after carrying out semantics recognition to the voice messaging
Sentence is fed back to user.Specifically, which can answer language locally obtaining according to the corpus itself stored
Sentence, can also be sent to server 102 for the voice messaging, obtain answer statement after carrying out semantics recognition by server 102,
Then user is fed back to by terminal 101.If semantics recognition is carried out in terminal, terminal can be used as semantics recognition equipment, if
Semantics recognition is carried out in server, then server can be used as semantics recognition equipment.The present embodiment to concrete implementation mode not
It is particularly limited, terminal 101 is local to be obtained answer statement and obtain answer statement all according to voice messaging by server 102
It can.
Fig. 2 is the structural block diagram for the semantics recognition equipment that further embodiment of this invention provides.As shown in Fig. 2, the equipment 20
It include: speech recognition module 201, processing module 202 and feedback module 203；The speech recognition module 201 is used for receiving
The voice messaging of family input, and the voice messaging is identified, text information is generated, and the text information is sent to
The processing module 202；The processing module 202 handles to obtain NLP as a result, and will for carrying out NLP to the text information
The NLP result is sent to the feedback module 203；The feedback module 203 is used for according to the NLP as a result, in conventional language
Material searches whether the answer statement for having with practical significance in library, and if it exists, the returning with practical significance that will then find
It answers sentence and feeds back to user, if it does not exist, then obtain answer statement of revealing all the details at random or according to pre-defined rule from corpus of revealing all the details,
And the answer statement of revealing all the details is fed back into user.
During specific implementation, the speech recognition module 201 receives voice messaging automatically, and to the voice messaging
It is identified, generates text information, and the text information is sent to the processing module 202；The processing module 202 is right
The text information carries out NLP and handles to obtain NLP as a result, and the NLP result is sent to the feedback module 203；It is described
Feedback module 203 according to the NLP as a result, searched whether in conventional corpus exist with practical significance answer statement,
If it exists, then the answer statement with practical significance found is fed back into user, if it does not exist, then from corpus of revealing all the details
Answer statement of revealing all the details is obtained at random or according to pre-defined rule, and the answer statement of revealing all the details is fed back into user.
It can be seen that NLP result is as answer statement of the foundation for feedback for searching answer statement in the process
Correctness plays an important role.However, in the prior art, the processing module 202 only receives the speech recognition module 201
To curent interrogation information handled, obtain the NLP result.Not in the presence of the answer statement fed back according to the NLP result
Appropriate situation, influences user experience.Based on this, the embodiment of the present invention provides a kind of method for recognizing semantics, to improve semantic knowledge
Other accuracy.
The method for recognizing semantics provided below using specific embodiment come the present invention will be described in detail embodiment.
Fig. 3 is the flow diagram for the method for recognizing semantics that yet another embodiment of the invention provides.As shown in figure 3, this method
Include:
S301, curent interrogation information is handled by NLP technology, obtains the first NLP result；Wherein described first
NLP result includes type, keyword and movement.
Optionally, the executing subject of this method can be the intelligent terminal that interactive voice can be carried out with people, such as: hand
Machine, plate, phone wrist-watch etc..
Optionally, the inquiry message can be voice messaging, can also be the text information of user's input.If the inquiry
Ask that information is voice messaging, then it is described that curent interrogation information is handled by NLP technology, it may include: to described current
Inquiry message carries out speech recognition, generates text information；The text information is handled by NLP technology.
Optionally, described that curent interrogation information is handled by NLP technology, it may include: to curent interrogation information
Carry out part-of-speech tagging.The process of noun, verb, adjective, adverbial word etc. is labeled as to the word in sentence, it can also be to part of speech
The result of mark carries out keyword (Key) and extracts, acts (Action) extraction, auxiliary word (Particle) extraction and type
(Type) it determines.
Such as: user input inquiry message " how is the weather of today? ", by NLP technology to the inquiry message, carry out
Can be obtained after processing corresponding NLP result be " Type=weather, Key1=today, Key2=weather, Action=inquiry,
How is Particle=".
S302, judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance.
In the present embodiment, for step S301 and step S302 execution sequence without limitation, step S301 and step
S302 can also can sequentially be executed with parallel execution
In the present embodiment, the previous inquiry message, refer to it is adjacent with curent interrogation information and curent interrogation information it
The inquiry message of preceding user's input.
Such as: it is directed to human-computer dialogue content below:
People: " how is the weather of today? "
Machine: " weather of today is clear to cloudy."
People: " tomorrow? "
If curent interrogation information will be used as " tomorrow ", " how is the weather of today " is then previous inquiry message.
In the present embodiment, the answer statement being of practical significance can be for can for the intelligent terminal of human-computer interaction
Corresponding answer statement is found in conventional corpus.Correspondingly, the not no answer statement of practical significance, it can be in conventional language
Expect to search in library less than corresponding answer statement, answers language at random or according to revealing all the details for pre-defined rule feedback from corpus of revealing all the details
Sentence.
Such as: be directed to previous inquiry message " how is the weather of today ", the NLP of generation as a result, can only terminal feedback
Module can find corresponding answer statement " weather of today is clear to cloudy ", the i.e. inquiry message in conventional corpus
There is the answer statement being of practical significance in " how is the weather of today ".
For curent interrogation information " tomorrow ", the NLP of generation is as a result, the feedback module of intelligent terminal can not be in conventional language
Material finds corresponding answer statement in library, is defined as chat of revealing all the details, and needs random in corpus of revealing all the details or is established rules according to pre-
The answer statement being of practical significance is not present in the meaningless answer statement of revealing all the details then fed back, the i.e. inquiry message " tomorrow ".
S303, the answer statement being of practical significance if it exists are then tied according to the first NLP result and the 2nd NLP
Fruit is replaced operation or padding operation, obtains the 3rd NLP result.
If the answer statement of previous inquiry message is the answer statement being of practical significance, rather than distribution is chatted for revealing all the details
Meaningless answer statement of revealing all the details, then according to previous inquiry message NLP result (first corresponding with curent interrogation information
NLP result and the 2nd NLP result), be replaced or cover processing, obtain for curent interrogation information final NLP as a result,
The i.e. described 3rd NLP result.
For example, for curent interrogation information " tomorrow ", being needed previous inquiry message " today in above-mentioned human-computer dialogue
Weather how? " corresponding 2nd NLP as a result, and corresponding first NLP of curent interrogation information " tomorrow " as a result, into
Row replacement operation or padding operation obtain the 3rd NLP result.
The replacement operation can be " bright in curent interrogation information will to replace with " today " in previous inquiry message
It ", then the 3rd NLP result can for " how is the weather of tomorrow? "
The padding operation can be to arrive (" how is weather ") cover in addition to " today " in previous inquiry message
In curent interrogation information " tomorrow " below, then the 3rd NLP result can for " how is weather tomorrow? "
S304, corresponding answer statement is obtained according to the 3rd NLP result.
Optionally, the feedback module of intelligent terminal can be according to the 3rd NLP result from the sheet for being stored in intelligent terminal
The corpus on ground obtains, or obtains from server side.
Method for recognizing semantics provided in this embodiment is handled curent interrogation information by using NLP technology, and raw
At the first NLP result；Wherein the first NLP result includes type, keyword and movement；Judge that previous inquiry message is corresponding
Whether there are the answer statements of practical significance for 2nd NLP result；The answer statement being of practical significance if it exists, then according to
First NLP result generates the final of the curent interrogation information by replacement operation or padding operation with the 2nd NLP result
Recognition result.The recognition result that the present embodiment is generated based on context semanteme can more accurately explain the semanteme of user, fit
The habit of speaking (such as use of pronoun, the use etc. of breviary sentence) in person-to-person communication is answered, in order in human-computer dialogue process
In, correct feedback is provided a user, user experience is improved.
In a specific embodiment, Fig. 4 be another embodiment of the present invention provides semantics recognition equipment structural representation
Figure, as shown in figure 4, the semantics recognition equipment 40 may include: that speech recognition module 201, processing module 202, NLP result are deposited
Store up module 206, operation module 205 and feedback module 203.The speech recognition module 201, for receiving the voice of user's input
Information, and the voice messaging is identified, text information is generated, and the text information is sent to the processing module
202.The processing module 202 handles to obtain NLP as a result, and by the NLP result for carrying out NLP to the text information
It is sent to the context Understanding Module and the NLP result cache module.The NLP result memory module 206, before storing
The corresponding NLP of one voice messaging is as a result, and be sent to the context Understanding Module for the corresponding NLP result of previous voice messaging.
The judgment module 204, for judging the 2nd NLP result of previous inquiry message, whether there are the answer languages of practical significance
Sentence, and if it exists, then the operation module 205 is according to current speech information and the corresponding NLP of previous voice messaging as a result, logical
Replacement operation or padding operation are crossed, obtains the 3rd NLP as a result, and the 3rd NLP result is sent to the feedback module
203.The feedback module 203, for according to the 3rd NLP as a result, searching in conventional corpus has practical significance
Answer statement, and the answer statement with practical significance found will be fed back to user.That is, speech recognition module
201 and processing module 202 for executing step S301；Context Understanding Module is for executing step S302 and S303；Feedback module
203 for executing step S304.
Optionally, the processing module 202 obtains JS object numbered musical notation after can handling inquiry message
The NLP result of (JavaScript Object Notation, Json) format.The NLP result memory module 206 can be with column
The form of table stores NLP result.
In a specific embodiment, method for recognizing semantics provided by the above embodiment can also include:
S305, the answer statement being of practical significance if it does not exist then obtain corresponding answer according to the first NLP result
Sentence.
If the answer statement being of practical significance is not present in previous voice messaging, show the previous voice for user's input
Information, intelligent terminal can only feed back to user from the answer statement of revealing all the details for the not practical significance that corpus obtains of revealing all the details, can be with
Determine previous voice messaging and current speech information and uncorrelated, therefore can be directly according to the first NLP of current speech information
As a result, obtaining corresponding answer statement.
Fig. 5 is the flow diagram for the method for recognizing semantics that yet another embodiment of the invention provides.As shown in figure 5, this method
Include:
S501, curent interrogation information is handled by NLP technology, obtains the first NLP result；Wherein described first
NLP result includes type, keyword and movement.
S502, judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance.
Step S501 and step S502 is similar with step S301 in above-described embodiment and step S302 in the present embodiment, this
Place repeats no more.
S503, according to the first NLP as a result, judge in curent interrogation information with the presence or absence of directive property pronoun or there is only
Noun.
Pronoun is the word for replacing noun and playing the role of noun.Pronoun may include: personal pronoun (you, I, he etc.), it is anti-
Body pronoun (myself, yourself, himself, own etc.), interrogative pronoun (where, who, when etc.), indefinite pronoun (it is some,
Very much, any, each etc.), directive property pronoun (this, that, here, there, which etc.).
Optionally, judge may include: with the presence or absence of directive property pronoun in curent interrogation information by curent interrogation information into
Row word segmentation processing, and judge whether each participle is directive property pronoun.Such as: to INQUIRE statement, " that sight spot is on what ground
Side? " " that " " sight spot " " " " what " " place " is obtained after carrying out word segmentation processing, part of speech is carried out to each participle respectively and is sentenced
Disconnected, available " that " is directive property pronoun.
In a specific embodiment, it is described according to the first NLP as a result, judging to whether there is in curent interrogation information
Directive property pronoun, comprising: judge whether the keyword in the first NLP result is directive property pronoun, if so, determining current
There are directive property pronouns in inquiry message.
Such as: it is directed to human-computer dialogue content below:
People: " tourist attractions what Shenzhen has joyful? "
Machine: " Window on the World, Splendid China local culture village, Shenzhen joy paddy, the botanical garden Xian Hu, little Mei sand, blue and green generation
Boundary."
People: " me is helped to navigate to first "
" Type=point of interest (Point in the corresponding first NLP result of curent interrogation information " me is helped to navigate to first "
Of Interest, POI), Key1=I, Key2=first, Action=navigation "." Key2=in first NLP result
One " it is directive property pronoun, then determine that there are directive property pronouns in curent interrogation information.
In a specific embodiment, it is described according to the first NLP as a result, judging whether only deposit in curent interrogation information
Before noun, further includes: judge in the first NLP result keyword in the type Yu the 2nd NLP result of keyword
Type it is whether identical.
It is described according to the first NLP as a result, judging whether there is only nouns in curent interrogation information, comprising: if described
First NLP result is identical as the type of keyword in the 2nd NLP result, then according to the first NLP as a result, judgement is current
Whether there is only nouns in inquiry message.
Such as: it is directed to human-computer dialogue content below:
People: " how is the weather of today? "
Machine: " weather of today is clear to cloudy."
People: " tomorrow? "
" Type=is chatted, Key=tomorrow, Action=in the corresponding first NLP result of curent interrogation information " tomorrow "
Casting, Particle=", " Type=days in the corresponding 2nd NLP result of previous inquiry message " how is the weather of today "
Gas, Key1=today, Key2=weather, Action=inquiry, how is Particle=".By will be in the first NLP result
" Key=tomorrow " compares with " Key1=today " in the 2nd NLP result, available, and two keywords belong to the time
Noun, type are identical.It therefore, can be further to whether there is only nouns to judge in curent interrogation information.
If S504, there is only nouns, according to the first NLP result and the 2nd NLP as a result, being replaced behaviour
Make, obtains the 3rd NLP result.
In a specific embodiment, it is described according to the first NLP result with the 2nd NLP as a result, being replaced
Operation, comprising: the keyword in the 2nd NLP result is replaced with to the keyword in the first NLP result.
Such as: in the corresponding first NLP result of curent interrogation information " tomorrow " " Type=is chatted, Key=tomorrow,
Action=casting, Particle=", in the corresponding 2nd NLP result of previous inquiry message " how is the weather of today "
" Type=weather, Key1=today, Key2=weather, Action=inquiry, how is Particle=".Due to curent interrogation
There is only the time noun " today " that time noun " tomorrow " will then replace in previous inquiry message tomorrow in information, then replace
Change the obtained corresponding 3rd NLP result of curent interrogation information of operation be " Type=weather, Key1=tomorrow, Key2=weather,
Action=inquiry, Particle=is how " namely " how is the weather of tomorrow ".So that intelligent terminal is according to the 3rd NLP
As a result answer statement is searched in conventional corpus.
S505, if it exists directive property pronoun, then according to the first NLP as a result, judging the class in the first NLP result
Whether type and movement are defined；If defined, according to the first NLP result with the 2nd NLP as a result, being mended
Bit manipulation obtains the 3rd NLP result.
In a specific embodiment, it is described according to the first NLP result and the 2nd NLP as a result, carrying out cover
Operation, comprising:
S5051, according to the corresponding answer statement of the 2nd NLP result, obtain and the key in the first NLP result
The corresponding content of word.
S5052, keyword in the first NLP result is deleted, and by the content cover to the first NLP result
Keyword at.
Such as:
In the corresponding 2nd NLP result of previous inquiry message " tourist attractions what Shenzhen has joyful " " Type=POI,
The Shenzhen Key1=, the sight spot Key2=, Key3=is joyful, Action=inquiry ".The feedback module of intelligent terminal can be according to
Two NLP results find answer statement " Window on the World, Splendid China local culture village, Shenzhen joy paddy, the botanical garden Xian Hu,
Little Mei sand, the blue and green world." in the corresponding first NLP result of curent interrogation information " me is helped to navigate to first " " Type=POI,
Key1=I, Key2=first, Action=navigation ".Since directive property pronoun is " Key2=first ", then by above-mentioned time
First sight spot " Window on the World " cover in sentence is answered at directive property pronoun " first " to get to the 3rd NLP result
" Type=POI, Key1=I, Key2=Window on the World, Action=navigation ", namely " me is helped to navigate to Window on the World ".With
Intelligent terminal is set to search answer statement in conventional corpus according to the 3rd NLP result.
S506, corresponding answer statement is obtained according to the 3rd NLP result.
Step S506 is similar with step S304 in above-described embodiment in the present embodiment, and details are not described herein again.
End-point detecting method provided in this embodiment, by specifically judging in curent interrogation information with the presence or absence of directive property generation
Word or there is only noun, determines and uses replacement operation or padding operation to curent interrogation information and previous inquiry message, can
Efficiently directive property pronoun will be present in realization or there is only the imperfect question sentences of noun to expand completely, so as to carry out semantics recognition
Intelligent terminal quickly and accurately obtains question and answer sentence.It avoids due to whether there is directive property pronoun in curent interrogation information or only depositing
It is imperfect caused by the noun, it is mistaken for chat of revealing all the details, and lead to not obtain correct answer statement, influence user experience.
Fig. 6 be another embodiment of the present invention provides semantics recognition equipment structural schematic diagram.As shown in fig. 6, the semanteme
Identify that equipment 60 includes: framing module 801, detection module 802 and determining module 803.
Processing module 202 obtains the first NLP result for handling by NLP technology curent interrogation information；Its
Described in the first NLP result include type, keyword and movement.
Optionally, the inquiry message can be voice messaging, can also be the text information of user's input.If the inquiry
Ask that information is voice messaging, then it is described that curent interrogation information is handled by NLP technology, it may include: to described current
Inquiry message carries out speech recognition, generates text information；The text information is handled by NLP technology.
Optionally, described that curent interrogation information is handled by NLP technology, it may include: to curent interrogation information
Carry out part-of-speech tagging.The process of noun, verb, adjective, adverbial word etc. is labeled as to the word in sentence, it can also be to part of speech
The result of mark carries out keyword (Key) and extracts, acts (Action) extraction, auxiliary word (Particle) extraction and type
(Type) it determines.
Such as: user input inquiry message " how is the weather of today? ", by NLP technology to the inquiry message, carry out
Can be obtained after processing corresponding NLP result be " Type=weather, Key1=today, Key2=weather, Action=inquiry,
How is Particle=".
Judgment module 204, for judging the 2nd NLP result of previous inquiry message, whether there are the answers of practical significance
Sentence.
The previous inquiry message refers to and curent interrogation information before user input adjacent with curent interrogation information
Inquiry message.
Such as: it is directed to human-computer dialogue content below:
People: " how is the weather of today? "
Machine: " weather of today is clear to cloudy."
People: " tomorrow? "
If curent interrogation information will be used as " tomorrow ", " how is the weather of today " is then previous inquiry message.
In the present embodiment, the answer statement being of practical significance can be for can for the intelligent terminal of human-computer interaction
Corresponding answer statement is found in conventional corpus.Correspondingly, the not no answer statement of practical significance, it can be in conventional language
Expect to search in library less than corresponding answer statement, answers language at random or according to revealing all the details for pre-defined rule feedback from corpus of revealing all the details
Sentence.
Such as: be directed to previous inquiry message " how is the weather of today ", the NLP of generation as a result, can only terminal feedback
Module can find corresponding answer statement " weather of today is clear to cloudy ", the i.e. inquiry message in conventional corpus
There is the answer statement being of practical significance in " how is the weather of today ".
For curent interrogation information " tomorrow ", the NLP of generation is as a result, the feedback module of intelligent terminal can not be in conventional language
Material finds corresponding answer statement in library, is defined as chat of revealing all the details, and needs random in corpus of revealing all the details or is established rules according to pre-
The answer statement being of practical significance is not present in the meaningless answer statement of revealing all the details then fed back, the i.e. inquiry message " tomorrow ".
Operation module 205, the answer statement for being of practical significance if it exists, then according to the first NLP result and institute
The 2nd NLP is stated as a result, being replaced operation or padding operation, obtains the 3rd NLP result.
If the answer statement of previous inquiry message is the answer statement being of practical significance, rather than distribution is chatted for revealing all the details
Meaningless answer statement of revealing all the details, then according to previous inquiry message NLP result (first corresponding with curent interrogation information
NLP result and the 2nd NLP result), be replaced or cover processing, obtain for curent interrogation information final NLP as a result,
The i.e. described 3rd NLP result.
For example, for curent interrogation information " tomorrow ", being needed previous inquiry message " today in above-mentioned human-computer dialogue
Weather how? " corresponding 2nd NLP as a result, and corresponding first NLP of curent interrogation information " tomorrow " as a result, into
Row replacement operation or padding operation obtain the 3rd NLP result.
The replacement operation can be " bright in curent interrogation information will to replace with " today " in previous inquiry message
It ", then the 3rd NLP result can for " how is the weather of tomorrow? "
The padding operation can be to arrive (" how is weather ") cover in addition to " today " in previous inquiry message
In curent interrogation information " tomorrow " below, then the 3rd NLP result can for " how is weather tomorrow? "
Feedback module 203, for obtaining corresponding answer statement according to the 3rd NLP result.
Optionally, the feedback module 203 can be according to the 3rd NLP result from being stored in the local of intelligent terminal
Corpus obtains, or obtains from server side.
Semantics recognition equipment provided in an embodiment of the present invention, processing module 202 believe curent interrogation by using NLP technology
Breath is handled, and generates the first NLP result；Wherein the first NLP result includes type, keyword and movement；Judge mould
Block 204 judges the corresponding 2nd NLP result of previous inquiry message, and whether there are the answer statements of practical significance；There is reality if it exists
The answer statement of border meaning, then operation module 205 passes through replacement behaviour according to the first NLP result and the 2nd NLP result
Work or padding operation generate the final recognition result of the curent interrogation information.The knowledge that the present embodiment is generated based on context semanteme
Not as a result, it is possible to more accurately explain the semanteme of user, (such as pronoun makes the habit of speaking for adapting in person-to-person communication
With the use etc. of breviary sentence), in order to during human-computer dialogue, provide a user correct feedback, improve user experience.
Fig. 7 is the structural schematic diagram for the semantics recognition equipment that yet another embodiment of the invention provides.As shown in fig. 7, the semanteme
Identification equipment 70 operation module 205 include:
Judging unit 2051, the answer statement for being of practical significance if it exists, according to the first NLP as a result, judgement
It whether there is directive property pronoun in curent interrogation information or there is only nouns；
Replacement operation unit 2052, if for there is only nouns, according to the first NLP result and the 2nd NLP
As a result, being replaced operation；
Padding operation unit 2053, for directive property pronoun if it exists, then according to the first NLP as a result, described in judgement
Whether the type and movement in the first NLP result are defined；If defined, according to the first NLP result with it is described
2nd NLP is as a result, carry out padding operation.
Optionally, the judging unit is specifically used for: judging in the first NLP result type of keyword and described the
Whether the type of keyword is identical in two NLP results；If they are the same, then judge whether there is only nouns for curent interrogation information.
Optionally, the replacement operation unit, is specifically used for: the keyword in the 2nd NLP result is replaced with institute
The keyword in the first NLP result is stated, obtains the 3rd NLP as a result, and obtaining corresponding time according to the 3rd NLP result
Answer sentence.
Optionally, the judging unit is specifically used for: judging whether the keyword in the first NLP result is directive property
Pronoun, if so, determining that there are directive property pronouns in curent interrogation information.
Optionally, the padding operation unit is specifically used for: according to the corresponding answer statement of the 2nd NLP result, obtaining
Obtain content corresponding with the keyword in the first NLP result；Delete the keyword in the first NLP result, and by institute
Content cover is stated at the keyword of the first NLP result.
Optionally, the feedback module is also used to: if the answer being of practical significance is not present for the 2nd NLP result
Sentence then obtains corresponding answer statement according to the first NLP result.
End-point detection equipment provided in an embodiment of the present invention, can be used for executing above-mentioned embodiment of the method, realization principle
Similar with technical effect, details are not described herein again for the present embodiment.
Fig. 8 is the hardware structural diagram for the semantics recognition equipment that one embodiment of the invention provides.As shown in figure 8, this reality
The semantics recognition equipment 80 for applying example offer includes: at least one processor 801 and memory 802.Wherein, processor 801, storage
Device 802 is connected by bus 803.
Optionally, which can also include communication component, and communication component passes through bus 803 and processor
801 and memory 802 connect.
During specific implementation, at least one processor 801 executes the computer execution that the memory 802 stores and refers to
It enables, so that at least one processor 801 executes method for recognizing semantics performed by semantics recognition equipment 80 as above.
When the method for recognizing semantics of the present embodiment is executed by server, which can be sent to inquiry message
Server, and the answer statement fed back to is received from server.
The specific implementation process of processor 801 can be found in above method embodiment, and it is similar that the realization principle and technical effect are similar,
Details are not described herein again for the present embodiment.
In above-mentioned embodiment shown in Fig. 8, it should be appreciated that processor can be central processing unit (English:
Central Processing Unit, referred to as: CPU), can also be other general processors, digital signal processor (English:
Digital Signal Processor, referred to as: DSP), specific integrated circuit (English: Application Specific
Integrated Circuit, referred to as: ASIC) etc..General processor can be microprocessor or the processor is also possible to
Any conventional processor etc..Hardware processor can be embodied directly in conjunction with the step of invention disclosed method to have executed
At, or in processor hardware and software module combination execute completion.
Memory may include high speed RAM memory, it is also possible to and it further include non-volatile memories NVM, for example, at least one
Magnetic disk storage.
Bus can be industry standard architecture (Industry Standard Architecture, ISA) bus, outer
Portion's apparatus interconnection (Peripheral Component, PCI) bus or extended industry-standard architecture (Extended
Industry Standard Architecture, EISA) bus etc..Bus can be divided into address bus, data/address bus, control
Bus etc..For convenient for indicating, the bus in illustrations does not limit only a bus or a type of bus.
The application also provides a kind of computer readable storage medium, and calculating is stored in the computer readable storage medium
Machine executes instruction, and when processor executes the computer executed instructions, realizes the semantic knowledge that semantics recognition equipment as above executes
Other method.
The application also provides a kind of computer readable storage medium, and calculating is stored in the computer readable storage medium
Machine executes instruction, and when processor executes the computer executed instructions, realizes the semantic knowledge that semantics recognition equipment as above executes
Other method.
Above-mentioned computer readable storage medium, above-mentioned readable storage medium storing program for executing can be by any kind of volatibility or non-
Volatile storage devices or their combination realize that, such as static random access memory (SRAM), electrically erasable is only
It reads memory (EEPROM), Erasable Programmable Read Only Memory EPROM (EPROM), programmable read only memory (PROM) is read-only to deposit
Reservoir (ROM), magnetic memory, flash memory, disk or CD.Readable storage medium storing program for executing can be general or specialized computer capacity
Any usable medium enough accessed.
A kind of illustrative readable storage medium storing program for executing is coupled to processor, to enable a processor to from the readable storage medium storing program for executing
Information is read, and information can be written to the readable storage medium storing program for executing.Certainly, readable storage medium storing program for executing is also possible to the composition portion of processor
Point.Processor and readable storage medium storing program for executing can be located at specific integrated circuit (Application Specific Integrated
Circuits, referred to as: ASIC) in.Certainly, processor and readable storage medium storing program for executing can also be used as discrete assembly and be present in equipment
In.
Those of ordinary skill in the art will appreciate that: realize that all or part of the steps of above-mentioned each method embodiment can lead to
The relevant hardware of program instruction is crossed to complete.Program above-mentioned can be stored in a computer readable storage medium.The journey
When being executed, execution includes the steps that above-mentioned each method embodiment to sequence；And storage medium above-mentioned include: ROM, RAM, magnetic disk or
The various media that can store program code such as person's CD.
Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention., rather than its limitations；To the greatest extent
Pipe present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: its according to
So be possible to modify the technical solutions described in the foregoing embodiments, or to some or all of the technical features into
Row equivalent replacement；And these are modified or replaceed, various embodiments of the present invention technology that it does not separate the essence of the corresponding technical solution
The range of scheme.
Claims (16)
1. a kind of method for recognizing semantics characterized by comprising
Curent interrogation information is handled by natural language processing NLP technology, obtains the first NLP result；Wherein described
One NLP result includes type, keyword and movement；
Judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance；
The answer statement being of practical significance if it exists, then according to the first NLP result with the 2nd NLP as a result, being replaced
Operation or padding operation are changed, the 3rd NLP result is obtained；
Corresponding answer statement is obtained according to the 3rd NLP result.
2. the method according to claim 1, wherein described according to the first NLP result and the 2nd NLP
As a result, being replaced operation or padding operation, comprising:
According to the first NLP as a result, judging in curent interrogation information with the presence or absence of directive property pronoun or there is only nouns；
If there is only noun, according to the first NLP result and the 2nd NLP as a result, being replaced operation；
Directive property pronoun if it exists, then according to the first NLP as a result, judging the type and movement in the first NLP result
It is whether defined；If defined, according to the first NLP result and the 2nd NLP as a result, carrying out padding operation.
3. according to the method described in claim 2, it is characterized in that, it is described according to the first NLP as a result, judging curent interrogation
Whether there is only before noun in information, further includes:
Judge whether the type of keyword in the type with the 2nd NLP result of keyword in the first NLP result is identical；
It is described according to the first NLP as a result, judging whether there is only nouns in curent interrogation information, comprising:
If the first NLP result is identical as the type of keyword in the 2nd NLP result, tied according to the first NLP
Fruit judges whether there is only nouns in curent interrogation information.
4. according to the method described in claim 3, it is characterized in that, described according to the first NLP result and the 2nd NLP
As a result, being replaced operation, comprising:
Keyword in the 2nd NLP result is replaced with to the keyword in the first NLP result.
5. according to the method described in claim 2, it is characterized in that, it is described according to the first NLP as a result, judging curent interrogation
It whether there is directive property pronoun in information, comprising:
Judge whether the keyword in the first NLP result is directive property pronoun, if so, determining to deposit in curent interrogation information
In directive property pronoun.
6. according to the method described in claim 2, it is characterized in that, described according to the first NLP result and the 2nd NLP
As a result, carrying out padding operation, comprising:
According to the corresponding answer statement of the 2nd NLP result, obtain in corresponding with the keyword in the first NLP result
Hold；
Delete the keyword in the first NLP result, and by the content cover to the keyword of the first NLP result
Place.
7. method according to claim 1-6, which is characterized in that the second of the previous inquiry message of judgement
Whether there are after the answer statement of practical significance for NLP result, further includes:
The answer statement being of practical significance if it does not exist then obtains corresponding answer statement according to the first NLP result.
8. a kind of semantics recognition equipment characterized by comprising
Processing module obtains the first NLP knot for handling by natural language processing NLP technology curent interrogation information
Fruit；Wherein the first NLP result includes type, keyword and movement；
Judgment module, for judging the 2nd NLP result of previous inquiry message, whether there are the answer statements of practical significance；
Operation module, the answer statement for being of practical significance if it exists, then according to the first NLP result and described second
NLP obtains the 3rd NLP result as a result, be replaced operation or padding operation；
Feedback module, for obtaining corresponding answer statement according to the 3rd NLP result.
9. equipment according to claim 8, which is characterized in that the operation module includes:
Judging unit, the answer statement for being of practical significance if it exists, according to the first NLP as a result, judging curent interrogation
It whether there is directive property pronoun in information or there is only nouns；
Replacement operation unit, if for there is only nouns, according to the first NLP result and the 2nd NLP as a result, carrying out
Replacement operation；
Padding operation unit, for directive property pronoun if it exists, then according to the first NLP as a result, judging the first NLP knot
Whether the type and movement in fruit are defined；If defined, tied according to the first NLP result and the 2nd NLP
Fruit carries out padding operation.
10. equipment according to claim 9, which is characterized in that the judging unit is specifically used for:
Judge whether the type of keyword in the type with the 2nd NLP result of keyword in the first NLP result is identical；
If they are the same, then judge whether there is only nouns for curent interrogation information.
11. equipment according to claim 10, which is characterized in that the replacement operation unit is specifically used for:
Keyword in the 2nd NLP result is replaced with into the keyword in the first NLP result, obtains the third
NLP is as a result, and obtain corresponding answer statement according to the 3rd NLP result.
12. equipment according to claim 9, which is characterized in that the judging unit is specifically used for:
Judge whether the keyword in the first NLP result is directive property pronoun, if so, determining to deposit in curent interrogation information
In directive property pronoun.
13. equipment according to claim 9, which is characterized in that the padding operation unit is specifically used for:
According to the corresponding answer statement of the 2nd NLP result, obtain in corresponding with the keyword in the first NLP result
Hold；
Delete the keyword in the first NLP result, and by the content cover to the keyword of the first NLP result
Place.
14. according to the described in any item equipment of claim 8 to 13, which is characterized in that the feedback module is also used to: if for
The answer statement being of practical significance is not present in the 2nd NLP result, then obtains corresponding answer according to the first NLP result
Sentence.
15. a kind of semantics recognition equipment characterized by comprising at least one processor and memory；
The memory stores computer executed instructions；
At least one described processor executes the computer executed instructions of the memory storage, so that at least one described processing
Device executes method for recognizing semantics as described in any one of claim 1 to 7.
16. a kind of computer readable storage medium, which is characterized in that be stored with computer in the computer readable storage medium
It executes instruction, when processor executes the computer executed instructions, realizes semanteme as described in any one of claim 1 to 7
Recognition methods.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910147719.0A CN110020429B (en) | 2019-02-27 | 2019-02-27 | Semantic recognition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910147719.0A CN110020429B (en) | 2019-02-27 | 2019-02-27 | Semantic recognition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110020429A true CN110020429A (en) | 2019-07-16 |
| CN110020429B CN110020429B (en) | 2023-05-23 |
Family
ID=67189089
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910147719.0A Active CN110020429B (en) | 2019-02-27 | 2019-02-27 | Semantic recognition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110020429B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111681647A (en) * | 2020-06-10 | 2020-09-18 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for recognizing word slot |
| CN113035179A (en) * | 2021-03-03 | 2021-06-25 | 科大讯飞股份有限公司 | Voice recognition method, device, equipment and computer readable storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6101490A (en) * | 1991-07-19 | 2000-08-08 | Hatton; Charles Malcolm | Computer system program for creating new ideas and solving problems |
| CN101593206A (en) * | 2009-06-25 | 2009-12-02 | 腾讯科技（深圳）有限公司 | Searching method and device based on answer in the question and answer interaction platform |
| CN102637192A (en) * | 2012-02-17 | 2012-08-15 | 清华大学 | Method for answering with natural language |
| CN105930452A (en) * | 2016-04-21 | 2016-09-07 | 北京紫平方信息技术股份有限公司 | Smart answering method capable of identifying natural language |
| CN107609101A (en) * | 2017-09-11 | 2018-01-19 | 远光软件股份有限公司 | Intelligent interactive method, equipment and storage medium |
| CN108920604A (en) * | 2018-06-27 | 2018-11-30 | 百度在线网络技术（北京）有限公司 | Voice interactive method and equipment |
-
2019
- 2019-02-27 CN CN201910147719.0A patent/CN110020429B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6101490A (en) * | 1991-07-19 | 2000-08-08 | Hatton; Charles Malcolm | Computer system program for creating new ideas and solving problems |
| CN101593206A (en) * | 2009-06-25 | 2009-12-02 | 腾讯科技（深圳）有限公司 | Searching method and device based on answer in the question and answer interaction platform |
| CN102637192A (en) * | 2012-02-17 | 2012-08-15 | 清华大学 | Method for answering with natural language |
| CN105930452A (en) * | 2016-04-21 | 2016-09-07 | 北京紫平方信息技术股份有限公司 | Smart answering method capable of identifying natural language |
| CN107609101A (en) * | 2017-09-11 | 2018-01-19 | 远光软件股份有限公司 | Intelligent interactive method, equipment and storage medium |
| CN108920604A (en) * | 2018-06-27 | 2018-11-30 | 百度在线网络技术（北京）有限公司 | Voice interactive method and equipment |
Non-Patent Citations (2)
| Title |
|---|
| A. CLEMENTEENA 等: ""A literature survey on question answering system in natural"", 《INTERNATIONAL JOURNAL OF ENGINEERING & TECHNOLOGY》 * |
| 王慧慧: ""基于自然语言处理的问答系统研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111681647A (en) * | 2020-06-10 | 2020-09-18 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for recognizing word slot |
| CN111681647B (en) * | 2020-06-10 | 2023-09-05 | 北京百度网讯科技有限公司 | Method, apparatus, device and storage medium for identifying word slots |
| CN113035179A (en) * | 2021-03-03 | 2021-06-25 | 科大讯飞股份有限公司 | Voice recognition method, device, equipment and computer readable storage medium |
| CN113035179B (en) * | 2021-03-03 | 2023-09-26 | 中国科学技术大学 | Voice recognition method, device, equipment and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110020429B (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109657054B (en) | Abstract generation method, device, server and storage medium | |
| CN108847241B (en) | Method for recognizing conference voice as text, electronic device and storage medium | |
| CN110276023B (en) | POI transition event discovery method, device, computing equipment and medium | |
| CN110148416A (en) | Audio recognition method, device, equipment and storage medium | |
| US20170206897A1 (en) | Analyzing textual data | |
| CN109522419B (en) | Session information completion method and device | |
| CN110232129B (en) | Scene error correction method, device, equipment and storage medium | |
| CN105931644A (en) | Voice recognition method and mobile terminal | |
| CN109635270A (en) | Two-way probabilistic natural language is rewritten and selection | |
| CN108288468A (en) | Audio recognition method and device | |
| CN109408622A (en) | Sentence processing method and its device, equipment and storage medium | |
| CN101815996A (en) | Detect name entities and neologisms | |
| CN110415679B (en) | Voice error correction method, device, equipment and storage medium | |
| CN110347790B (en) | Text duplicate checking method, device and equipment based on attention mechanism and storage medium | |
| CN106570180A (en) | Artificial intelligence based voice searching method and device | |
| JP2020004382A (en) | Method and device for voice interaction | |
| CN114556328A (en) | Data processing method and device, electronic equipment and storage medium | |
| CN111488468A (en) | Geographic information knowledge point extraction method and device, storage medium and computer equipment | |
| CN114154487A (en) | Text automatic error correction method and device, electronic equipment and storage medium | |
| CN105161095A (en) | Method and device for picture composition of speech recognition syntax tree | |
| CN109710732A (en) | Information query method, device, storage medium and electronic equipment | |
| CN108153875B (en) | Corpus processing method and device, intelligent sound box and storage medium | |
| CN111508497B (en) | Speech recognition method, device, electronic equipment and storage medium | |
| CN110020429A (en) | Method for recognizing semantics and equipment | |
| CN108268443B (en) | Method and device for determining topic point transfer and acquiring reply text |
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 | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20211103 Address after: 100176 101, floor 1, building 1, yard 7, Ruihe West 2nd Road, Beijing Economic and Technological Development Zone, Daxing District, Beijing Applicant after: Apollo Zhilian (Beijing) Technology Co.,Ltd. Address before: 100085 Baidu Building, 10 Shangdi Tenth Street, Haidian District, Beijing Applicant before: BAIDU ONLINE NETWORK TECHNOLOGY (BEIJING) Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |