CN110020429B - Semantic recognition method and device - Google Patents

Semantic recognition method and device Download PDF

Info

Publication number
CN110020429B
CN110020429B CN201910147719.0A CN201910147719A CN110020429B CN 110020429 B CN110020429 B CN 110020429B CN 201910147719 A CN201910147719 A CN 201910147719A CN 110020429 B CN110020429 B CN 110020429B
Authority
CN
China
Prior art keywords
nlp result
nlp
result
keywords
answer sentence
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910147719.0A
Other languages
Chinese (zh)
Other versions
CN110020429A (en
Inventor
于盛进
尹健刚
揭朋朋
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Apollo Intelligent Connectivity Beijing Technology Co Ltd
Original Assignee
Apollo Intelligent Connectivity Beijing Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Apollo Intelligent Connectivity Beijing Technology Co Ltd filed Critical Apollo Intelligent Connectivity Beijing Technology Co Ltd
Priority to CN201910147719.0A priority Critical patent/CN110020429B/en
Publication of CN110020429A publication Critical patent/CN110020429A/en
Application granted granted Critical
Publication of CN110020429B publication Critical patent/CN110020429B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/332—Query formulation
    • G06F16/3329—Natural language query formulation or dialogue systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/30—Semantic analysis
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/26—Speech to text systems

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Mathematical Physics (AREA)
  • Human Computer Interaction (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Machine Translation (AREA)

Abstract

The embodiment of the invention provides a semantic recognition method and equipment, wherein the method comprises the steps of processing current inquiry information through an NLP technology to obtain a first NLP result; wherein the first NLP result comprises a type, keywords and actions; judging whether a second NLP result of the previous query information has an answer sentence with practical meaning or not; if the answer sentence with the practical meaning exists, carrying out replacement operation or bit filling operation according to the first NLP result and the second NLP result to obtain a third NLP result; and obtaining a corresponding answer sentence according to the third NLP result. The embodiment of the invention can more accurately explain the semantics of the user based on the recognition result generated by the context semantics, is suitable for speaking habits (such as the use of pronouns, the use of abbreviated sentences and the like) in the communication between people, and is convenient for providing correct feedback for the user in the human-computer conversation process and improving the user experience.

Description

Semantic recognition method and device
Technical Field
The embodiment of the invention relates to the technical field of artificial intelligence, in particular to a semantic recognition method and semantic recognition equipment.
Background
The semantic recognition is a branch of artificial intelligence, and the semantic recognition technology can analyze big bright data in webpages, files, mails, audios, forums and social media, has wide application fields, and can be directly applied to industries such as medical treatment, education, finance and the like. The method can also be applied to all intelligent voice interaction scenes through a technical interface, such as intelligent home, vehicle-mounted voice, wearable equipment, VR, robot and the like, and can also be divided into: questions and answers, knowledge retrieval, classification questions, and the like. Intelligent voice interactions are seen as the most promising application scenario in artificial intelligence technology.
In the existing semantic recognition scheme based on context, a fixed sentence pattern of weather inquiry is recognized, and if the subsequent occurrence time or city name occurs, weather information corresponding to the time or city is generated.
However, the above scheme lacks flexibility in terms of the field Jing Shanyi and the fixed question pattern, and cannot adapt to the speaking habit of the person (e.g., use of pronouns, use of abbreviated sentences, etc.).
Disclosure of Invention
The embodiment of the invention provides a semantic recognition method and a semantic recognition device, which are used for improving the flexibility of semantic recognition and adapting to the speaking habit of a person.
In a first aspect, an embodiment of the present invention provides a semantic recognition method, including:
processing the current query information through an NLP technology to obtain a first NLP result; wherein the first NLP result comprises a type, keywords and actions;
judging whether a second NLP result of the previous query information has an answer sentence with practical meaning or not;
if the answer sentence with the practical meaning exists, carrying out replacement operation or bit filling operation according to the first NLP result and the second NLP result to obtain a third NLP result;
and obtaining a corresponding answer sentence according to the third NLP result.
In one possible design, the performing a replacement operation or a bit filling operation according to the first NLP result and the second NLP result includes:
Judging whether directional pronouns or nouns exist in the current query information according to the first NLP result;
if only nouns exist, performing replacement operation according to the first NLP result and the second NLP result;
if directional pronouns exist, judging whether types and actions in the first NLP result are defined according to the first NLP result; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result.
In one possible design, before determining whether only nouns exist in the current query information according to the first NLP result, the method further includes:
judging whether the types of the keywords in the first NLP result are the same as the types of the keywords in the second NLP result;
the judging whether only nouns exist in the current query information according to the first NLP result comprises the following steps:
if the types of the keywords in the first NLP result and the second NLP result are the same, judging whether only nouns exist in the current query information according to the first NLP result.
In one possible design, the performing a replacing operation according to the first NLP result and the second NLP result includes:
And replacing the keywords in the second NLP result with the keywords in the first NLP result.
In one possible design, the determining whether the directional pronoun exists in the current query information according to the first NLP result includes:
judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
In one possible design, the performing a bit filling operation according to the first NLP result and the second NLP result includes:
obtaining content corresponding to the keywords in the first NLP result according to the answer sentence corresponding to the second NLP result;
and deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
In one possible design, after the determining whether the answer sentence having the practical meaning exists in the second NLP result of the previous query information, the method further includes:
and if no answer sentence with practical meaning exists, obtaining a corresponding answer sentence according to the first NLP result.
In a second aspect, an embodiment of the present invention provides a semantic recognition apparatus, including:
The processing module is used for processing the current inquiry information through an NLP technology to obtain a first NLP result; wherein the first NLP result comprises a type, keywords and actions;
the judging module is used for judging whether an answer sentence with practical significance exists in a second NLP result of the previous query information;
the operation module is used for carrying out replacement operation or bit filling operation according to the first NLP result and the second NLP result if the answer sentence with the practical meaning exists, so as to obtain a third NLP result;
and the feedback module is used for obtaining a corresponding answer sentence according to the third NLP result.
In one possible design, the operation module includes:
the judging unit is used for judging whether directional pronouns or nouns exist in the current query information or not according to the first NLP result if answer sentences with practical meanings exist;
a replacing operation unit, configured to perform a replacing operation according to the first NLP result and the second NLP result if only nouns exist;
the bit filling operation unit is used for judging whether the type and the action in the first NLP result are defined according to the first NLP result if directional pronouns exist; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result.
In one possible design, the judging unit is specifically configured to:
judging whether the types of the keywords in the first NLP result are the same as the types of the keywords in the second NLP result; if so, judging whether the current inquiry information only has nouns.
In one possible design, the replacement operating unit is specifically configured to:
and replacing the keywords in the second NLP result with the keywords in the first NLP result to obtain the third NLP result, and obtaining corresponding answer sentences according to the third NLP result.
In one possible design, the judging unit is specifically configured to:
judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
In one possible design, the bit-filling operation unit is specifically configured to:
obtaining content corresponding to the keywords in the first NLP result according to the answer sentence corresponding to the second NLP result;
and deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
In one possible design, the feedback module is further configured to obtain a corresponding answer sentence according to the first NLP result if the second NLP result does not have an answer sentence with a practical meaning.
In a third aspect, an embodiment of the present invention provides a semantic recognition apparatus, including: at least one processor and memory;
the memory stores computer-executable instructions;
the at least one processor executes the computer-executable instructions stored by the memory such that the at least one processor performs the method as described above in the first aspect and the various possible designs of the first aspect.
In a fourth aspect, embodiments of the present invention provide a computer-readable storage medium having stored therein computer-executable instructions which, when executed by a processor, implement the method as described in the first aspect and the various possible designs of the first aspect.
The semantic recognition method and the device provided by the embodiment process the current query information by adopting an NLP technology and generate a first NLP result; wherein the first NLP result comprises a type, keywords and actions; judging whether a second NLP result corresponding to the previous query information has an answer sentence with practical meaning or not; and if the answer sentence with the practical meaning exists, generating a final identification result of the current query information through a replacement operation or a bit filling operation according to the first NLP result and the second NLP result. The method and the device can more accurately explain the semantics of the user based on the recognition result generated by the context semantics, adapt to speaking habits (such as the use of pronouns, the use of abbreviated sentences and the like) in the communication between people, and are convenient for providing correct feedback for the user in the human-computer conversation process and improving the user experience.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings that are needed in the embodiments or the description of the prior art will be briefly described below, it will be obvious that the drawings in the following description are some embodiments of the present invention, and that other drawings can be obtained according to these drawings without inventive effort to a person skilled in the art.
FIG. 1 is a schematic diagram of a human-computer voice interaction system according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention;
FIG. 3 is a flowchart illustrating a semantic recognition method according to another embodiment of the present invention;
fig. 4 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention;
FIG. 5 is a flowchart illustrating a semantic recognition method according to another embodiment of the present invention;
FIG. 6 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention;
FIG. 7 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention;
fig. 8 is a schematic hardware structure of a semantic recognition device according to an embodiment of the present invention.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 is a schematic architecture diagram of a human-computer voice interaction system according to an embodiment of the invention. As shown in fig. 1, the system provided in this embodiment includes a terminal 101 and a server 102. The terminal 101 may be a child story machine, a mobile phone, a tablet, a vehicle-mounted terminal, etc. The implementation of the terminal 101 is not particularly limited in this embodiment, as long as the terminal 101 can perform voice interaction with the user.
The voice interaction (Speech Interaction) is based on voice recognition, natural language processing (Natural Language Processing, NLP), voice synthesis and other technologies, and gives the terminal an intelligent human-computer interaction experience of 'listening, speaking and understanding you' under various actual application scenes. The method is suitable for multiple application scenes, including intelligent question-answering, intelligent playing, intelligent searching and other scenes.
The terminal 101 receives voice information input by a user, performs semantic recognition on the voice information, and then obtains an answer sentence to feed back to the user. Specifically, the terminal 101 may obtain the answer sentence locally according to the corpus stored in the terminal 101, or may send the voice information to the server 102, perform semantic recognition by the server 102, obtain the answer sentence, and then feed back to the user through the terminal 101. If the semantic recognition is performed at the terminal, the terminal can be used as the semantic recognition device, and if the semantic recognition is performed at the server, the server can be used as the semantic recognition device. The specific implementation of the present embodiment is not particularly limited, and the terminal 101 may obtain the answer sentence locally and the server 102 may obtain the answer sentence according to the voice information.
Fig. 2 is a block diagram of a semantic recognition device according to another embodiment of the present invention. As shown in fig. 2, the apparatus 20 includes: the voice recognition module 201, the processing module 202 and the feedback module 203; the voice recognition module 201 is configured to receive voice information input by a user, recognize the voice information, generate text information, and send the text information to the processing module 202; the processing module 202 is configured to perform NLP processing on the text information to obtain an NLP result, and send the NLP result to the feedback module 203; the feedback module 203 is configured to search whether an answer sentence with an actual meaning exists in a conventional corpus according to the NLP result, if so, feed back the searched answer sentence with an actual meaning to a user, and if not, obtain an spam answer sentence from the spam corpus randomly or according to a predetermined rule, and feed back the spam answer sentence to the user.
In a specific implementation process, the voice recognition module 201 automatically receives voice information, recognizes the voice information, generates text information, and sends the text information to the processing module 202; the processing module 202 performs NLP processing on the text information to obtain an NLP result, and sends the NLP result to the feedback module 203; the feedback module 203 searches whether there is an answer sentence with an actual meaning in a conventional corpus according to the NLP result, if yes, feeds back the found answer sentence with an actual meaning to the user, and if not, obtains the spam answer sentence from the spam corpus randomly or according to a predetermined rule, and feeds back the spam answer sentence to the user.
It follows that the NLP result plays an important role in this process as a basis for finding the answer sentence for the correctness of the fed back answer sentence. However, in the prior art, the processing module 202 only processes the current query information received by the speech recognition module 201, so as to obtain the NLP result. And the situation that the answer sentence fed back according to the NLP result is inappropriate exists, so that the user experience is affected. Based on the above, the embodiment of the invention provides a semantic recognition method to improve the accuracy of semantic recognition.
The semantic recognition method provided by the embodiment of the invention is described in detail below by adopting a specific embodiment.
Fig. 3 is a flowchart of a semantic recognition method according to still another embodiment of the present invention. As shown in fig. 3, the method includes:
s301, processing current inquiry information through an NLP technology to obtain a first NLP result; wherein the first NLP result includes a type, a keyword, and an action.
Alternatively, the execution subject of the method may be an intelligent terminal capable of voice interaction with a person, for example: cell phones, tablets, phone watches, etc.
Alternatively, the query information may be voice information, and may also be text information input by the user. If the query information is voice information, the processing the current query information by NLP technology may include: performing voice recognition on the current query information to generate text information; and processing the text information through NLP technology.
Optionally, the processing the current query information through NLP technology may include: and marking the part of speech of the current query information. Namely, the process of labeling words in sentences as nouns, verbs, adjectives, adverbs and the like can also carry out keyword (Key) extraction, action (Action) extraction, auxiliary word (Particle) extraction and Type (Type) determination on the part-of-speech labeling results.
For example: the user inputs query information "what is today's weather? The corresponding NLP result is 'type=weather, key1=today, key2=weather, action=query, particle=how' after the query information is processed through NLP technology.
S302, judging whether an answer sentence with practical meaning exists in a second NLP result of the previous query information.
In the present embodiment, the execution order of the steps S301 and S302 is not limited, and the steps S301 and S302 may be executed in parallel or sequentially
In this embodiment, the previous query information refers to query information that is adjacent to the current query information and is input by the user before the current query information.
For example: the following man-machine conversation content is aimed at:
people: "what is today? "
And (3) machine: "today's weather is sunny and cloudy. "
People: "tomorrow? "
If "tomorrow" is used as the current inquiry information, "how weather today is" is the previous inquiry information.
In this embodiment, the answer sentence with practical meaning may be that the intelligent terminal for man-machine interaction can find the corresponding answer sentence in the conventional corpus. Correspondingly, the answer sentence without practical meaning can be the corresponding answer sentence which cannot be found in the conventional corpus, and the spam answer sentence which is fed back randomly or according to a preset rule from the spam corpus.
For example: aiming at the previous query information 'how much is the weather today', the feedback module of the generated NLP result only can find the corresponding answer sentence 'how much is the weather today' in a clear manner in the conventional corpus, namely the query information 'how much is the weather today' has the answer sentence with practical significance.
Aiming at the NLP result generated by the current query information 'Mingshi', the feedback module of the intelligent terminal cannot find the corresponding answer sentence in the conventional corpus, is defined as spam boring, and needs to feed back meaningless spam answer sentences randomly or according to a preset rule in the spam corpus, namely the query information 'Mingshi' does not have any answer sentence with practical meaning.
And S303, if an answer sentence with practical significance exists, performing replacement operation or bit filling operation according to the first NLP result and the second NLP result to obtain a third NLP result.
If the answer sentence of the previous inquiry information is an answer sentence with practical meaning, but not a nonsensical spam answer sentence distributed for spam, performing substitution or bit filling processing according to the NLP results (the first NLP result and the second NLP result) respectively corresponding to the previous inquiry information and the current inquiry information, so as to obtain a final NLP result for the current inquiry information, namely the third NLP result.
For example, in the man-machine conversation, for the current inquiry information "tomorrow", it is necessary to use the previous inquiry information "how is the weather today? And carrying out replacement operation or bit filling operation on the corresponding second NLP result and the first NLP result corresponding to the current query information 'Mingshun', and obtaining a third NLP result.
The replacing operation may be to replace "today" in the previous query with "tomorrow" in the current query, and the third NLP result may be "what is the weather of tomorrow? "
The filling operation may be to fill in ("how weather is") other than "today" in the previous query to "tomorrow" in the current query, and then the third NLP result may be "what is the tomorrow weather? "
S304, obtaining a corresponding answer sentence according to the third NLP result.
Optionally, the feedback module of the intelligent terminal may be obtained from a local corpus stored in the intelligent terminal according to the third NLP result, or obtained from a server side.
According to the semantic recognition method provided by the embodiment, the current query information is processed by adopting an NLP technology, and a first NLP result is generated; wherein the first NLP result comprises a type, keywords and actions; judging whether a second NLP result corresponding to the previous query information has an answer sentence with practical meaning or not; and if the answer sentence with the practical meaning exists, generating a final identification result of the current query information through a replacement operation or a bit filling operation according to the first NLP result and the second NLP result. The method and the device can more accurately explain the semantics of the user based on the recognition result generated by the context semantics, adapt to speaking habits (such as the use of pronouns, the use of abbreviated sentences and the like) in the communication between people, and are convenient for providing correct feedback for the user in the human-computer conversation process and improving the user experience.
In a specific embodiment, fig. 4 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention, as shown in fig. 4, the semantic recognition device 40 may include: the system comprises a voice recognition module 201, a processing module 202, an NLP result storage module 206, an operation module 205 and a feedback module 203. The voice recognition module 201 is configured to receive voice information input by a user, recognize the voice information, generate text information, and send the text information to the processing module 202. The processing module 202 is configured to perform NLP processing on the text information to obtain an NLP result, and send the NLP result to the context understanding module and the NLP result caching module. The NLP result storage module 206 is configured to store an NLP result corresponding to the previous voice message, and send the NLP result corresponding to the previous voice message to the context understanding module. The judging module 204 is configured to judge whether the second NLP result of the previous query information has an answer sentence with practical meaning, if so, the operating module 205 obtains a third NLP result by a replacement operation or a bit filling operation according to the NLP results corresponding to the current voice information and the previous voice information, and sends the third NLP result to the feedback module 203. And the feedback module 203 is configured to search for an answer sentence with an actual meaning in a conventional corpus according to the third NLP result, and feedback the found answer sentence with the actual meaning to the user. That is, the voice recognition module 201 and the processing module 202 are for performing step S301; the context understanding module is used for executing steps S302 and S303; the feedback module 203 is configured to execute step S304.
Optionally, the processing module 202 may process the query information to obtain an NLP result in Json object numbered musical notation (JavaScript Object Notation, json) format. The NLP result storage module 206 may store NLP results in the form of a list.
In a specific embodiment, the semantic recognition method provided in the foregoing embodiment may further include:
s305, if no answer sentence with practical meaning exists, a corresponding answer sentence is obtained according to the first NLP result.
If the previous voice information does not have the answer sentence with the practical meaning, the fact that the previous voice information is input by the user is indicated, the intelligent terminal can only feed back the spam answer sentence with the non-practical meaning obtained by the user from the spam corpus, and the fact that the previous voice information is not related to the current voice information can be judged, so that the corresponding answer sentence can be obtained directly according to the first NLP result of the current voice information.
Fig. 5 is a flowchart of a semantic recognition method according to still another embodiment of the present invention. As shown in fig. 5, the method includes:
s501, processing current inquiry information through an NLP technology to obtain a first NLP result; wherein the first NLP result includes a type, a keyword, and an action.
S502, judging whether an answer sentence with practical meaning exists in a second NLP result of the previous query information.
Step S501 and step S502 in this embodiment are similar to step S301 and step S302 in the above embodiment, and will not be described here again.
S503, judging whether directional pronouns or nouns exist in the current query information according to the first NLP result.
Pronouns are words that replace nouns and act as nouns. The pronouns may include: the human pronouns (you, me, he, etc.), the anti-pronouns (me own, you own, he own, his own, etc.), the query pronouns (where, who, when, etc.), the indefinite pronouns (some, many, any, each, etc.), the directional pronouns (this, that, here, there, what, etc.).
Optionally, determining whether the directional pronoun exists in the current query information may include: and carrying out word segmentation processing on the current query information, and judging whether each word segment is a directional pronoun. For example: for the query statement "where is that sight? The word segmentation processing is carried out to obtain 'that' scenic spot 'and' what 'place' respectively carries out part-of-speech judgment on each segmented word, and the 'that' can be obtained as a directional pronoun.
In a specific embodiment, the determining, according to the first NLP result, whether a directional pronoun exists in the current query information includes: judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
For example: the following man-machine conversation content is aimed at:
people: "what is a funny tourist attraction in Shenzhen? "
And (3) machine: "Window of the world, speical China folk culture village, shenzhen Happy valley, xianhu vegetable garden, small Meisha and Qingqing world". "
People: help me navigate to first'
Current query information "help me navigate to" type=point of interest (Point of Interest, POI), key1=me, key2=first, action=navigate in the first "corresponding first NLP result". And if 'Key2=first' in the first NLP result is a directional pronoun, judging that the directional pronoun exists in the current inquiry information.
In a specific embodiment, before determining whether only nouns exist in the current query information according to the first NLP result, the method further includes: judging whether the type of the keyword in the first NLP result is the same as the type of the keyword in the second NLP result.
The judging whether only nouns exist in the current query information according to the first NLP result comprises the following steps: if the types of the keywords in the first NLP result and the second NLP result are the same, judging whether only nouns exist in the current query information according to the first NLP result.
For example: the following man-machine conversation content is aimed at:
people: "what is today? "
And (3) machine: "today's weather is sunny and cloudy. "
People: "tomorrow? "
In the first NLP result corresponding to the current query information "tomorrow", the type=boring, the key=tomorrow, the action=broadcasting, the particle=woolen ", in the second NLP result corresponding to the previous query information" how weather today "is, the type=weather, the key1=today, the key2=weather, the action=query, the particle=how. By comparing "key=tomorrow" in the first NLP result with "key1=today" in the second NLP result, it can be obtained that the two keywords belong to the same time noun and are of the same type. Therefore, it is possible to further judge whether only nouns exist in the current query information.
S504, if only nouns exist, performing replacement operation according to the first NLP result and the second NLP result to obtain a third NLP result.
In a specific embodiment, the performing a replacing operation according to the first NLP result and the second NLP result includes: and replacing the keywords in the second NLP result with the keywords in the first NLP result.
For example: in the first NLP result corresponding to the current query information "tomorrow", the type=boring, the key=tomorrow, the action=broadcasting, the particle=woolen ", in the second NLP result corresponding to the previous query information" how weather today "is, the type=weather, the key1=today, the key2=weather, the action=query, the particle=how. Since only the time noun 'tomorrow' exists in the current inquiry information, the tomorrow is replaced by the time noun 'today' in the previous inquiry information, and then the third NLP result corresponding to the current inquiry information obtained by the replacing operation is 'type=weather, key1=tomorrow, key2=weather, action=query, particle=how' i.e. what 'tomorrow' is. And the intelligent terminal searches the answer sentence in the conventional corpus according to the third NLP result.
S505, if directional pronouns exist, judging whether types and actions in the first NLP result are defined according to the first NLP result; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result to obtain a third NLP result.
In a specific embodiment, the performing a bit filling operation according to the first NLP result and the second NLP result includes:
s5051, obtaining contents corresponding to the keywords in the first NLP result according to the answer sentences corresponding to the second NLP result.
S5052, deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
For example:
in the second NLP result corresponding to the previous query information "what is interesting in Shenzhen" the type=poi, key1=Shenzhen, key2=attraction, key3=interesting, action=query ". The feedback module of the intelligent terminal can find out the window of the world, the golden Chinese folk culture village, shenzhen Happy valley, the garden of the Xianhu, the small plum sand and the Qing world of the answer sentence according to the second NLP result. "current query information" helps me navigate to "type=poi, key1=me, key2=first, action=navigate" in the first NLP result corresponding to the first ". Since the directional pronoun is "key2=first", the first sight spot "window of the world" in the answer sentence is complemented to the position of the directional pronoun "first", so as to obtain a third NLP result "type=poi, key1=me, key2=window of the world, action=navigation", namely "help me navigate to window of the world". And the intelligent terminal searches the answer sentence in the conventional corpus according to the third NLP result.
S506, obtaining a corresponding answer sentence according to the third NLP result.
Step S506 in this embodiment is similar to step S304 in the above embodiment, and will not be repeated here.
According to the endpoint detection method provided by the embodiment, whether directional pronouns exist or only nouns exist in the current query information is determined, and replacement operation or bit filling operation is adopted for the current query information and the previous query information, so that incomplete question sentences with directional pronouns or only nouns can be quickly and completely expanded, and intelligent terminals for carrying out semantic recognition can accurately and quickly acquire question and answer sentences. The problem that correct answer sentences cannot be obtained and user experience is affected because incomplete caused by whether directional pronouns exist or only nouns exist in the current query information is avoided, and misjudgment is performed as spam.
Fig. 6 is a schematic structural diagram of a semantic recognition device according to another embodiment of the present invention. As shown in fig. 6, the semantic recognition device 60 includes: a framing module 801, a detection module 802, and a determination module 803.
The processing module 202 is configured to process the current query information through an NLP technology, and obtain a first NLP result; wherein the first NLP result includes a type, a keyword, and an action.
Alternatively, the query information may be voice information, and may also be text information input by the user. If the query information is voice information, the processing the current query information by NLP technology may include: performing voice recognition on the current query information to generate text information; and processing the text information through NLP technology.
Optionally, the processing the current query information through NLP technology may include: and marking the part of speech of the current query information. Namely, the process of labeling words in sentences as nouns, verbs, adjectives, adverbs and the like can also carry out keyword (Key) extraction, action (Action) extraction, auxiliary word (Particle) extraction and Type (Type) determination on the part-of-speech labeling results.
For example: the user inputs query information "what is today's weather? The corresponding NLP result is 'type=weather, key1=today, key2=weather, action=query, particle=how' after the query information is processed through NLP technology.
A judging module 204, configured to judge whether the second NLP result of the previous query information has an answer sentence with a practical meaning.
The previous inquiry information refers to inquiry information which is adjacent to the current inquiry information and is input by a user before the current inquiry information.
For example: the following man-machine conversation content is aimed at:
people: "what is today? "
And (3) machine: "today's weather is sunny and cloudy. "
People: "tomorrow? "
If "tomorrow" is used as the current inquiry information, "how weather today is" is the previous inquiry information.
In this embodiment, the answer sentence with practical meaning may be that the intelligent terminal for man-machine interaction can find the corresponding answer sentence in the conventional corpus. Correspondingly, the answer sentence without practical meaning can be the corresponding answer sentence which cannot be found in the conventional corpus, and the spam answer sentence which is fed back randomly or according to a preset rule from the spam corpus.
For example: aiming at the previous query information 'how much is the weather today', the feedback module of the generated NLP result only can find the corresponding answer sentence 'how much is the weather today' in a clear manner in the conventional corpus, namely the query information 'how much is the weather today' has the answer sentence with practical significance.
Aiming at the NLP result generated by the current query information 'Mingshi', the feedback module of the intelligent terminal cannot find the corresponding answer sentence in the conventional corpus, is defined as spam boring, and needs to feed back meaningless spam answer sentences randomly or according to a preset rule in the spam corpus, namely the query information 'Mingshi' does not have any answer sentence with practical meaning.
And an operation module 205, configured to perform a replacement operation or a bit filling operation according to the first NLP result and the second NLP result if there is an answer sentence with an actual meaning, so as to obtain a third NLP result.
If the answer sentence of the previous inquiry information is an answer sentence with practical meaning, but not a nonsensical spam answer sentence distributed for spam, performing substitution or bit filling processing according to the NLP results (the first NLP result and the second NLP result) respectively corresponding to the previous inquiry information and the current inquiry information, so as to obtain a final NLP result for the current inquiry information, namely the third NLP result.
For example, in the man-machine conversation, for the current inquiry information "tomorrow", it is necessary to use the previous inquiry information "how is the weather today? And carrying out replacement operation or bit filling operation on the corresponding second NLP result and the first NLP result corresponding to the current query information 'Mingshun', and obtaining a third NLP result.
The replacing operation may be to replace "today" in the previous query with "tomorrow" in the current query, and the third NLP result may be "what is the weather of tomorrow? "
The filling operation may be to fill in ("how weather is") other than "today" in the previous query to "tomorrow" in the current query, and then the third NLP result may be "what is the tomorrow weather? "
And the feedback module 203 is configured to obtain a corresponding answer sentence according to the third NLP result.
Alternatively, the feedback module 203 may obtain, according to the third NLP result, from a local corpus stored in the intelligent terminal, or from a server side.
According to the semantic recognition device provided by the embodiment of the invention, the processing module 202 processes the current query information by adopting an NLP technology and generates a first NLP result; wherein the first NLP result comprises a type, keywords and actions; the judging module 204 judges whether the second NLP result corresponding to the previous query information has an answer sentence with practical meaning; if there is an answer sentence with practical meaning, the operation module 205 generates a final recognition result of the current query information through a replacement operation or a bit filling operation according to the first NLP result and the second NLP result. The method and the device can more accurately explain the semantics of the user based on the recognition result generated by the context semantics, adapt to speaking habits (such as the use of pronouns, the use of abbreviated sentences and the like) in the communication between people, and are convenient for providing correct feedback for the user in the human-computer conversation process and improving the user experience.
Fig. 7 is a schematic structural diagram of a semantic recognition device according to still another embodiment of the present invention. As shown in fig. 7, the operation module 205 of the semantic recognition device 70 includes:
a judging unit 2051, configured to judge whether a directional pronoun or only a noun exists in the current query information according to the first NLP result if there is an answer sentence with an actual meaning;
a replacing operation unit 2052, configured to perform a replacing operation according to the first NLP result and the second NLP result if only nouns exist;
the bit filling operation unit 2053 is configured to determine, if a directional pronoun exists, whether a type and an action in the first NLP result are defined according to the first NLP result; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result.
Optionally, the judging unit is specifically configured to: judging whether the types of the keywords in the first NLP result are the same as the types of the keywords in the second NLP result; if so, judging whether the current inquiry information only has nouns.
Optionally, the replacement operation unit is specifically configured to: and replacing the keywords in the second NLP result with the keywords in the first NLP result to obtain the third NLP result, and obtaining corresponding answer sentences according to the third NLP result.
Optionally, the judging unit is specifically configured to: judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
Optionally, the bit filling operation unit is specifically configured to: obtaining content corresponding to the keywords in the first NLP result according to the answer sentence corresponding to the second NLP result; and deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
Optionally, the feedback module is further configured to: and if the second NLP result does not have the answer sentence with the practical meaning, obtaining the corresponding answer sentence according to the first NLP result.
The endpoint detection apparatus provided in the embodiment of the present invention may be used to execute the above method embodiment, and its implementation principle and technical effects are similar, and this embodiment will not be described herein.
Fig. 8 is a schematic hardware structure of a semantic recognition device according to an embodiment of the present invention. As shown in fig. 8, the semantic recognition device 80 provided in the present embodiment includes: at least one processor 801 and a memory 802. The processor 801 and the memory 802 are connected by a bus 803.
Optionally, the semantic recognition device may further comprise a communication component, which is connected to the processor 801 and the memory 802 via a bus 803.
In a specific implementation, the at least one processor 801 executes computer-executable instructions stored in the memory 802, so that the at least one processor 801 performs the semantic recognition method as performed by the semantic recognition device 80.
When the semantic recognition method of the present embodiment is executed by the server, the communication section may transmit the query information to the server and receive the answer sentence fed back from the server.
The specific implementation process of the processor 801 may refer to the above-mentioned method embodiment, and its implementation principle and technical effects are similar, and this embodiment will not be described herein again.
In the embodiment shown in fig. 8, it should be understood that the processor may be a central processing unit (english: central Processing Unit, abbreviated as CPU), or may be other general purpose processors, digital signal processors (english: digital Signal Processor, abbreviated as DSP), application specific integrated circuits (english: application Specific Integrated Circuit, abbreviated as ASIC), or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with the present invention may be embodied directly in a hardware processor for execution, or in a combination of hardware and software modules in a processor for execution.
The memory may comprise high speed RAM memory or may further comprise non-volatile storage NVM, such as at least one disk memory.
The bus may be an industry standard architecture (Industry Standard Architecture, ISA) bus, an external device interconnect (Peripheral Component, PCI) bus, or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, among others. The buses may be divided into address buses, data buses, control buses, etc. For ease of illustration, the buses in the drawings of the present application are not limited to only one bus or one type of bus.
The application also provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and when a processor executes the computer execution instructions, the semantic recognition method executed by the semantic recognition device is realized.
The application also provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and when a processor executes the computer execution instructions, the semantic recognition method executed by the semantic recognition device is realized.
The computer readable storage medium described above may be implemented by any type of volatile or non-volatile memory device or combination thereof, such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk. A readable storage medium can be any available medium that can be accessed by a general purpose or special purpose computer.
An exemplary readable storage medium is coupled to the processor such the processor can read information from, and write information to, the readable storage medium. In the alternative, the readable storage medium may be integral to the processor. The processor and the readable storage medium may reside in an application specific integrated circuit (Application Specific Integrated Circuits, ASIC for short). The processor and the readable storage medium may reside as discrete components in a device.
Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the method embodiments described above may be performed by hardware associated with program instructions. The foregoing program may be stored in a computer readable storage medium. The program, when executed, performs steps including the method embodiments described above; and the aforementioned storage medium includes: various media that can store program code, such as ROM, RAM, magnetic or optical disks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention.

Claims (16)

1. A semantic recognition method, comprising:
processing the current query information through a natural language processing NLP technology to obtain a first NLP result; wherein the first NLP result comprises a type, keywords and actions;
judging whether a second NLP result of the previous query information has an answer sentence with practical meaning or not; the answer sentences with practical significance are answer sentences which can be found in a conventional corpus;
if the answer sentence with the practical meaning exists, carrying out replacement operation or bit filling operation according to the first NLP result and the second NLP result to obtain a third NLP result; obtaining a corresponding answer sentence according to the third NLP result;
the performing the replacing operation or the bit filling operation according to the first NLP result and the second NLP result includes:
replacing keywords with the same keyword type as the keywords in the first NLP result in the second NLP result with the keywords in the first NLP result; or,
filling the content corresponding to the keyword in the first NLP result in the answer sentence corresponding to the second NLP result to the position of the keyword in the first NLP result; or,
And supplementing the content of the second NLP result except for the keywords with the same keyword type as the keywords in the first NLP result to the first NLP result.
2. The method of claim 1, wherein the performing a replacement operation or a fill operation based on the first NLP result and the second NLP result comprises:
judging whether directional pronouns or nouns exist in the current query information according to the first NLP result;
if only nouns exist, performing replacement operation according to the first NLP result and the second NLP result;
if directional pronouns exist, judging whether types and actions in the first NLP result are defined according to the first NLP result; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result.
3. The method of claim 2, wherein the determining whether only nouns are present in the current query based on the first NLP result further comprises:
judging whether the types of the keywords in the first NLP result are the same as the types of the keywords in the second NLP result;
the judging whether only nouns exist in the current query information according to the first NLP result comprises the following steps:
If the types of the keywords in the first NLP result and the second NLP result are the same, judging whether only nouns exist in the current query information according to the first NLP result.
4. The method of claim 2, wherein the determining whether a directional pronoun exists in the current query information according to the first NLP result comprises:
judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
5. The method of claim 1, wherein the filling content corresponding to the keyword in the first NLP result in the answer sentence corresponding to the second NLP result to the position of the keyword in the first NLP result comprises:
and deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
6. The method as recited in claim 5, further comprising:
and obtaining the content corresponding to the keyword in the first NLP result according to the answer sentence corresponding to the second NLP result.
7. The method according to any one of claims 1-6, wherein after determining whether the answer sentence having the actual meaning exists in the second NLP result of the previous query information, further comprising:
And if no answer sentence with practical meaning exists, obtaining a corresponding answer sentence according to the first NLP result.
8. A semantic recognition apparatus, comprising:
the processing module is used for processing the current query information through a natural language processing NLP technology to obtain a first NLP result; wherein the first NLP result comprises a type, keywords and actions;
the judging module is used for judging whether an answer sentence with practical significance exists in a second NLP result of the previous query information; the answer sentences with practical significance are answer sentences which can be found in a conventional corpus;
the operation module is used for carrying out replacement operation or bit filling operation according to the first NLP result and the second NLP result if the answer sentence with the practical meaning exists, so as to obtain a third NLP result;
the feedback module is used for obtaining a corresponding answer sentence according to the third NLP result;
the operation module is specifically configured to:
replacing keywords with the same keyword type as the keywords in the first NLP result in the second NLP result with the keywords in the first NLP result; or,
filling the content corresponding to the keyword in the first NLP result in the answer sentence corresponding to the second NLP result to the position of the keyword in the first NLP result; or,
And supplementing the content of the second NLP result except for the keywords with the same keyword type as the keywords in the first NLP result to the first NLP result.
9. The apparatus of claim 8, wherein the operation module comprises:
the judging unit is used for judging whether directional pronouns or nouns exist in the current query information or not according to the first NLP result if answer sentences with practical meanings exist;
a replacing operation unit, configured to perform a replacing operation according to the first NLP result and the second NLP result if only nouns exist;
the bit filling operation unit is used for judging whether the type and the action in the first NLP result are defined according to the first NLP result if directional pronouns exist; and if the first NLP result and the second NLP result are defined, performing bit filling operation according to the first NLP result and the second NLP result.
10. The apparatus according to claim 9, wherein the judging unit is specifically configured to:
judging whether the types of the keywords in the first NLP result are the same as the types of the keywords in the second NLP result; if so, judging whether the current inquiry information only has nouns.
11. The apparatus according to claim 9, wherein the judging unit is specifically configured to:
judging whether the keyword in the first NLP result is a directional pronoun, if so, judging that the directional pronoun exists in the current inquiry information.
12. The apparatus according to claim 9, wherein the bit-filling operation unit is specifically configured to:
and deleting the keywords in the first NLP result, and filling the content into the keywords of the first NLP result.
13. The apparatus according to claim 12, characterized in that the bit-filling operation unit is specifically configured to:
and obtaining the content corresponding to the keyword in the first NLP result according to the answer sentence corresponding to the second NLP result.
14. The apparatus of any one of claims 8 to 13, wherein the feedback module is further configured to: and if the second NLP result does not have the answer sentence with the practical meaning, obtaining the corresponding answer sentence according to the first NLP result.
15. A semantic recognition apparatus, comprising: at least one processor and memory;
the memory stores computer-executable instructions;
the at least one processor executing computer-executable instructions stored in the memory causes the at least one processor to perform the semantic recognition method of any one of claims 1 to 7.
16. A computer-readable storage medium, in which computer-executable instructions are stored, which, when executed by a processor, implement the semantic recognition method according to any one of claims 1 to 7.
CN201910147719.0A 2019-02-27 2019-02-27 Semantic recognition method and device Active CN110020429B (en)

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 CN110020429A (en) 2019-07-16
CN110020429B true 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)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113488036A (en) * 2020-06-10 2021-10-08 海信集团有限公司 Multi-round voice interaction method, terminal and server
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
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

Citations (6)

* Cited by examiner, † Cited by third party
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

Patent Citations (6)

* Cited by examiner, † Cited by third party
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)

* Cited by examiner, † Cited by third party
Title
"A literature survey on question answering system in natural";A. Clementeena 等;《International Journal of Engineering & Technology》;20180630;第7卷(第2期);452-455 *
"基于自然语言处理的问答系统研究";王慧慧;《中国优秀硕士学位论文全文数据库 信息科技辑》;20061215;I138-1564 *

Also Published As

Publication number Publication date
CN110020429A (en) 2019-07-16

Similar Documents

Publication Publication Date Title
CN108287858B (en) Semantic extraction method and device for natural language
JP6667504B2 (en) Orphan utterance detection system and method
CN112417102B (en) Voice query method, device, server and readable storage medium
CN111625635A (en) Question-answer processing method, language model training method, device, equipment and storage medium
CN110276023B (en) POI transition event discovery method, device, computing equipment and medium
AU2017408800B2 (en) Method and system of mining information, electronic device and readable storable medium
CN101815996A (en) Detect name entities and neologisms
CN110457449B (en) Method, device, equipment and storage medium for training model online
CN108345612B (en) Problem processing method and device for problem processing
CN110020429B (en) Semantic recognition method and device
CN107112009B (en) Method, system and computer-readable storage device for generating a confusion network
JP2020004382A (en) Method and device for voice interaction
CN111046060A (en) Data retrieval method, device, equipment and medium based on elastic search
US11947909B2 (en) Training a language detection model for language autodetection from non-character sub-token signals
CN112686051A (en) Semantic recognition model training method, recognition method, electronic device, and storage medium
CN107424612B (en) Processing method, apparatus and machine-readable medium
CN111428011A (en) Word recommendation method, device, equipment and storage medium
CN111984774A (en) Search method, device, equipment and storage medium
CN111984775A (en) Question and answer quality determination method, device, equipment and storage medium
CN111508497B (en) Speech recognition method, device, electronic equipment and storage medium
CN111555960A (en) Method for generating information
CN114242047A (en) Voice processing method and device, electronic equipment and storage medium
CN108268443B (en) Method and device for determining topic point transfer and acquiring reply text
CN111444321A (en) Question answering method, device, electronic equipment and storage medium
US10963641B2 (en) Multi-lingual tokenization of documents and associated queries

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
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 Intelligent Connectivity (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.

GR01 Patent grant
GR01 Patent grant