WO2024106626A1 - Système de questions et de réponses interactif basé sur l'intelligence artificielle - Google Patents

Système de questions et de réponses interactif basé sur l'intelligence artificielle Download PDF

Info

Publication number
WO2024106626A1
WO2024106626A1 PCT/KR2023/002677 KR2023002677W WO2024106626A1 WO 2024106626 A1 WO2024106626 A1 WO 2024106626A1 KR 2023002677 W KR2023002677 W KR 2023002677W WO 2024106626 A1 WO2024106626 A1 WO 2024106626A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
query
response
inquiry
question
Prior art date
Application number
PCT/KR2023/002677
Other languages
English (en)
Korean (ko)
Inventor
이경일
방그린
Original Assignee
주식회사 솔트룩스
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 주식회사 솔트룩스 filed Critical 주식회사 솔트룩스
Publication of WO2024106626A1 publication Critical patent/WO2024106626A1/fr

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
    • 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
    • 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
    • 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
    • G06F16/3325—Reformulation based on results of preceding query
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/3332—Query translation
    • G06F16/3338—Query expansion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3343—Query execution using phonetics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/33—Querying
    • G06F16/3331—Query processing
    • G06F16/334—Query execution
    • G06F16/3344—Query execution using natural language analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/43—Querying
    • G06F16/432—Query formulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/43—Querying
    • G06F16/432—Query formulation
    • G06F16/433—Query formulation using audio data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/268—Morphological analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/20—Natural language analysis
    • G06F40/279—Recognition of textual entities
    • G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
    • 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
    • G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
    • 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

Definitions

  • the present invention relates to an artificial intelligence-based interactive question-answering system. Specifically, it provides an artificial intelligence-based interactive question-answering technology that accurately understands the user's query intent and quickly provides a response to the query intent. It's related to that.
  • the present invention provides an artificial intelligence-based interactive question answering technology that accurately understands the user's query intention and quickly provides a response to the query intention.
  • the purpose is to request a re-inquiry with a more specific query phrase from the user so that the user can obtain the desired response.
  • an artificial intelligence-based interactive question and answer system implemented with a computing device including one or more processors and one or more memories that store instructions executable by the processor according to an embodiment of the present invention.
  • a query receiver that receives the user's query in natural language based on speech recognition technology (Speech to text, STT);
  • a response possibility determination unit that analyzes and refines the user's query received in natural language through a natural language preprocessing model to determine the possibility of responding to the user's inquiry;
  • a response providing unit that provides a response to the user's inquiry using a preset artificial intelligence algorithm when the response possibility determination unit analyzes that the possibility of responding to the user's inquiry is greater than a threshold standard;
  • a material request unit that requests the user to re-inquire about the previously received user's inquiry.
  • the possibility of response to the user's inquiry is less than the threshold standard, but the basis for determining that the possibility of response is less than the threshold standard is that two or more query intentions are analyzed in the user's query, and two or more query intentions are analyzed. If the correlation between query intentions is determined to be less than the threshold, the material request unit presents two or more analyzed query intentions to the user and requests the user to select one of the two or more query intentions. It is desirable to allow the user to re-inquire about the previously received user's inquiry.
  • the material request unit generates a list of queries corresponding to responses that exist in excess of the preset number as a recommended query list and provides it to the user, allowing the user to request materials for the previously received user's query. It is desirable to ensure that .
  • the above-described material request unit provides a different identification code to the user when providing a plurality of queries included in the recommended query list, and provides a different identification code to each query phrase, and the query receiver performs the function of the material request unit.
  • a sub-query based on a recommended query list is performed from a user, it is desirable to receive and process the user's sub-query by means including at least one of a query phrase and an identification code assigned to the query phrase.
  • the above-described response providing unit when providing a response to a user's inquiry, provides a first response, which is an artificial intelligence algorithm-based response, and a text message from a database that manages the history of query responses performed among multiple users on a preset website. It is desirable to additionally provide a second response, which is a response extracted through mining analysis.
  • the inquiry receiver is installed on the user interface of the user terminal linked to the preset network based on voice recognition technology. By transmitting the content of the recognized user's query, it is desirable to correct the content of the user's query transmitted to the user interface using the input means of the user terminal.
  • the above-mentioned response possibility determination unit preferably includes a morpheme analysis unit, a syntax analysis unit, a semantic analysis unit, and a pragmatic analysis unit in order to analyze the user's query received in natural language form through a natural language pre-processing model.
  • the above-described artificial intelligence-based interactive question and answer system preferably further includes a question and answer management unit that manages user inquiries and responses to user inquiries.
  • the above-mentioned question and answer management unit matches and manages the user's query initially received in the query receiver and the query phrase derived by analyzing and refining the user's inquiry through a natural language preprocessing model in the response possibility analysis unit. It is desirable to analyze the similarity between phrases and manage the user's query data set used to derive query phrases showing similarity greater than the critical similarity as query data related to the query phrase.
  • an artificial intelligence-based interactive question answering technology that accurately understands the user's inquiry intention and quickly provides a response to the inquiry intention, particularly in the user's inquiry If the intent of more than one query is identified, or if the user's query is abstract and there is a possibility that an excessive number of responses will be provided, we request a re-inquiry with a more specific query phrase from the user so that the user can obtain the desired response. This has the effect of increasing satisfaction with the use of the question-and-answer system.
  • the present invention provides a response to a user's inquiry as a first response containing expert knowledge extracted from an expert knowledge database and a second response containing experiential knowledge based on the users' experience. , It has the effect of being able to implement a differentiated question and answer service.
  • FIG 1 and 2 are diagrams of the configuration of an artificial intelligence-based interactive question and answer system according to an embodiment of the present invention.
  • 3 and 4 are examples of a material inquiry being performed by a material request unit according to the analysis results of the response possibility analysis unit according to an embodiment of the present invention.
  • Figure 5 is a conceptual diagram of requesting direct correction of the query to the user terminal when analysis of the user's query fails according to an embodiment of the present invention.
  • Figure 6 is a conceptual diagram of matching and managing query data sets and query phrases in the question and answer management unit according to an embodiment of the present invention.
  • FIG. 7 is an example of the internal configuration of a computing device according to an embodiment of the present invention.
  • first, second, etc. may be used to describe various components, but the components are not limited by the terms. The above terms are used only for the purpose of distinguishing one component from another. For example, a first component may be referred to as a second component, and similarly, the second component may be referred to as a first component without departing from the scope of the present invention.
  • the term and/or includes any of a plurality of related stated items or a combination of a plurality of related stated items.
  • the present invention relates to an artificial intelligence-based interactive question-answering system, and specifically, to provide an artificial intelligence-based interactive question-answering technology that accurately understands the user's query intent and quickly provides a response to the query intent.
  • an artificial intelligence-based interactive question-answering technology that accurately understands the user's query intent and quickly provides a response to the query intent.
  • the user can be asked to re-inquire with a more specific query phrase and receive the response desired by the user.
  • the purpose is to enable you to obtain.
  • the main component of the artificial intelligence-based interactive question and answer system 10 of the present invention is a natural language form based on speech recognition technology (Speech to text, STT). It includes a query receiving unit 11 that receives a user's query.
  • the above-mentioned query receiver 11 functions to receive a user's query in the form of natural language by having a computer interpret the voice language spoken by a person using known voice recognition technology and convert it into text data.
  • the user's inquiry received in natural language is analyzed and refined through a natural language preprocessing model to determine the possibility of responding to the user's inquiry. It includes a response possibility determination unit 12 that determines the possibility of response.
  • the above-described response possibility determination unit 12 includes a morpheme analysis unit 121 and a syntax analysis unit to analyze the user's query received in natural language through a natural language preprocessing model. (122), a semantic analysis unit 123, and a pragmatic analysis unit 124 may be included.
  • the above-described morpheme analysis unit 121 understands that when a user's query is converted into text data by the above-mentioned voice recognition technology, a process of analyzing the converted text data and separating it into minimum semantic units called morphemes is performed. It can be.
  • the morpheme analysis unit 121 described above performs the function of identifying the linguistic structure in the user's query converted to text data and distinguishing the root, prefix/suffix, and part of speech (POS). As an example, When a user query, ‘What is the weather in Seoul today?’ is received, the morpheme analysis unit 121 separates today, Seoul, and weather as roots (substantive morphemes), and uses ⁇ of and ⁇ as suffixes (grammatical morphemes). After separation, it can be used to analyze the user's query intention using the root corresponding to the actual morpheme.
  • the above-mentioned syntax analysis unit 122 performs the function of analyzing (or parsing) the structure of the sentence. To put it simply, the sentence components such as subject, verb, and object of the user's query received by the query receiver 11 It can be understood that the function of analyzing the sentence structure according to sentence components is performed by determining whether or not it contains.
  • the semantic analysis unit 123 described above functions to analyze the specific meaning of the user's query based on the results generated by the morpheme analysis unit 121 and the syntactic analysis unit 122, and
  • the pragmatic analysis unit 124 functions to understand the user's query intention by interpreting sentences through knowledge related to language use.
  • the functions of the morpheme analysis unit 121, syntax analysis unit 122, semantic analysis unit 123, and pragmatic analysis unit 124 are performed sequentially, or some are performed sequentially and some are performed in parallel. It may be possible, and the present invention is not limited thereto.
  • the response possibility determination unit 12 analyzes and refines the user's query through a natural language preprocessing model as described above, and then determines the possibility of responding to the user's query.
  • the determination of the possibility of responding as described above is made when the user's inquiry intention (for example, less than two inquiry intentions) is clearly identified as a result of the analysis of the user's inquiry, and the number of responses to the user's inquiry is less than a preset number. (For example, if it is determined that there are less than 5 responses to the inquiry), the possibility of response may be determined to be above the threshold standard.
  • the determination of the possibility of responding as described above may be made when, as a result of analysis of the user's inquiry, the intention of the user's inquiry is identified as two or more, and is not clearly identified, or when the number of replies to the user's inquiry is more than a preset number. If it is determined that there are excessive numbers (for example, 5 or more), the possibility of response may be determined to be below the threshold standard.
  • the response provider 13 uses voice synthesis technology (Text to speech, Using text-to-speech (TTS), it will be possible to transmit a text-based response to the user as a human voice.
  • voice synthesis technology Text to speech, Using text-to-speech (TTS)
  • TTS text-to-speech
  • the response possibility determination unit 12 described above analyzes the possibility of responding to the user's inquiry as being less than the threshold standard, a request for material is made to request a re-inquiry for the previously received inquiry from the user. It may further include part 14.
  • the above-described material request unit 14 performs a function of requesting a more detailed and specific inquiry from the user who made the inquiry, as the analysis of the previously received inquiry is not clearly performed. It may be possible, and various embodiments may exist in this material method.
  • the material request unit 14 of the present invention presents the two or more analyzed query intentions to the user and provides two or more query intentions. By requesting a selection input for any one of the above query intentions, the user can re-inquire about the user's previously received query.
  • the content of the user's first query (101a) is 'This is my first time going to Jeju Island in the winter, so I'm concerned about where to go.'
  • a query containing more than two sentences such as 'When will Asiana online check-in be possible?', was performed, and that the two sentences were not highly related.
  • the first sentence's intention to request a response to 'Winter Jeju Island travel' can be understood through a natural language preprocessing model, but the second sentence requests a response to 'Asiana online check-in time' through a natural language preprocessing model.
  • the intention of the query can be identified, and responses to these queries are performed in the geographic information database where knowledge data about geographic information is stored and the aircraft information database where knowledge data about aircraft information is stored, respectively. Since the response comes from a different knowledge database, it may be judged that there is no significant correlation between the query intentions.
  • the material request unit 14 of the present invention can present a plurality of query intentions, such as 102a, to the user, and perform a material request to select one of the plurality of query intentions. According to the performance of the function of the material request unit 14, when a query is made for one of the plurality of query intentions identified by the user, as in 101b, a response to the query can be provided, as in 102b.
  • the material request unit 14 As a result of performing the function of the response possibility determination unit 12 described above, it is determined that the possibility of responding to the user's inquiry is less than the threshold standard, and the basis for determining that the possibility of response is less than the threshold standard is the user's If there are too many responses to the inquiry exceeding the preset number, the material request unit 14 generates a list of queries corresponding to the responses exceeding the preset number as a recommended query list and then sends it to the user. By providing it to the user, it is possible to enable the user to re-inquire about the user's previously received inquiry.
  • the request for the material of the present invention can be made by requesting a material inquiry to ask more specifically what you want to know about the universe, as in 202a, and at the same time, a list of questions for subcategories of answers related to the universe is created as a list of recommended questions for the user. can be provided to.
  • a list of queries on space-related knowledge by category stored in the space-related knowledge database such as the etymology of the universe, physical characteristics of the universe, composition of the universe, changes in the universe view, and various views of the universe, is provided, It can be provided as a recommended query list, and through this, it can function to clearly determine the intention of the user's query by allowing the user to perform a specific query again.
  • a query to know the etymology of the universe is re-queried as in 201b in response to the request for the material of 202a, and accordingly, a response to the etymology of the universe is provided in the space-related knowledge database. It can be seen by referring to the embodiment provided.
  • the present invention when the user's query is performed as a long sentence or multiple sentences and the intention of the query is identified as multiple, and when the range of responses to the user's query is too wide (i.e., when the user's query is an abstract query) In this case, the response to the query is not given up, but the query request unit 14 is encouraged to perform a query that allows the user to determine the actual intention of the query for which the user wants a response. , the effect of implementing a question-and-answer system that allows users to quickly and accurately provide the desired response can be expected.
  • each query phrase is provided to the user. It can be provided after assigning a different identification code to each query phrase, and when a material inquiry based on the recommended query list is performed from the user according to the function of the material request unit 14, the query receiving unit 11 provides the query phrase and It may function to receive and process a user's query by means of including at least one of the identification codes assigned to the query phrase.
  • the material query performed by the user is divided into materials included in the material list. It may be received as a phrase, but in abbreviated form, when performing a query with an identification code given to the phrase of the material, that is, a query phrase about the origin of the universe, the user can receive the same as 202b just by entering the Arabic numeral 1. A response can be obtained, and by performing this function, the inconvenience on the part of the user performing the query is greatly reduced, thereby improving the convenience of the query answering system.
  • the above-described response provider 13 determines the intention of the query from the user's query analyzed and refined through a natural language preprocessing model, and stores the response in the database (20). ) and functions to provide the first response, which is a response derived by artificial intelligence.
  • the present invention can connect to an external database that provides weather services and provide weather information according to region and time to the user as a response. If the intention of the query identified from the user's query is a query related to geography, the present invention can provide a first response by accessing an external database that provides geographic services through a plug-in or the like.
  • the response provider 13 of the present invention is connected to a plurality of external databases in which various expert knowledge services are accumulated in addition to the above-mentioned services according to the analyzed query intent, and provides a response corresponding to the query intent to the user. It may function to provide, but the present invention is not limited thereto.
  • responses to queries related to mathematical operations, unit conversion, etc. can be performed on their own without referring to a separate knowledge service.
  • the present invention in addition to providing the above-described first response as a response to a user's inquiry, the present invention manages the inquiry and response history performed between multiple users on a preset website.
  • the database may function to additionally provide a second response by analyzing the question and response details accumulated in the database using a text mining technique.
  • these second responses can be generally understood as sharing knowledge obtained from the user's experience, which has the effect of making it possible to easily acquire professional knowledge through the first response and experiential knowledge through the second response. There is.
  • the function of the material request unit 14 is performed repeatedly a threshold number of times, but the material request unit 14 is performed more than the threshold number of times. Even after performing the function of (14) repeatedly, if the analysis of the user's query fails, the query receiver ( 11) delivers the content of the user's query recognized based on voice recognition technology, and correction of the content of the user's query transmitted to the user interface using the input means of the user terminal (e.g. keyboard, touchpad, etc.) It is desirable to make it happen.
  • the input means of the user terminal e.g. keyboard, touchpad, etc.
  • 300 in FIG. 5 shows an example in which the query receiver 11 delivers the content of the user's query recognized based on voice recognition technology as analysis of the user's query fails.
  • an example of requesting correction for the part where voice recognition failed, 'Hoo? Jip' is shown
  • 310 in FIG. 5 is an example in which correction processing for the part where voice recognition failed is performed through the user terminal, and the input of the user terminal is
  • An example is shown in which the part where voice recognition failed, 'Hoo? House', is corrected and submitted to 'Hoo Pot House'.
  • 5 shows an example of searching for a response to the query 'Tell me about a delicious hotpot restaurant near here' with reference to the details corrected in the above-described embodiment of 310, through which the user can provide a response. This has the effect of allowing you to more clearly understand the content of the query you want to obtain.
  • the processor of the present invention stores the query data that failed voice recognition and the correction data corrected from the user terminal, so that when voice data similar to the query data that failed voice recognition is received later, the previously stored correction data By functioning to present to the user, it may function to reduce the inconvenience of entering a query phrase or phrase that fails voice recognition in the user terminal, but the present invention is not limited to this.
  • the above-described artificial intelligence-based interactive question and answer system 10 provides user inquiries and responses to user inquiries. It may further include a question-and-answer management unit.
  • the above-mentioned question response management unit matches and manages the user's inquiry initially received by the inquiry receiver 11 and the query phrase derived by analyzing and refining the user's inquiry through a natural language preprocessing model in the response possibility analysis unit, By analyzing the similarity between a plurality of query phrases, it can function to manage the user's query data set used to derive query phrases showing similarity greater than a critical similarity as query data related to the query phrase.
  • the similarity between the above-mentioned query phrases may be determined using an algorithm that includes at least one of the Jacquard similarity determination algorithm and the cosine similarity determination algorithm, which are algorithms that determine how similar the text included in the user's query is. there is.
  • the Jacquard similarity judgment algorithm described above uses a method of making two sentences each into a set of words and then measuring the similarity through the two sets.
  • the number of common words which is the intersection of the two sets, is calculated as the union of the two sets, that is, the number of all words. Similarity is determined by dividing by number, and the result has a value between 0 and 1 depending on the number of common elements, and the closer it is to 1, the higher the similarity is.
  • the first sentence is [today, Seoul, weather, what about]
  • the tokens (concepts of each word) of the sentence are separated into five
  • Sentence 2 is [today, Seoul, weather]
  • the tokens of the sentence are separated into 3, of which there are 3 common intersections of the tokens, and the total tokens of the two sentences are 8, giving a similarity of 38%.
  • the first sentence is divided into four sentence tokens as [today, Seoul, weather, how are you], and the second sentence is [today, Seoul, weather].
  • the token is divided into three, so the intersection of the two sentences is three, and the total tokens based on the actual morphemes of the two sentences are seven, resulting in a similarity of 43%.
  • the above-described cosine similarity judgment algorithm is an algorithm that obtains vectors for each word in a sentence and determines similarity using the cosine angle between the vectors.
  • the cosine value has a value between -1 and 1, so the more the two vectors have the same direction, the more similar they can be considered. If the angle is vertical, 0 is returned, and if they are completely different, -1 is returned to judge the similarity. I do it.
  • the first sentence is 'What musicals are playing at Seoul Arts Center in November?'
  • the second sentence is 'What musicals are playing at Suwon Gyeonggi Arts Center in November?'
  • the first sentence is separated into real morpheme-based tokens as [November, Seoul Arts Center, Musical]
  • the second sentence is [November, [Suwon, Gyeonggi Arts Center, Musical] tokens based on actual morphemes are separated and can be expressed as Table 1 below.
  • At least one of the two algorithms described above can be used to determine similarity, and for example, user queries with a similarity of 40% or more can be grouped and managed as a query data set.
  • a response corresponding to the corresponding query phrase is extracted from the related knowledge database. This allows an immediate response to be sent.
  • the response corresponding to the user's query is managed in the query data set, so that the user's query is analyzed and refined by applying a natural language preprocessing model each time. Since the time required can be shortened, it is possible to build a question-and-answer system that can provide quick responses to user inquiries.
  • the present invention accurately understands the user's inquiry intention and quickly provides a response to the inquiry intention.
  • the intent of more than one query is identified in the user's query, or when the user's query is abstract and there is a possibility that excessively many responses will be provided, we request a re-inquiry with a more specific query phrase from the user. This has the effect of increasing users' satisfaction with the question-and-answer system by allowing them to obtain the answers they want.
  • the present invention provides a response to a user's inquiry as a first response containing expert knowledge extracted from an expert knowledge database and a second response containing experiential knowledge based on the users' experience. , It has the effect of being able to implement a differentiated question and answer service.
  • FIG. 7 illustrates an example of the internal configuration of a computing device according to an embodiment of the present invention, and in the following description, overlapping with the description of FIGS. 1 to 6 described above Description of unnecessary embodiments will be omitted.
  • the computing device 10000 includes at least one processor 11100, a memory 11200, a peripheral interface 11300, and an input/output subsystem ( It may include at least an I/O subsystem (11400), a power circuit (11500), and a communication circuit (11600). At this time, the computing device 10000 may correspond to a user terminal (A) connected to a tactile interface device or the computing device (B) described above.
  • the memory 11200 may include, for example, high-speed random access memory, magnetic disk, SRAM, DRAM, ROM, flash memory, or non-volatile memory. there is.
  • the memory 11200 may include software modules, instruction sets, or various other data required for the operation of the computing device 10000.
  • access to the memory 11200 from other components may be controlled by the processor 11100.
  • the peripheral interface 11300 may couple input and/or output peripherals of the computing device 10000 to the processor 11100 and the memory 11200.
  • the processor 11100 may execute a software module or set of instructions stored in the memory 11200 to perform various functions for the computing device 10000 and process data.
  • the input/output subsystem 11400 can couple various input/output peripheral devices to the peripheral interface 11300.
  • the input/output subsystem 11400 may include a controller for coupling peripheral devices such as a monitor, keyboard, mouse, printer, or, if necessary, a touch screen or sensor to the peripheral device interface 11300.
  • peripheral devices such as a monitor, keyboard, mouse, printer, or, if necessary, a touch screen or sensor to the peripheral device interface 11300.
  • input/output peripheral devices may be coupled to the peripheral interface 11300 without going through the input/output subsystem 11400.
  • Power circuit 11500 may supply power to all or some of the terminal's components.
  • power circuit 11500 may include a power management system, one or more power sources such as batteries or alternating current (AC), a charging system, a power failure detection circuit, a power converter or inverter, a power status indicator, or a power source. It may contain arbitrary other components for creation, management, and distribution.
  • the communication circuit 11600 may enable communication with another computing device using at least one external port.
  • the communication circuit 11600 may include an RF circuit to transmit and receive RF signals, also known as electromagnetic signals, to enable communication with other computing devices.
  • FIG. 7 is only an example of the computing device 10000, and the computing device 11000 may omit some components shown in FIG. 7, further include additional components not shown in FIG. 7, or 2. It may have a configuration or arrangement that combines more than one component.
  • a computing device for a communication terminal in a mobile environment may further include a touch screen or a sensor in addition to the components shown in FIG. 7, and may include various communication methods (WiFi, 3G, LTE) in the communication circuit 1160. , Bluetooth, NFC, Zigbee, etc.) may also include a circuit for RF communication.
  • Components that may be included in the computing device 10000 may be implemented as hardware, software, or a combination of both hardware and software, including one or more signal processing or application-specific integrated circuits.
  • Methods according to embodiments of the present invention may be implemented in the form of program instructions that can be executed through various computing devices and recorded on a computer-readable medium.
  • the program according to this embodiment may be composed of a PC-based program or a mobile terminal-specific application.
  • the application to which the present invention is applied can be installed on a user terminal through a file provided by a file distribution system.
  • the file distribution system may include a file transmission unit (not shown) that transmits the file according to a request from the user terminal.
  • devices and components described in embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), It may be implemented using one or more general-purpose or special-purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions.
  • a processing device may execute an operating system (OS) and one or more software applications that run on the operating system.
  • OS operating system
  • software applications that run on the operating system.
  • a processing device may access, store, manipulate, process, and generate data in response to the execution of software.
  • a single processing device may be described as being used; however, those skilled in the art will understand that a processing device may include multiple processing elements and/or multiple types of processing elements. It can be seen that it may include.
  • a processing device may include a plurality of processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are possible.
  • Software may include a computer program, code, instructions, or a combination of one or more of these, which may configure a processing unit to operate as desired, or may be processed independently or collectively. You can command the device.
  • Software and/or data may be used by any type of machine, component, physical device, virtual equipment, computer storage medium or device to be interpreted by or to provide instructions or data to a processing device. It can be embodied permanently or temporarily.
  • Software may be distributed over networked computing devices and stored or executed in a distributed manner.
  • Software and data may be stored on one or more computer-readable recording media.
  • the method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium.
  • the computer-readable medium may include program instructions, data files, data structures, etc., singly or in combination.
  • Program instructions recorded on the medium may be specially designed and configured for the embodiment or may be known and available to those skilled in the art of computer software.
  • Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks.
  • program instructions include machine language code, such as that produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc.
  • the hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Landscapes

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

Abstract

La présente invention concerne un système de question et de réponse interactif basé sur l'intelligence artificielle et, plus particulièrement, comprend : une unité de réception de question pour recevoir une question d'un utilisateur sous la forme d'un langage naturel sur la base de paroles converties en texte (STT) ; une unité de détermination de possibilité de réponse pour déterminer, de façon à fournir une réponse à la question de l'utilisateur reçue par l'unité de réception de question, une possibilité de réponse à la question de l'utilisateur par analyse et affinage, par l'intermédiaire d'un modèle de prétraitement de langage naturel, de la question de l'utilisateur reçue sous la forme d'un langage naturel ; une unité de fourniture de réponse pour fournir une réponse à la question de l'utilisateur à l'aide d'un algorithme d'intelligence artificielle prédéfini lorsque l'unité de détermination de possibilité de réponse analyse que la possibilité de réponse à la question de l'utilisateur est égale ou supérieure à une référence de seuil ; et une unité de demande de question récursive pour demander à l'utilisateur une question récursive pour une question reçue au préalable de l'utilisateur lorsque l'unité de détermination de possibilité de réponse analyse que la possibilité de réponse à la question de l'utilisateur est inférieure à la référence de seuil.
PCT/KR2023/002677 2022-11-14 2023-02-24 Système de questions et de réponses interactif basé sur l'intelligence artificielle WO2024106626A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR10-2022-0151191 2022-11-14
KR1020220151191A KR20240069959A (ko) 2022-11-14 2022-11-14 인공지능 기반의 대화형 질의 응답 시스템

Publications (1)

Publication Number Publication Date
WO2024106626A1 true WO2024106626A1 (fr) 2024-05-23

Family

ID=91084679

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2023/002677 WO2024106626A1 (fr) 2022-11-14 2023-02-24 Système de questions et de réponses interactif basé sur l'intelligence artificielle

Country Status (2)

Country Link
KR (1) KR20240069959A (fr)
WO (1) WO2024106626A1 (fr)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160225370A1 (en) * 2015-01-30 2016-08-04 Microsoft Technology Licensing, Llc Updating language understanding classifier models for a digital personal assistant based on crowd-sourcing
KR20180071931A (ko) * 2016-12-20 2018-06-28 삼성전자주식회사 전자 장치, 그의 사용자 발화 의도 판단 방법 및 비일시적 컴퓨터 판독가능 기록매체
KR20210125399A (ko) * 2020-04-08 2021-10-18 (주)아이컴시스 대화형 음성봇 서버 및 이를 이용한 무인 상담 방법
KR102319648B1 (ko) * 2021-02-19 2021-11-01 (주)아와소프트 빅데이터를 기반으로 생성된 학습데이터를 제공하는 챗봇 시스템 및 그 방법
KR102383810B1 (ko) * 2021-09-28 2022-04-11 주식회사 퍼니웍 챗봇을 이용한 다국어 지원 시스템 및 방법

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160225370A1 (en) * 2015-01-30 2016-08-04 Microsoft Technology Licensing, Llc Updating language understanding classifier models for a digital personal assistant based on crowd-sourcing
KR20180071931A (ko) * 2016-12-20 2018-06-28 삼성전자주식회사 전자 장치, 그의 사용자 발화 의도 판단 방법 및 비일시적 컴퓨터 판독가능 기록매체
KR20210125399A (ko) * 2020-04-08 2021-10-18 (주)아이컴시스 대화형 음성봇 서버 및 이를 이용한 무인 상담 방법
KR102319648B1 (ko) * 2021-02-19 2021-11-01 (주)아와소프트 빅데이터를 기반으로 생성된 학습데이터를 제공하는 챗봇 시스템 및 그 방법
KR102383810B1 (ko) * 2021-09-28 2022-04-11 주식회사 퍼니웍 챗봇을 이용한 다국어 지원 시스템 및 방법

Also Published As

Publication number Publication date
KR20240069959A (ko) 2024-05-21

Similar Documents

Publication Publication Date Title
WO2021049706A1 (fr) Système et procédé de réponse aux questions d'ensemble
JP3114181B2 (ja) 異言語交信用翻訳方法およびシステム
WO2023029506A1 (fr) Procédé et appareil d'analyse d'état de maladie, dispositif électronique et support de stockage
US11157536B2 (en) Text simplification for a question and answer system
US20100121630A1 (en) Language processing systems and methods
Shah et al. NLKBIDB-Natural language and keyword based interface to database
JP2021507350A (ja) 複雑な回答の補強証拠取り出し
KR20040025642A (ko) 확인 문장을 검색하기 위한 방법 및 시스템
KR20160007040A (ko) 단문／복문 구조의 자연어 질의에 대한 검색 및 정보 제공 방법 및 시스템
Alqahtani et al. Arabic Quranic search tool based on ontology
WO2024090712A1 (fr) Système de conversation par intelligence artificielle pour psychothérapie par empathie
WO2020242086A1 (fr) Serveur, procédé et programme informatique pour supposer l'avantage comparatif de multi-connaissances
WO2016190495A1 (fr) Procédé de gestion de règles à base de données non structurées et dispositif associé
KR20170107282A (ko) 자연어 이해 및 질의응답 기반의 의사결정 지원 장치 및 방법
KR20200037077A (ko) Vqa 트레이닝 데이터를 생성하는 방법, 장치, 기기 및 컴퓨터 판독 가능 매체
Liebeskind et al. Semiautomatic construction of cross-period thesaurus
WO2018147543A1 (fr) Système de questions-réponses basé sur un graphe de concept et procédé de recherche de contexte l'utilisant
CN113157887A (zh) 知识问答意图识别方法、装置、及计算机设备
WO2014148664A1 (fr) Système de recherche en plusieurs langues, procédé de recherche en plusieurs langues et système de recherche d'image basé sur la signification d'un mot
WO2024106626A1 (fr) Système de questions et de réponses interactif basé sur l'intelligence artificielle
CN117131176A (zh) 交互式问答处理方法、装置、电子设备和存储介质
Tarbell et al. Towards understanding the generalization of medical text-to-sql models and datasets
WO2020138608A1 (fr) Procédé et appareil de réponse à une question faisant appel à une pluralité de robots conversationnels
Hosseini Pozveh et al. FNLP‐ONT: A feasible ontology for improving NLP tasks in Persian
WO2019112223A1 (fr) Procédé de récupération de document électronique et serveur associé

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23891703

Country of ref document: EP

Kind code of ref document: A1