WO2024218839A1 - コンテンツ生成システム - Google Patents

コンテンツ生成システム Download PDF

Info

Publication number
WO2024218839A1
WO2024218839A1 PCT/JP2023/015385 JP2023015385W WO2024218839A1 WO 2024218839 A1 WO2024218839 A1 WO 2024218839A1 JP 2023015385 W JP2023015385 W JP 2023015385W WO 2024218839 A1 WO2024218839 A1 WO 2024218839A1
Authority
WO
WIPO (PCT)
Prior art keywords
communicative
information
content
human language
generation system
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/JP2023/015385
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
陽至 日野
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Yamaha Motor Co Ltd
Original Assignee
Yamaha Motor Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Yamaha Motor Co Ltd filed Critical Yamaha Motor Co Ltd
Priority to PCT/JP2023/015385 priority Critical patent/WO2024218839A1/ja
Priority to JP2025514907A priority patent/JPWO2024218839A1/ja
Priority to CN202380096996.8A priority patent/CN120917439A/zh
Priority to ES202590064A priority patent/ES3041845A2/es
Priority to GB2518241.1A priority patent/GB2644497A/en
Priority to US18/637,896 priority patent/US20240346237A1/en
Priority to DE102024203563.7A priority patent/DE102024203563A1/de
Publication of WO2024218839A1 publication Critical patent/WO2024218839A1/ja
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9538—Presentation of query results
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00—Handling natural language data
    • G06F40/10—Text processing
    • G06F40/166—Editing, e.g. inserting or deleting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—Two-dimensional [2D] image generation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T13/00—Animation
    • G06T13/80—Two-dimensional [2D] animation, e.g. using sprites

Definitions

  • the present invention relates to a content generation system.
  • Non-Patent Document 1 A text generation system has been proposed that uses the large-scale language models (LLMs) GPT-2 and GPT-3 (Patent Document 1). Recently, GPT-4 was made public, and its usage is being widely studied (Non-Patent Document 1).
  • GPT-4 is finally released. Explaining how to use GPT-4 and its performance", [online], March 15, 2023, ChatGPT Research Institute, [Retrieved March 21, 2023], Internet ⁇ URL: https://chatgpt-lab.com/n/n7facbf0f8890>
  • the object of the present invention is to use LLM to efficiently generate content that is easier for humans to understand. More specifically, the object of the present invention is to both promote user understanding of the content and improve the efficiency of content generation, thereby reducing the load on the hardware resources that make up the content generation system.
  • the present invention can provide the following content generation system:
  • a content generation system comprising: The content generation system includes: at least one user interface; At least one memory; at least one processor coupled to the memory and configured to execute at least one program stored in the memory;
  • the at least one program comprises: (A) utilizing each of a large scale language model that operates using communicative human language information input via the user interface, and visualization software that is programmed to output visual information using textual information provided; (B) in (A), obtaining communicative human language information for inclusion in the content using at least a portion of the communicative human language information input via the user interface using the large-scale language model; (C) in (A), acquiring communicative human language information for selecting visualization software capable of generating visual information to be included in the content by the large-scale language model using at least a portion of the communicative human language information input via the user interface, and selecting the visualization software based on the acquired communicative human language information; (D) in (A), operating the visualization software selected in (C) with text information obtained based on the output of the large-scale language model in
  • This content generation system uses the LLM to efficiently generate content that is easier for humans to understand.
  • this visualization software the user does not need to select the visualization software, and the content generation system can select visualization software according to the user's needs. This can both promote the user's understanding of the content and improve the efficiency of content generation.
  • This content generation system does not assume the use of specific visualization software, and since the content generation system selects visualization software according to the user's needs, there is no need to configure the content generation system for each visualization software. As a result, versatility can be improved and the load on the hardware resources that make up the content generation system can be reduced. In other words, if the same hardware resources are used, more advanced processing can be performed.
  • the content generation system includes at least one user interface, at least one memory, and at least one processor.
  • the at least one user interface, at least one memory, and at least one processor are connected to each other so as to be able to communicate with each other.
  • the at least one user interface, at least one memory, and at least one processor may be provided in a single device to form a centralized processing system, or may be provided in a distributed manner in multiple devices connected to each other so as to form a distributed processing system.
  • the distributed processing system referred to here can be interpreted as a broad concept including local or cloud, or a combination thereof.
  • the user terminal device may be included in the multiple devices that make up the distributed processing system.
  • the user terminal device may function as a content generation system by installing a program related to the content generation system in the user terminal device and executing it in at least one processor provided in the user terminal device.
  • the user interface is an input device and an output device provided in the user terminal device.
  • the input device is, for example, a keyboard or a pointing device.
  • the output device is, for example, a display, a projector, a printer, a speaker, or an earphone.
  • the content generation system is capable of accepting input of communicative human language information via a user interface and capable of outputting content.
  • the content generation system may be based on a large-scale language model.
  • the large-scale language model may be multimodal.
  • the content generation system functions as a chatbot capable of inputting and outputting communicative human language information via a user interface.
  • the chatbot is based on a large-scale language model (LLM-based chatbot).
  • the content generation system is capable of repeatedly transmitting communicative human language information via a user interface.
  • the content generation system is capable of repeatedly transmitting communicative human language information to and from the large-scale language model.
  • the content generation system is capable of maintaining a context during the repeated transmission of information via the user interface or to and from the large-scale language model.
  • the content generation system is not incorporated in a specific visualization software and is not associated solely with a specific visualization software.
  • the visualization software to be selected is not determined before the content generation system is started to be used.
  • the contents include communicative human language information and visual information.
  • the contents are generated based on "communicative human language information inputted through a user interface".
  • the "communicative human language information inputted through a user interface” has the nature of a query, such as "Please create explanatory materials about technical matter X" in the embodiment described below, and includes the subject of the content.
  • the query here consists of a command, request, inquiry, or any combination of these to the content generation system (e.g., "Please create explanatory materials about ⁇ ").
  • the query may include, for example, a command, request, inquiry, or any combination of these about the specifications of the content (e.g., amount, language, layout).
  • the subject is, for example, a matter that a user wants to explain or communicate to a person (e.g., "technical matter X").
  • the subject includes, for example, a theme and a topic.
  • the "communicative human language information inputted through a user interface” does not necessarily have to be inputted all at once, and may be inputted as a result of multiple inputs and outputs via the user interface between the user and the content generation system.
  • the "communicative human language information input via a user interface” does not include information indicating the name of the visualization software.
  • the visualization software is selected by the content generation system, not by the user. However, information indicating the name of the visualization software may be input to the content generation system.
  • the content is created to, for example, explain or describe the subject matter in response to the query.
  • the communicative human language information and visual information contained in the content are content-wise related to each other and to the subject matter.
  • the content is created to be capable of enhancing a person's understanding of the subject matter.
  • the content is created to explain or describe the subject matter in more detail.
  • the communicative human language information in the content has a greater amount of text than the communicative human language information entered via a user interface.
  • the visual information is added, for example, to complement the explanation or description provided by the communicative human language information.
  • Contents include, for example, documents, presentation materials, and videos.
  • Documents include, for example, technical explanation documents, intellectual property rights-related documents submitted to administrative agencies or courts, intellectual property rights appraisal documents, and intellectual property rights research documents.
  • the content generation system may be capable of generating these documents themselves or their drafts.
  • Intellectual property rights-related documents include, for example, patent specifications, OA (office action) response documents, and arbitration or court documents.
  • Intellectual property rights appraisal documents may be about infringement or non-infringement, or about the validity of rights.
  • Intellectual property rights research documents may be about infringement or non-infringement, prior art research, or trend research. Note that information that is primarily visual and has communicative human language information added to it does not fall under the "content” defined here. For example, movies, dramas, game videos, sports videos, etc. do not fall under the "content” defined here.
  • Communicative human language information is language information that can be understood, recognized, and memorized by humans. Communicative human language information is based on the language system that humans use in daily conversation. Communicative human language information is transmitted as text information or audio information. Programming languages do not fall under communicative human language information. Programming languages here include not only low-level languages (machine languages and assembler languages) but also high-level languages (interpreted languages and compiled languages). Although communicative human language information is used as the information below, the information below may include a programming language in addition to the communicative human language information. The programming language in this case is not used for execution by a processor, but is used for presenting information to a person. In one embodiment, the information below does not include a programming language transmitted to a processor for execution. Information input to a content generation system via a user interface Information input to a large-scale language model to operate the large-scale language model Information included in the content together with visual information
  • the visual information may be an image or a video.
  • An image is static visual information.
  • a video is dynamic visual information.
  • An image may be a visualization or a non-visualization.
  • a visualization is an image obtained from data or information. Data is, for example, the source of information.
  • Examples of visualizations include figures, graphs, charts, diagrams, plots, histograms, tables, and matrices. Examples of figures include contour maps, contour plots, vector diagrams, isosurface maps, mechanical drawings, engineering drawings, and patent drawings. Examples of non-visualizations include photographic images that have actually been taken.
  • a visualization may be generated based on one or more non-visualizations.
  • a video may be an animation or simulation, or may be a live-action capture.
  • At least a portion of the visual information included in the content may have been generated, for example, by visualization software or by a content generation system that modifies the product of the visualization software.
  • Visual information obtained by searching the Internet does not, in itself, constitute visual information generated by visualization software in a content generation system.
  • An animation or simulation may be generated based on one or more live action captures.
  • An animation or simulation may include one or more non-visualizations and/or one or more visualizations.
  • a user interface refers to a device and/or equipment included in a content generation system that transmits information between the content generation system and a user.
  • the content generation system is configured without including a user terminal device and is connected to a user terminal device so that it can communicate with the user terminal device
  • an example of the user interface is a communication module included in the content generation system and capable of communicating with the user terminal device.
  • the content generation system is configured to include a user terminal device
  • an example of the user interface is an input device and an output device of the user terminal device.
  • Processors include central processing units (CPUs), microprocessors, general-purpose processors, digital signal processors (DSPs), graphics processing units (GPUs), controllers, microcontrollers, programmable logic devices (PLDs), field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), and the like. Multiple processors can be configured by any combination of these. Execution of a program by at least one processor is not limited to execution of the entire program by one processor, but may also be execution of the program by parallel processing of a multiprocessor constructed by multiple processors. In a multiprocessor, either symmetric or asymmetric multiprocessing may be performed. A multiprocessor may be either tightly coupled or loosely coupled. A processor may be a multicore processor.
  • Memory includes RAM, ROM, non-volatile RAM, PROM, EPROM, EEPROM, flash memory, magnetic data storage, optical data storage, registers, or combinations thereof. There may be one or more memories.
  • a large-scale language model is a language model that has a large number of parameters.
  • a large-scale language model is capable of executing natural language processing tasks.
  • a large-scale language model may be a natural language generation model.
  • a natural language generation model is based on a large-scale language model and is capable of creating sentences consisting of new communicative human language information from input communicative human language information. Note that creating sentences consisting of new communicative human language information from input communicative human language information is an example of a natural language processing task.
  • the number of parameters of the large-scale language model may be, for example, 1 billion or more, 10 billion or more, or 100 billion or more.
  • the language model refers to a communicative human language modeled using the occurrence probability of words.
  • the large-scale language model is a language model capable of inference using a method such as zero-shot learning (ZSL), one-shot learning (OSL), or few-shot learning (FSL) without fine tuning.
  • the large-scale language model is configured to execute a task and output in response to an input prompt.
  • the prompt includes communicative human language information. Examples of large-scale language models are not particularly limited, and include, for example, GPT-3, GPT-4, GShard, Switch Transformer, Gopher, and HyperCLOVA.
  • the large-scale language model may be included in the content generation system, or may be capable of communicating with the content generation system without being included in the content generation system.
  • the visualization software is not particularly limited as long as it is software that generates visual information using input text information, and may be, for example, data visualization software or a visual foundation model (VFM).
  • VFM visual foundation model
  • each piece of visual information may be generated by a different visualization software.
  • the processor may select visualization software for each piece of visual information.
  • Process (A) includes processes (B) to (D). In the embodiment described below, processes (B), (C), and (D) are executed in this order. In addition, in each of processes (B) to (D), communicative human language information is supplied from the content generation system to the large-scale language model. However, the order of processes (B) to (D) is not particularly limited. Processes (B) to (D) do not need to be strictly distinguished from one another in terms of time and/or processing content. It is sufficient that process (A) (i.e. processes (B) to (D)) and process (E) are completed before content is output in process (F). When each of processes (B) to (D) includes multiple processes, the processes related to processes (B) to (D) may be executed alternately. The supply of communicative human language information from the content generation system to the large-scale language model may be executed in common for all or any two of processes (B) to (D). In any case, the result of the process executed earlier can be used for the process executed later.
  • a part of the communicative human language information inputted through the user interface may be used, or the whole of it may be used.
  • the same or different communicative human language information as the communicative human language information inputted through the user interface may be supplied to the large-scale language model.
  • the communicative human language information inputted through the user interface is mainly used as the information supplied to the large-scale language model in the process (B).
  • a task may be set based on a subject included in the communicative human language information inputted through the user interface, and a prompt based on the task may be supplied to the large-scale language model.
  • a prompt may be supplied to the large-scale language model based on the subject included in the communicative human language information inputted through the user interface and the predetermined task.
  • the prompt can be interpreted as the input matter, and the content-oriented communicative human language information is, for example, a detailed explanation of the input matter.
  • the process (B) may be repeatedly executed multiple times. In this case, first, one or more items (e.g., words, phrases, expressions, or sentences) may be extracted from the description of the subject as communicative human language information obtained from the large-scale language model, and the extracted items may be supplied to the large-scale language model.
  • items e.g., words, phrases, expressions, or sentences
  • an explanation of the subject when an explanation of the subject is obtained from the large-scale language model, items of lower-level concepts related to the subject may be extracted from the explanation, and the extracted items may be supplied to the large-scale language model.
  • the content generated through such a process can include a detailed explanation or description of the subject, as well as a more detailed explanation or description of the items included in the explanation or description.
  • a part of the communicative human language information inputted through the user interface may be used, or the whole may be used.
  • the communicative human language information inputted through the user interface is mainly used as the information supplied to the large-scale language model in process (C).
  • Selecting visualization software based on the communicative human language information acquired from the large-scale language model includes, for example, the following aspects. If the communicative human language information includes information indicating the name of the visualization software, the processor can directly select the visualization software by the name. In another aspect, the processor can use the communicative human language information to perform a web search to obtain communicative human language information and/or visual information, and select the visualization software based on this information.
  • the web search may be a search for visual information (e.g., an image search) or a search for communicative human language information (e.g., a text search).
  • the visualization software selectable by the content generation system does not have to be specified before the content generation system is started to be used.
  • the visualization software may be selected from one or more visualization software available via a network such as the Internet.
  • multiple visualization software selectable by the content generation system may be identified before starting to use the content generation system.
  • the visualization software may be selected based on a subject included in communicative human language information input via a user interface. Since the visualization software is selected by the content generation system, it is not necessary for the user to select the visualization software according to the desired content.
  • the text information obtained based on the output of the large-scale language model may be the same as the information output from the large-scale language model, or may be modified from the information output from the large-scale language model.
  • This text information has information according to the communicative human language information input via the user interface, i.e., information corresponding to or related to the communicative human language information. However, this text information is different from the communicative human language information itself input via the user interface.
  • the input of the text information in a format that can be input to the visualization software is not required from the user, but is executed by the processor in process (D).
  • the "text information obtained based on the output of the large-scale language model" in process (D) means that, of the input via the user interface and the output of the large-scale language model, which may be text information sources in the content generation system, the output of the large-scale language model is the basis of the text information.
  • the input via the user interface may also be the basis of the text information.
  • process (E) content is generated that includes at least a portion of the communicative human language information acquired in process (B) or a modified version of the communicative human language information, and at least a portion of the visual information acquired in process (D) or a modified version of the visual information.
  • “at least a portion” means that the content does not necessarily include all of the communicative human language information acquired in process (B), and does not necessarily include all of the visual information acquired in process (D).
  • the "modified version” refers to information acquired in process (B) or (D) that has been modified by the content generation system. In other words, the content may include a modified version instead of or in addition to the information acquired in process (B) or (D).
  • the modification here means that a change that does not substantially change the information to be transmitted is allowed.
  • the layout of the communicative human language information and visual information in the content is not particularly limited, and may be determined automatically by a processor, for example, or in response to a request input via a user interface.
  • the manner in which the content is output is not particularly limited, and may involve providing a file of the content or displaying the content itself.
  • the present invention provides a content generation system that utilizes LLM to efficiently generate content that is easier for humans to understand.
  • FIG. 1A is a system outline diagram for explaining a content generation system according to an embodiment of the present disclosure
  • FIG. 1B is a flowchart for explaining processing performed by the content generation system.
  • FIG. 1(a) is a system overview diagram for explaining a content generation system 1 according to an embodiment of the present disclosure.
  • the content generation system 1 includes a processor 2, a memory 3 and a communication module 4 communicatively connected to the processor 2.
  • the memory 3 stores programs for executing processes (A) to (F).
  • the processor 2 executes the programs.
  • the communication module 4 corresponds to a user interface.
  • the number of processors 2 and memories 3 is not particularly limited, and each may be either one or more.
  • the hardware configuration of the content generation system 1 is not particularly limited.
  • the content generation system 1 may be configured by a single server device, or by multiple server devices communicatively connected to each other. In this case, the multiple server devices may be configured to provide a cloud computing service.
  • the communication module 4 enables communication between the processor 2 and each of the user terminal device 11, the large-scale language model 6, and the visualization software 7.
  • the multiple user terminal devices 11 are each capable of communicating with the content generation system 1 through the network 12. There is no particular limit to the number of user terminal devices 11.
  • a PC terminal device, a tablet terminal device, and a mobile phone device are shown as examples of the user terminal devices 11.
  • the user terminal devices 11 are not limited to these examples.
  • Various types of terminal devices that can be used by users can be used as the user terminal devices 11.
  • the network 12 enables communication between a plurality of user terminal devices 11 and the content generation system 1.
  • the network 12 is not particularly limited, and can be constructed using various wired networks, various wireless networks, or a combination of these. There is no particular limit to the communication method. There is no particular limit to the communication protocol.
  • the large-scale language model 6 is stored in one server device, or in multiple server devices that are connected to each other so that they can communicate with each other.
  • the one or multiple server devices are capable of communicating with the content generation system 1.
  • the large-scale language model 6 outputs to the content generation system 1 in response to an input from the content generation system 1.
  • the input is, for example, communicative human language information.
  • the output is, for example, communicative human language information in response to the input.
  • the output is, for example, text information that is a response to the input and is used as input to the visualization software 7.
  • the multiple visualization software 7 are each stored in one server device, or in multiple server devices that are communicatively connected to each other.
  • the one or more server devices in which each visualization software 7 is stored can communicate with the content generation system 1 via a network such as the Internet.
  • each visualization software 7 is available to the content generation system 1 via a network 12 such as the Internet.
  • One or more visualization software 7 are selected by the content generation system 1 from the multiple visualization software 7.
  • the selected visualization software 7 outputs to the content generation system 1 in response to an input from the content generation system 1.
  • the input is, for example, text information obtained from the large-scale language model 6.
  • the output is, for example, visual information generated using the text information.
  • FIG. 1(b) is a flowchart for explaining the processing performed by the content generation system 1.
  • a user inputs "Please create explanatory materials about technical matter X" into the user terminal device 11 as communicative human language information, but this is an example, and the input communicative human language information is not limited to this example.
  • the user inputs communicative human language information "Please create explanatory materials about technical matter X" to the user terminal device 11 (step S111).
  • the user terminal device 11 for example, software or an application related to the provision of a service by the content generation system 1 is running, or a website related to the provision of the service is displayed in a web browser.
  • the user inputs communicative human language information.
  • the input communicative human language information has a subject (here, "Please create explanatory materials about technical matter X").
  • the input communicative human language information is transmitted by the user terminal device 11 to the content generation system 1 via the network 12 (step S112).
  • the processor 2 receives the communicative human language information via the communication module 4 (step S11).
  • the processor 2 uses each of the large-scale language model 6 and the visualization software 7 (processing (A)).
  • the visualization software 7 is selected and managed using the output of the large-scale language model 6 by the content generation system 1.
  • the management is targeted at the selected visualization software 7. That is, the management is to utilize the selected visualization software 7 for explaining or describing a subject included in the communicative human language information inputted via a user interface.
  • the process (A) includes the following processes (B) to (D). Note that in the process (A), iterative processing is possible between the content generation system 1 and the large-scale language model 6 and/or the visualization software 7.
  • step S11 the processor 2 acquires communicative human language information to be included in the content by the large-scale language model 6 based on the communicative human language information input through the communication module 4 in step S11 (processing (B)).
  • the processor 2 first supplies the communicative human language information to the large-scale language model 6 (step S12).
  • step S12 the communicative human language information for the content is supplied from the large-scale language model 6 to the content generation system 1 (step S62).
  • step S13 the processor 2 acquires communicative human language information to be included in the content by the large-scale language model 6 (step S13).
  • the process (B) is repeatedly executed in this embodiment, but is not limited to this example.
  • the processor 2 acquires a detailed description of the technical matter X as the communicative human language information to be included in the content by the process (B) that is repeated multiple times.
  • This description includes a general formula for explaining the technical matter X and variables included in the general formula.
  • the processor 2 uses the communicative human language information input via the communication module 4 to obtain communicative human language information for selecting visualization software 7 capable of generating visual information to be included in the content using the large-scale language model 6, and selects the visualization software 7 based on the obtained communicative human language information (Process (C)).
  • the processor 2 supplies communicative human language information to the large-scale language model 6 (step S14).
  • the communicative human language information acquired in the process (B) is used in the process (C).
  • the processor 2 acquires a general formula for explaining the technical matter X and the variables contained in the general formula.
  • the processor 2 supplies, as the communicative human language information, to the large-scale language model 6, a query about visualization software 7 capable of generating visual information using the general formula and the variables.
  • communicative human language information for selecting visualization software 7 is supplied from the large-scale language model 6 to the content generation system 1 (step S64).
  • the processor 2 acquires communicative human language information for selecting visualization software 7 from the large-scale language model 6 (step S15).
  • the communicative human language information acquired in step S15 includes the name of visualization software 7 capable of generating visual information using the general formula and variables.
  • the processor 2 selects visualization software 7 based on the acquired communicative human language information (step S16).
  • the processor 2 obtains visual information to be included in the content by operating the visualization software 7 selected in step S16 using text information obtained based on the output of the large-scale language model 6 in response to the communicative human language information input via the communication module 4 (processing (D)).
  • process (D) first, the processor 2 supplies communicative human language information to the large-scale language model 6 (step S17).
  • processes (B) and (C) are performed before process (D), so the communicative human language information acquired in processes (B) and (C) is used in process (D). That is, the processor 2 supplies, as communicative human language information, to the large-scale language model 6 an inquiry about the type and value of data to be input to the visualization software 7 to obtain visual information showing technical matter X.
  • text information is supplied from the large-scale language model 6 to the content generation system 1 (step S67).
  • the processor 2 obtains the text information from the large-scale language model 6 (step S18).
  • the type and value of data to be input are supplied in a format that can be input to the visualization software 7.
  • the processor 2 then supplies the text information to the visualization software 7 (step S19).
  • the visualization software 7 is operably stored in one or more server devices.
  • the visualization software 7 operates using the text information to generate visual information.
  • the text information includes the type and value of data to be input to the visualization software 7.
  • the text information is acquired in a format that can be input to the visualization software 7. Therefore, the processor 2 can supply this text information to the visualization software 7 and operate the visualization software 7 using this text information.
  • visual information is generated by the visualization software 7.
  • the visual information generated in this embodiment is a simulation result and a graph showing a specific example of technical item X.
  • the visual information generated in this manner is supplied from the visualization software 7 to the content generation system 1 (step S79).
  • the processor 2 acquires visual information to be included in the content from the visualization software 7 (step S20).
  • Process (E) After processes (B) to (D), the processor 2 generates content including at least a portion of the communicative human language information acquired in process (B) and at least a portion of the visual information acquired in process (D) (process (E)).
  • the processor 2 outputs the content generated in the process (E) from the communication module 4 to the user terminal device 11 via the network 12 (process (F)).
  • the user terminal device 11 receives the content (step S121) and outputs the content (step S122).
  • the content output manner is not particularly limited.
  • the content output may be a display of the content on a display provided in the user terminal device 11, or a data file of the content may be provided.
  • the present invention is not limited to the embodiment described above.
  • the present invention can be implemented in other embodiments, and various modifications can be added.
  • Content generation system 2 Processor 3: Memory 4: Communication module 6: Large-scale language model 7: Visualization software 11: User terminal device 12: Network

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Computational Linguistics (AREA)
  • Artificial Intelligence (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • User Interface Of Digital Computer (AREA)
PCT/JP2023/015385 2023-04-17 2023-04-17 コンテンツ生成システム Ceased WO2024218839A1 (ja)

Priority Applications (7)

Application Number Priority Date Filing Date Title
PCT/JP2023/015385 WO2024218839A1 (ja) 2023-04-17 2023-04-17 コンテンツ生成システム
JP2025514907A JPWO2024218839A1 (https=) 2023-04-17 2023-04-17
CN202380096996.8A CN120917439A (zh) 2023-04-17 2023-04-17 内容生成系统
ES202590064A ES3041845A2 (es) 2023-04-17 2023-04-17 Sistema de generacion de contenido
GB2518241.1A GB2644497A (en) 2023-04-17 2023-04-17 Content generation system
US18/637,896 US20240346237A1 (en) 2023-04-17 2024-04-17 Content generation system
DE102024203563.7A DE102024203563A1 (de) 2023-04-17 2024-04-17 Content-Erzeugungssystem

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2023/015385 WO2024218839A1 (ja) 2023-04-17 2023-04-17 コンテンツ生成システム

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/637,896 Continuation-In-Part US20240346237A1 (en) 2023-04-17 2024-04-17 Content generation system

Publications (1)

Publication Number Publication Date
WO2024218839A1 true WO2024218839A1 (ja) 2024-10-24

Family

ID=92908739

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2023/015385 Ceased WO2024218839A1 (ja) 2023-04-17 2023-04-17 コンテンツ生成システム

Country Status (7)

Country Link
US (1) US20240346237A1 (https=)
JP (1) JPWO2024218839A1 (https=)
CN (1) CN120917439A (https=)
DE (1) DE102024203563A1 (https=)
ES (1) ES3041845A2 (https=)
GB (1) GB2644497A (https=)
WO (1) WO2024218839A1 (https=)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20250103797A1 (en) * 2023-09-26 2025-03-27 Dropbox, Inc. Generating field objects for auto-populating fillable documents utilizing a large language model

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR102897132B1 (ko) * 2024-12-06 2025-12-08 주식회사 일만백만 인공지능 기반 유저의 텍스트로부터 감성을 추론하고 애니메이션과 모션 스타일을 설정하여 애니메이션을 생성하는 시스템 및 방법

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003076678A (ja) * 2001-08-31 2003-03-14 Just Syst Corp 情報処理装置、情報処理方法、及び情報処理プログラム
CN111881307A (zh) * 2020-07-28 2020-11-03 平安科技（深圳）有限公司 一种演示文稿生成方法、装置、计算机设备及存储介质
JP2022179507A (ja) * 2018-06-21 2022-12-02 株式会社Tsunagu.AI ウェブコンテンツ自動生成システム

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10896214B2 (en) * 2018-06-01 2021-01-19 Accenture Global Solutions Limited Artificial intelligence based-document processing
US11741306B2 (en) 2019-12-18 2023-08-29 Microsoft Technology Licensing, Llc Controllable grounded text generation
US12536557B2 (en) * 2022-04-21 2026-01-27 Merchant & Gould P.C. Risk assessment management system and method
US12536387B2 (en) * 2023-03-10 2026-01-27 Microsoft Technology Licensing, Llc Task decomposition for LLM integrations with spreadsheet environments

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003076678A (ja) * 2001-08-31 2003-03-14 Just Syst Corp 情報処理装置、情報処理方法、及び情報処理プログラム
JP2022179507A (ja) * 2018-06-21 2022-12-02 株式会社Tsunagu.AI ウェブコンテンツ自動生成システム
CN111881307A (zh) * 2020-07-28 2020-11-03 平安科技（深圳）有限公司 一种演示文稿生成方法、装置、计算机设备及存储介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20250103797A1 (en) * 2023-09-26 2025-03-27 Dropbox, Inc. Generating field objects for auto-populating fillable documents utilizing a large language model
US12462096B2 (en) * 2023-09-26 2025-11-04 Dropbox, Inc. Generating field objects for auto-populating fillable documents utilizing a large language model

Also Published As

Publication number Publication date
CN120917439A (zh) 2025-11-07
ES3041845A2 (es) 2025-11-14
US20240346237A1 (en) 2024-10-17
DE102024203563A1 (de) 2024-10-17
GB2644497A (en) 2026-04-15
JPWO2024218839A1 (https=) 2024-10-24

Similar Documents

Publication Publication Date Title
JP7206419B2 (ja) 人工知能推奨モデルの特徴処理方法、装置、電子機器、及びコンピュータプログラム
CA2915619C (en) Method and apparatus for customized software development kit (sdk) generation
US20150040098A1 (en) Systems and methods for developing and delivering platform adaptive web and native application content
WO2017162024A1 (zh) 组件和模板的可视化开发方法及系统、存储介质、设备
WO2024218839A1 (ja) コンテンツ生成システム
CN118410133A (zh) 用于数字助手创建的方法、装置、设备和存储介质
Zhang et al. UniRLTest: Universal platform-independent testing with reinforcement learning via image understanding
US12147769B2 (en) Method for aligning text with media material, apparatus and storage medium
CN114546220A (zh) 页面配置方法及装置、存储介质、计算机设备
Iglesias et al. Search-Based Negative Prompt Optimisation for Text-to-Image Generation
WO2025026189A1 (zh) 用于提示词管理的方法、装置、设备和存储介质
CN120457410A (zh) 使用语法映射来生成合成训练数据
Dogdu et al. A data-model driven web application development framework
US20100223594A1 (en) Method and apparatus for implementing a composable control architecture
KR102211962B1 (ko) 실시간 서버 푸시형 웹 스크래핑 방법
Chelemen Modeling a web application for cloud content adaptation with ASMs
JP7843799B2 (ja) 機能マップ生成方法及びシステム
CN115552405A (zh) 制造系统设计验证装置
TWI914925B (zh) 動態生成投射信息提供者消息的基於人工智能的答案的方法及系統
US20240370236A1 (en) Managing an app, developing an app including an event artifact, method, and system
US20250247249A1 (en) Device and method for extracting and structuring verfiable personal data from user input and output data captured on a device based on a multi-modal and language ai model
Soylu et al. Mashups and widget orchestration
CN118535705A (zh) 文本生成方法、装置、设备、计算机可读存储介质及产品
Bibichkov et al. Ontological approach to development of web-content generation method
Kaltz et al. Supporting Systematic Usage of Context in Web Applications.

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: 23933986

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 202380096996.8

Country of ref document: CN

ENP Entry into the national phase

Ref document number: 2025514907

Country of ref document: JP

Kind code of ref document: A

WWE Wipo information: entry into national phase

Ref document number: 2025514907

Country of ref document: JP

WWE Wipo information: entry into national phase

Ref document number: P202590064

Country of ref document: ES

WWE Wipo information: entry into national phase

Ref document number: 112025000160084

Country of ref document: IT

ENP Entry into the national phase

Ref document number: 202518241

Country of ref document: GB

Kind code of ref document: A

Free format text: PCT FILING DATE = 20230417

WWE Wipo information: entry into national phase

Ref document number: 202517106387

Country of ref document: IN

WWP Wipo information: published in national office

Ref document number: 202380096996.8

Country of ref document: CN

WWP Wipo information: published in national office

Ref document number: P202590064

Country of ref document: ES

NENP Non-entry into the national phase

Ref country code: DE

WWP Wipo information: published in national office

Ref document number: 202517106387

Country of ref document: IN

122 Ep: pct application non-entry in european phase

Ref document number: 23933986

Country of ref document: EP

Kind code of ref document: A1