Large language model-based apparatus for preparing electronic document through natural language conversation and method for preparing electronic document by using same

The electronic document writing device uses a large language model to generate keywords and write content data directly into documents, addressing the inefficiencies of existing systems by eliminating the need for additional learning processes, thus enabling efficient and cost-effective document creation across multiple fields.

WO2024248451A1PCT designated stage expired Publication Date: 2024-12-05FORCS

Patent Information

Application Number
PCT/KR2024/007158
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-05-30
Filing Date
2024-05-27
Publication Date
2024-12-05

AI Technical Summary

Technical Problem

Current electronic document creation systems using deep learning-based language models require additional learning processes and data collection for each field, leading to increased costs and inefficiencies as the number of fields increases, necessitating a solution that eliminates the need for these processes.

Method used

An electronic document writing device utilizing a large language model (LLM) that receives user input, generates keywords, and writes content data into electronic documents without requiring additional learning processes, allowing for efficient creation of documents across various fields without the need for model fine-tuning.

Benefits of technology

Enables the creation of electronic documents in multiple formats across various fields without the need for separate learning processes, reducing costs and time, and providing a user-friendly interface for document creation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure KR2024007158_05122024_PF_FP_ABST
    Figure KR2024007158_05122024_PF_FP_ABST
Patent Text Reader

Abstract

The present application relates to an apparatus for preparing an electronic document, comprising: an input unit for receiving, from a user terminal, a response input including content data to be input to a writing item of the electronic document; a processing unit that generates a keyword corresponding to the content data from the response input by using a large language model; and a generation unit that extracts the content data corresponding to the writing item on the basis of the keyword and writes the content data in the electronic document.
Need to check novelty before this filing date? Find Prior Art

Description

A large-scale language model-based electronic document creation device using natural language conversation and an electronic document creation method using the same

[0001] This invention relates to a device for creating electronic documents based on a large-scale language model through natural language conversation, and a method for creating electronic documents using the same. For example, this invention relates to a device for automatically creating electronic documents by generating keywords necessary for document creation from user responses based on a large-scale language model, and a method for creating electronic documents using the same.

[0002] Recently, situations requiring the creation of documents remotely, even without face-to-face interaction, have become increasingly common. Furthermore, as the types of documents users desire to create, depending on the time and situation, are diversifying, demand for companies that provide electronic documents across various fields (domains) is increasing.

[0003] To meet user demands, electronic document creation systems currently in use utilize deep learning-based language models to automatically generate electronic documents for specific fields.

[0004] However, to provide smooth service, fine-tuning of deep learning-based document creation models for each field is inevitable, and as the number of fields provided increases, costs such as training time and the collection of training data for fine-tuning become inevitable.

[0005] Accordingly, there is an increasing need for technology for electronic document creation devices that provide the system with the minimum information necessary for document creation, thereby configuring an environment for creating documents by field, and thus not requiring additional learning processes for the model.

[0006] The technology underlying this application is disclosed in Korean Patent Publication No. 10-1705228.

[0007] The present invention aims to solve the problems of the prior art described above, and to provide an electronic document creation device and an electronic document creation method using the same, which do not require an additional learning process for the model by utilizing a large language model (LLM) rather than a general language model.

[0008] However, the technical tasks to be achieved by the embodiments of the present invention are not limited to the technical tasks described above, and other technical tasks may exist.

[0009] As a technical means for achieving the above-described technical task, an electronic document creation device according to one embodiment of the present invention may include an input unit that receives a response input including content data to be entered into an entry item of the electronic document from a user terminal; a processing unit that generates a keyword corresponding to the content data from the response input using a large language model; and a generation unit that extracts the content data corresponding to the entry item based on the keyword and enters the content data into the electronic document.

[0010] According to one embodiment of the present invention, the input unit may convert the response input into an input form recognizable by the large-scale language model, but is not limited thereto.

[0011] According to one embodiment of the present invention, the response input may be characterized by including, but is not limited to, at least one of a voice input, a text input, an image input, and a video input.

[0012] According to one embodiment of the present invention, the processing unit may generate the keyword from the response input using query information generated based on a prompt including output format information for generating the keyword in a form processable by the generating unit and meta information including input item information for entry items of the electronic document, but is not limited thereto.

[0013] According to one embodiment of the present invention, the processing unit may be characterized by obtaining the prompt and the query information together with the response input as inputs of the large-scale language model, but is not limited thereto.

[0014] According to one embodiment of the present invention, the prompt and the query information may be updatable by a verification process that uses test data including example conversations in which questions and correct answers corresponding to the questions are predetermined, and determines at least one of whether the keyword derived by inputting the prompt and the query information into the large-scale language model matches the predetermined correct answer for the example conversation and whether the keyword matches the output format information, but is not limited thereto.

[0015] According to one embodiment of the present invention, the electronic document having the content data entered therein may include an output unit that outputs the electronic document through the user terminal, but is not limited thereto.

[0016] Meanwhile, a method for creating an electronic document using an electronic document creation device according to one embodiment of the present invention may include the steps of: receiving a response input including content data to be entered into an entry item of the electronic document from a user terminal; generating a keyword corresponding to the content data from the response input using a large-scale language model; and extracting the content data corresponding to the entry item based on the keyword and entering the content data into the electronic document.

[0017] According to one embodiment of the present invention, the receiving step may be, but is not limited to, converting the response input into an input form recognizable by a large-scale language model.

[0018] According to one embodiment of the present invention, the response input may be characterized by including, but is not limited to, at least one of a voice input, a text input, an image input, and a video input.

[0019] According to one embodiment of the present invention, the step of generating the keyword may be, but is not limited to, generating the keyword from the response input using query information generated based on a prompt including output format information for generating the keyword in a form processable by the electronic document creation device and meta information including input item information for entry items of the electronic document.

[0020] According to one embodiment of the present invention, the electronic document creation device may be characterized by receiving the prompt and the query information in advance from a service provider terminal, but is not limited thereto.

[0021] According to one embodiment of the present invention, the method may further include, but is not limited to, a step of outputting the electronic document in which the content data is entered through the user terminal.

[0022] According to one embodiment of the present invention, when the entry of the content data for all entry items of the electronic document is completed and the creation of the electronic document is completed, the step of storing the electronic document may be further included, but is not limited thereto.

[0023] The above-described problem-solving methods are merely exemplary and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, additional embodiments may be included in the drawings and detailed description of the invention.

[0024] According to the aforementioned means for solving the problem of the present invention, by utilizing a large language model (LLM) rather than a general language model, an electronic document creation device and an electronic document creation method using the same can be provided that do not require an additional learning process for the model.

[0025] According to the aforementioned means of solving the problem of this institute, by utilizing a large language model (LLM) in the process of providing a system that finds keywords from a user's response and reflects / writes them in a document, an electronic document writing environment in various formats by field can be provided to the user in a short period of time without a separate learning process.

[0026] However, the effects that can be obtained from this center are not limited to the effects described above, and other effects may exist.

[0027] Figure 1 is a schematic diagram of an electronic document creation system according to one embodiment of the present invention.

[0028] Figure 2 is a conceptual diagram for explaining the operation process of an electronic document creation system according to one embodiment of the present invention.

[0029] Figure 3 is a diagram illustrating an example of an output result according to the input of a large-scale language model.

[0030] Figure 4 is a schematic diagram of an electronic document creation device according to one embodiment of the present invention.

[0031] Figure 5 is a flowchart of a process for verifying a prompt of an electronic document creation device according to one embodiment of the present invention.

[0032] Figure 6 is a flowchart of a method for creating an electronic document using an electronic document creation device according to one embodiment of the present invention.

[0033] Figure 7 is a flowchart of an electronic document creation method according to one embodiment of the present invention. Specifically, this flowchart illustrates a process for creating an electronic document in which a user responds to a query from an electronic document creation device.

[0034] Figure 8 is a flowchart of an electronic document creation method according to one embodiment of the present invention. Specifically, this flowchart illustrates a process in which an electronic document is created by a user responding to a counselor's inquiries.

[0035] Below, with reference to the attached drawings, embodiments of the present invention are described in detail to facilitate easy implementation by those skilled in the art. However, the present invention can be implemented in various different forms and is not limited to the embodiments described herein. In the drawings, irrelevant parts have been omitted for clarity, and similar reference numerals have been used throughout the specification to indicate similar elements.

[0036] Throughout this specification, when a part is said to be "connected" to another part, this includes not only the case where it is "directly connected," but also the case where it is "electrically connected" or "indirectly connected" with another element in between.

[0037] Throughout this specification, when it is said that a member is located “on,” “above,” “upper,” “lower,” “lower” or “lower” another member, this includes not only cases where the member is in contact with the other member, but also cases where another member exists between the two members.

[0038] Throughout this specification, whenever a part is said to "include" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise specifically stated.

[0039] Hereinafter, the electronic document creation device of this invention and the electronic document creation method using the same will be described in detail with reference to implementation examples, examples, and drawings. However, the invention is not limited to these implementation examples, examples, and drawings.

[0040] This invention relates to a device for creating electronic documents based on a large-scale language model through natural language conversation, and a method for creating electronic documents using the same. For example, this invention relates to a device for automatically creating electronic documents by generating keywords necessary for document creation from user responses based on a large-scale language model, and a method for creating electronic documents using the same.

[0041] Figure 1 is a schematic diagram of an electronic document creation system according to one embodiment of the present invention.

[0042] Referring to FIG. 1, an electronic document creation system (10) according to one embodiment of the present invention may include an electronic document creation device (100), a user terminal (200), and a service provider terminal (300) according to one embodiment of the present invention.

[0043] An electronic document creation device (100), a user terminal (200), and a service provider terminal (300) can communicate with each other through a network (20). The network (20) refers to a connection structure that enables information exchange between each node, such as terminals and servers. Examples of such a network (20) include, but are not limited to, a 3GPP (3rd Generation Partnership Project) network, an LTE (Long Term Evolution) network, a 5G network, a WIMAX (World Interoperability for Microwave Access) network, the Internet, a LAN (Local Area Network), a Wireless LAN (Wireless Local Area Network), a WAN (Wide Area Network), a PAN (Personal Area Network), a wifi network, a Bluetooth network, a satellite broadcasting network, an analog broadcasting network, a DMB (Digital Multimedia Broadcasting) network, etc.

[0044] The user terminal (200) may be any type of wireless communication device, such as, for example, a smartphone, a smart pad, a tablet PC, a PCS (Personal Communication System), a GSM (Global System for Mobile communication), a PDC (Personal Digital Cellular), a PHS (Personal Handyphone System), a PDA (Personal Digital Assistant), an IMT (International Mobile Telecommunication)-2000, a CDMA (Code Division Multiple Access)-2000, a W-CDMA (W-Code Division Multiple Access), or a Wibro (Wireless Broadband Internet) terminal.

[0045] In particular, the user terminal (200) may include various audio devices and driving devices using newly emerging communication technologies, such as smart speakers, smart cars, smart appliances, wearable devices, and augmented reality (VR / MR) devices.

[0046] In addition, referring to FIG. 1, in the description of the embodiment of the present invention, the user terminal (200) may include a plurality of terminals corresponding to each speaker to provide a response input including content data generated (converted) into an electronic document form by the electronic document creation device (100) disclosed in the present invention.

[0047] For example, referring to FIG. 1, an electronic document creation device (100) can obtain a conversation input including a speech of a first user (1) and a speech of a second user (2) from a user terminal (200), and create (generate) an electronic document based on the obtained conversation input.

[0048] As another example, although not shown in the drawing, according to an implementation example of the present invention, the electronic document creation device (100) may create an electronic document by extracting and entering content data based on a conversation input including a speech of a first user (1) received from a first user terminal (210) and a speech of a second user (2) received from a second user terminal (220).

[0049] In addition, referring to FIG. 1, as described in detail below, the electronic document creation device (100) disclosed herein may be equipped with a large language model (LLM; 101) for obtaining query information and prompts necessary for creating an electronic document in advance and generating keywords through the meta information and prompts.

[0050] Figure 2 is a conceptual diagram for explaining the operation process of an electronic document creation system according to one embodiment of the present invention.

[0051] Referring to FIG. 2, the electronic document creation device (100) disclosed herein can operate to obtain, as input, a natural language conversation (11a) including conversation inputs performed by a first user (1) and a second user (2), and to provide, as output, an electronic document in which each input item of an electronic document format (form) to be created is entered based on a keyword derived from the natural language conversation.

[0052] Additionally, the large language model (LLM; 101) possessed by the electronic document creation device (100) can be used for the purpose of generating keywords to be used to enter input items of an electronic document from natural language conversation content.

[0053] More specifically, referring to (a) of FIG. 2, the electronic document creation device (100) can configure input data (12) of a large-scale language model (101). Specifically, the input data of the large-scale language model (101) can include a prompt (12a), a conversation content (12b), and a query (12c). The prompt (12a) is a part of a phrase transmitted to the large-scale language model (101), and can include a phrase that instructs the model (in other words, the large-scale language model (101), etc.) on what role it should perform. For example, the prompt (12a) can include content (phrase) that specifies an output format, content on restrictions that prevent a result other than a specified output format from being output, etc.

[0054] In addition, the conversation content (12b) can be composed in a manner of sequentially connecting the contents of speech between the first user (1) and the second user (2) or the speech between the user and the system, and the query (12c) can be composed based on meta information (11b) including the name (name) of each item (entry item) that constitutes the electronic document, the type of item value, the code for each type of item value, etc., as information that can be confirmed through the format of the electronic document to be created.

[0055] In addition, referring to (b) of FIG. 2, the large-scale language model (101) of the electronic document creation device (100) can receive input data (12) and output data (e.g., text data (13), etc.) that satisfies the output format defined in the prompt (12a). More specifically, the large-scale language model (101) can generate an answer suitable for a query (12c) from the input conversation content (12b) and derive an output according to the format defined in the prompt (12a, e.g., a string in JSON format, etc.).

[0056] In this regard, Fig. 3 is a diagram exemplifying output results according to the input of a large-scale language model. More specifically, Fig. 3 illustrates output data derived from a large-scale language model (101) when input data (12) is applied to the large-scale language model (101).

[0057] Referring to FIG. 3, it can be confirmed that an output having a different format from the output format described above through (a) of FIG. 2 is derived by a large-scale language model (101), and in a case where the format of the output data does not satisfy the predefined format, the electronic document creation device (100) can perform a modification / supplementation process to derive an output of the desired format by processing (enhancing) the phrase of the prompt (12a), the technical format of the query (12c), etc.

[0058] In addition, referring to (b) of FIG. 2, the electronic document creation device (100) can transmit the output derived by the large-scale language model (101) to the electronic document generator (102). For reference, the electronic document generator (102) may refer to the same configuration as at least one of the processing unit (120) and the generating unit (130) described below according to the implementation example of the present application, or may refer to any one of the lower modules provided in the processing unit (120) and the generating unit (130).

[0059] The electronic document generator (102) can generate keywords for each input item of an electronic document to be created by parsing the output of a large-scale language model (101) and reflect them in the electronic document.

[0060] Figure 4 is a schematic diagram of an electronic document creation device according to one embodiment of the present invention.

[0061] Referring to FIG. 4, an electronic document creation device (100) according to one embodiment of the present invention may include an input unit (110); a processing unit (120); and a generation unit (130).

[0062] The above input unit (110) can receive a response input including content data to be entered in the entry item of the electronic document from the user terminal (200).

[0063] In addition, the user terminal (200) can be understood as various devices equipped with a display screen capable of displaying information about an electronic document that has been completed. If the user terminal (200) is not equipped with a display screen, an additional terminal including a display capable of displaying information about an electronic document that has been completed can be equipped.

[0064] Meanwhile, the user terminal (200) may include an application unit (not shown) that is linked to the electronic document creation device (100) to transmit user input entered by the user to the electronic document creation device (100) or receive a processing result for the user input.

[0065] According to one embodiment of the present invention, the application unit may include a virtual artificial intelligence assistant application. The virtual artificial intelligence assistant application may include, but is not limited to, virtual artificial intelligence assistant services such as Siri, Google Assistant, Alexa, Cortana, Bixby, Nugu, Clova, etc.

[0066] According to another embodiment of the present invention, the application unit may include a messenger application. Examples of the messenger application may include, but are not limited to, messenger services such as KakaoTalk, MyPeople, Line, TikTok, BuddyBuddy, SayClub, MSN Messenger, Yahoo Messenger, NateOn, Daum Wiz Genie, Daum Messenger, KTiman Messenger, Facebook, Telegram, and WhatsApp.

[0067] According to one embodiment of the present invention, the response input may be characterized by including, but not limited to, at least one of a voice input, a text input, an image input, and a video input.

[0068] An electronic document creation device (100) according to the present invention can receive various types of input from a user terminal (200) and create an electronic document based on the input.

[0069] According to one embodiment of the present invention, the input unit (120) may convert the received response input into an input form recognizable by a large-scale language model (101), but is not limited thereto.

[0070] The electronic document creation device (100) according to the present invention is a device based on a large-scale language model, and the main input form of the large-scale language model is text. Accordingly, if the response input is not a text input, the input unit (110) can convert the response input into a form recognizable by the large-scale language model and transmit it.

[0071] For example, if the user's response input is voice input, it can be converted into text through Speech-To-Text (STT), and if it is image input, text information can be extracted through Optical Character Recognition (OCR), but is not limited thereto.

[0072] Meanwhile, the processing unit (120) can generate keywords corresponding to the content data from the response input using a large language model.

[0073] The above keyword includes the content data that must be written in the entry field of the electronic document, and the keyword generated by the processing unit (120) is transmitted to the generation unit (130) and processed into content data, and then can be written in the entry field of the electronic document.

[0074] According to one embodiment of the present invention, the processing unit (120) may be characterized by obtaining, as input of the large-scale language model, a prompt including query information configured based on meta information including input item information for entry items of the electronic document together with the response input and output format information for generating the keyword in a form processable by the generating unit (130), but is not limited thereto.

[0075] The above prompt is specifically part of a phrase that is transmitted to a large-scale language model (101), and may additionally include, in addition to the output format information, phrases that instruct the large-scale language model (101) what role it should perform, and contents that impose restrictions so that no results other than the output format are produced, and the like. By the above prompt, keyword generation and output format control can be easily performed in the processing unit (120) during the electronic document creation process.

[0076] The form of the above prompt may include natural language that is easily understandable to humans, as well as formulas, symbols, programming languages, encrypted content, etc., and is not particularly limited in form as long as it can be recognized by the processing unit (120).

[0077] The above output format is a form that can be parsed by the generation unit (130), and can be various formats such as JSON, xml, yaml, user-defined output, etc., and there is no particular limitation on the format as long as it can be processed by the generation unit (130).

[0078] The above meta information is information that can be confirmed through the format of the electronic document, and may include the name of an item that constitutes the electronic document, the type of the item value, the code for each type of the item value, etc., and query information may be configured based on input item information for the entry items of the electronic document among the meta information, and the electronic document creation device (100) may generate an appropriate query to induce a response including content data for creating an electronic document from the user by obtaining the query information in advance from the service provider.

[0079] For example, when writing a vacation application form as an electronic document, the processing unit (120) can obtain query information in advance based on meta information including input item information for entry items of the vacation application form including 'vacation start date', 'vacation end date', 'vacation reason', etc. by the service provider, and the electronic document writing device (100) can generate appropriate queries such as 'When is the vacation start date?', 'When is the vacation end date?', 'What is the vacation reason?', etc. to the user in order to induce a response including content data that must be entered in entry items such as 'vacation start date', 'vacation end date', 'vacation type', etc.

[0080] According to one embodiment of the present invention, the processing unit (120) may generate the keyword from the response input using the prompt and the query information, but is not limited thereto.

[0081] When generating a keyword from the response input, the processing unit (120) can clearly determine which entry item the response input is for through the query information, thereby preventing the problem of the generation unit (130) not being able to determine which entry item the content data extracted from the keyword should be reflected in.

[0082] The above processing unit (120) may also obtain the query information together with the response input through the input unit (110). In this case, the input unit (110) may transmit the response input and the query information to the processing unit (120) by configuring them in a format set through a pre-obtained prompt so that the processing unit (120) can clearly recognize that they are input data.

[0083] For example, if a prompt including information for generating a noun keyword from a string following “Q:” and extracting a value from a string following “A:” is obtained together with a response input from the processing unit (120), before transmitting the response input from the input unit (110) to the processing unit (120), the meta information and the response input can be transmitted with the prefixes “Q:” and “A:” attached to them, respectively, so that the processing unit (120) can clearly identify which entry item the response input is for.

[0084] Meanwhile, the above-mentioned generation unit (130) extracts the content data corresponding to the entry item based on the keyword and writes the content data into the electronic document.

[0085] According to one embodiment of the present invention, the electronic document creation device (100) may include an output unit (140) that outputs the electronic document in which the content data is entered through the user terminal (200), but is not limited thereto.

[0086] In this regard, the output unit (140) can check whether the content data has been properly written in the electronic document through the user terminal (200), and if there is an error in the entry of the content data, it can be corrected.

[0087] Meanwhile, the electronic document creation device (100) according to the present invention can operate so that the electronic document is created more accurately by having a function of verifying the output form of content data output through a prompt obtained in advance by the service provider.

[0088] According to one embodiment of the present invention, the prompt and the query information may be updatable by a verification process that uses test data including example conversations in which questions and correct answers corresponding to the questions are predetermined, and determines at least one of whether the keyword derived by inputting the prompt and the query information into the large-scale language model matches the predetermined correct answer for the example conversation and whether the keyword matches the output format information, but is not limited thereto.

[0089] More specifically, the electronic document creation device (100) can apply a verification process to confirm whether the processing unit (120) produces output as intended before actually providing the electronic document creation service disclosed herein, and when performing this verification process, a prompt and question composition method is applied using test data including example conversations in which questions and corresponding answers are specified in advance rather than actual conversations, to construct the input of a large-scale language model (101), and when the constructed input is transmitted, it can be verified whether the predefined answers and output formats are correctly reflected in the keywords derived from the processing unit (120).

[0090] In this regard, if an unintended form of output (keyword) is derived from the processing unit (120) considering the predefined correct answer and output format, the electronic document creation device (100) can repeatedly apply the verification process until the desired (intended) form of output is derived as a keyword by applying changes in the prompt phrase, changes in the order of the prompt phrase, advancements in the method / format of composing the query (query information), changes in the arrangement order between the prompt (12a), the text (conversation content, 12b), and the query (query information, 12c) when composing the input of the large-scale language model (101).

[0091] The verification target of the above verification process may include, but is not limited to, individual input components such as text, dates, combo boxes, and check boxes that constitute an electronic document.

[0092] Figure 5 is a flowchart of a process for verifying a prompt of an electronic document creation device according to one embodiment of the present invention.

[0093] Referring to FIG. 5, the service provider terminal (300) can apply a process of verifying and improving prompts and queries so that the processing unit (120) can derive a desired format of output (keyword) through the electronic document creation device (100) disclosed herein, and the processing unit (120) of the electronic document creation device (100) can obtain optimized (updated) prompts and queries to correctly provide the intended output by repeatedly applying the verification process illustrated in FIG. 5. More specifically, the service provider terminal (300) can confirm (verify) an output (e.g., an output text illustrated in FIG. 5, etc.) derived based on a test input input by the service provider, etc.

[0094] In addition, referring to FIG. 5, the overall flow of the verification process for updating prompt and query information can be confirmed when the content data derived in response to the test input does not conform to the preset output format (output text ≠ correct answer text). The above process can be repeatedly performed until the content data derived through the processing unit (120) conforms to the preset output format and correct answer, and the accuracy of the electronic document created through this verification process can be further improved.

[0095] Figure 6 is a flowchart of a method for creating an electronic document using an electronic document creation device according to one embodiment of the present invention.

[0096] Referring to FIG. 6, a method for creating an electronic document using an electronic document creation device according to one embodiment of the present invention may include a step (S100) of receiving a response input including content data to be entered into an entry item of the electronic document from a user terminal; a step (S200) of generating a keyword corresponding to the content data from the response input using a large-scale language model; and a step (S300) of extracting the content data corresponding to the entry item based on the keyword and entering the content data into the electronic document.

[0097] In the above description, steps S100 to S300 may be further divided into additional steps or combined into fewer steps, depending on the implementation example of the present invention. Furthermore, some steps may be omitted as needed, and the order of the steps may be changed.

[0098] An electronic document creation method according to one embodiment of the present invention may be implemented in the form of program commands that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program commands, data files, data structures, etc., either singly or in combination. The program commands recorded on the medium may be those specifically designed and configured for the present invention or may be those known and usable by 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; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program commands, such as ROMs, RAMs, and flash memories. Examples of program commands include not only machine language codes generated by a compiler, but also high-level language codes 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 present invention, and vice versa.

[0099] Additionally, the aforementioned electronic document creation method can also be implemented in the form of a computer program or application executed by a computer and stored in a recording medium.

[0100] In addition, the method for creating an electronic document according to the present invention may be performed in a form in which a user responds to an inquiry of an electronic document creation device (100), a form in which a user responds to an inquiry of a counselor, etc., but is not limited thereto.

[0101] Below, we will describe in more detail the methods for creating each electronic document. The following explanation assumes that the electronic document is a vacation request form. However, this is for convenience only and is not limited to a vacation request form.

[0102] Figure 7 is a flowchart of an electronic document creation method according to one embodiment of the present invention. Specifically, Figure 7 is a flowchart illustrating a process for creating an electronic document in which a user responds to a query from an electronic document creation device.

[0103] Referring to Fig. 7, first, a response input including content data to be entered in the entry item of the electronic document is received from a user terminal (S100).

[0104] In addition, referring to FIG. 7, a query for inducing a user's response including content data to be entered in an entry field of an electronic document for creating an electronic document can be transmitted to a user terminal, and the user can respond to the query on the user terminal and transmit the response input to the input unit (110).

[0105] At this time, there are cases where the user's response input does not include the input item information of the electronic document, so the generation unit (130) does not know which item the value should be reflected in.

[0106] Therefore, in order to prevent the problem of not being able to determine where to reflect the value in the generation unit (130), when transmitting data from the input unit (110) to the processing unit (120), meta information such as a system query is transmitted together with the response input so that the processing unit (120) can clearly detect which input item the answer is for. That is, as prompts and system queries are added, the data transmitted from the input unit (110) to the processing unit (120) may include prompts, query sentences, and user answers (text) converted into text form as input for a large-scale language model (101).

[0107] Meanwhile, before combining meta information with the above response input and transmitting the data to the processing unit (120), it must be configured in an input format that conforms to the prompt so that the processing unit (120) can clearly recognize that it is input data.

[0108] Referring to FIG. 7, when a prompt including information instructing the processing unit (120) to generate a noun keyword from the string following "Q:" and to extract a value from the string following "A:" is obtained, before transmitting data from the input unit (110) to the processing unit (120), the processing unit (120) can transmit the system query (meta information) and the user response (response input) with a prefix such as "Q:" and "A:", respectively, so that they can be clearly identified.

[0109] Next, a large-scale language model is used to generate keywords corresponding to the content data from the response input. (S200)

[0110] The processing unit (120) generates a keyword corresponding to content data through the response input and meta information received through the input unit (110), and converts the keyword specified through the prompt into an output format for generating it in a form that can be processed by the generation unit (130) and transmits it to the generation unit (130).

[0111] Referring to FIG. 7, if the prompt includes the content of setting the output format to JSON and displaying the date in the YYYY-MM-DD format, when properly processed by the processing unit (120), a keyword including the JSON string {'vacation start date': '2023-02-28'} can be transmitted to the generation unit (130).

[0112] Next, content data corresponding to the entry items is extracted based on the keywords, and the content data is entered into an electronic document. (S300)

[0113] The generation unit (130) can extract content data corresponding to the entry items of an electronic document based on the received keywords and enter the content data into the actual electronic document input items.

[0114] According to one embodiment of the present invention, the method may further include, but is not limited to, a step of outputting the electronic document in which the content data is entered through the user terminal.

[0115] Referring to FIG. 7, the output unit can output an electronic document with the "vacation start date" field updated through the user terminal so that the user can confirm the electronic document in which the content data derived in response to the response input is entered. In addition, for items other than the "vacation start date" that have not yet been answered, when the user responds to the system's additional query through the input unit (110), the output unit can confirm the same through the series of processes described above.

[0116] According to one embodiment of the present invention, when the entry of the content data for all entry items of the electronic document is completed and the creation of the electronic document is completed, the method may further include, but is not limited to, a step of storing the electronic document.

[0117] The electronic document creation device (100) can store a completed electronic document when the creation of the electronic document is completed by completing the entry of content data for all entry items included in the electronic document.

[0118] Figure 8 is a flowchart of an electronic document creation method according to one embodiment of the present invention. Specifically, Figure 8 is a flowchart illustrating the creation of an electronic document in which a user responds to a counselor's inquiries.

[0119] Referring to FIG. 8, a query for eliciting a user response, including content data to be entered in an entry field of an electronic document for creating an electronic document, may be transmitted from a counselor terminal to a user terminal. Simultaneously, the query may be transmitted to an electronic document creation device (100), and an unupdated electronic document based on the query may be transmitted to the user terminal. The user may respond to the query by inputting a response input to the input unit (110) on the user terminal.

[0120] For reference, although FIG. 8 illustrates that the first user terminal (210) corresponding to the first user (user) and the second user terminal (220) corresponding to the second user (counselor) are separate devices, this is to distinguish each speaker included in the conversation input, and can be understood to encompass a state in which the conversation input by the first user and the second user is obtained using a single device (e.g., the user terminal (200) illustrated in FIG. 1, etc.) according to the implementation example of the present invention.

[0121] In addition, referring to FIG. 8, when transmitting a response input from the input unit (110) to the processing unit (120), the generation unit (130) may transmit together meta information including input item information of the electronic document in a form identifiable by the processing unit (130) regarding which input item of the electronic document to fill in. That is, as prompts and system queries are added, the data transmitted from the input unit (110) to the processing unit (120) may include prompts, queries, and text-type data (fingerprints) as inputs to the large-scale language model (101).

[0122] For example, when a prompt including information instructing to find an answer to a question following “QN:” (N is a natural number) or to infer an answer is obtained from a fingerprint following “P:” in the processing unit (120), the system query (meta information) and the user-counselor conversation can be transmitted with a prefix such as “QN:”, “P:”, etc., so that the processing unit (120) can clearly identify them before transmitting the data from the input unit (110) to the processing unit (120).

[0123] Referring to FIG. 8, the counselor's inquiry and the user's response input can be converted into text in the input unit (110) and then connected to the text previously converted in the input unit (110) and transmitted to the processing unit (120).

[0124] The above description of the present invention is for illustrative purposes only, and those skilled in the art will readily appreciate that the present invention can be readily modified into other specific forms without altering the technical spirit or essential characteristics of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single entity may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined manner.

[0125] The scope of the present invention is indicated by the claims described below rather than the detailed description above, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present invention.

Claims

1. In the electronic document creation device, An input unit that receives a response input including content data to be entered in an entry field of the electronic document from a user terminal; A processing unit that generates keywords corresponding to the content data from the response input using a large language model; and A generation unit that extracts the content data corresponding to the entry item based on the keyword and writes the content data into the electronic document; An electronic document creation device including:

2. In paragraph 1, The above input section, Converting the above response input into an input form recognizable by the large-scale language model, Electronic document creation device.

3. In paragraph 2, An electronic document creation device, characterized in that the above response input includes at least one of voice input, text input, image input, and video input.

4. In paragraph 1, The above processing unit, The keyword is generated from the response input using query information generated based on a prompt including output format information for generating the keyword in a form that can be processed by the generation unit and meta information including input item information for the entry items of the electronic document. Electronic document creation device.

5. In paragraph 4, The above processing unit, An electronic document creation device characterized in that the prompt and the query information are acquired as inputs of the large-scale language model together with the response input.

6. In paragraph 5, The above prompt and query information are, An electronic document creation device characterized in that it is updateable by a verification process that uses test data including example dialogues in which a question and an answer corresponding to the question are predetermined, and determines at least one of whether the keyword derived by inputting the prompt and the question information into the large-scale language model corresponds to the predetermined answer for the example dialogue and whether the keyword corresponds to the output format information.

7. In paragraph 1, An output unit that outputs the electronic document containing the above content data through the user terminal; An electronic document creation device including:

8. In a method for creating an electronic document using an electronic document creation device, A step of receiving a response input including content data to be entered in an entry item of the electronic document from a user terminal; A step of generating keywords corresponding to the content data from the response input using a large-scale language model; and A step of extracting the content data corresponding to the entry item based on the keyword and entering the content data into the electronic document; A method for creating an electronic document, including:

9. In paragraph 8, The above receiving step is, Converting the above response input into an input form that can be recognized by a large-scale language model, How to create an electronic document.

10. In paragraph 9, A method for creating an electronic document, characterized in that the above response input includes at least one of voice input, text input, image input, and video input.

11. In paragraph 8, The steps for generating the above keywords are: Generating the keyword from the response input using query information generated based on a prompt including output format information for generating the keyword in a form processable by the electronic document creation device and meta information including input item information for the entry items of the electronic document. How to create an electronic document.

12. In paragraph 11, The above electronic document creation device, A method for creating an electronic document, characterized in that the prompt and the query information are received in advance from a service provider terminal.

13. In paragraph 12, The above prompt and query information are, An electronic document creation method, wherein a test input including test data is transmitted to the electronic document creation device, and the content data derived in response to the test input is updated through a verification process that determines whether the content data conforms to a preset output format.

14. In paragraph 8, A step of outputting the electronic document in which the above content data is entered through the user terminal; A method of creating an electronic document, which further includes:

15. In paragraph 14, When the above content data for all entry items of the above electronic document is completed and the creation of the electronic document is completed, the step of saving the above electronic document; A method of creating an electronic document, which further includes:

Citation Information

Patent Citations

  • Data generation method using language model, computer device, and computer program

    JP2023018624A

  • Method and system for summarizing document using hyperscale language model

    JP2023053867A

  • Friction material

    KR1020240158486A

  • Finishing composition for protecting surface of concrete structure and steel structure having excellent visibility and finishing method for protecting surface of concrete structure and steel structure using the same

    KR102352951B1

  • Apparatus and method for creating interactive and document-type electronic document, and method for providing electronic document that may be created interactively

    WO2021100989A1

Cited By

  • Search optimization method, device and medium based on large model

    CN119782494A

  • Electric power business work order information intelligent identification method based on OCR technology

    CN120182979A