Method and system for providing special document sharing platform

Abstract

The invention relates to a method and a system for providing a special document sharing platform. The method of providing a dedicated document sharing platform through a platform application executed by at least one processor of a terminal comprises the steps of: obtaining a first dedicated document; using the question and answer language model to generate question and answer content for the first special document; generating promotion content including an online promotion material for the first dedicated document based on the question and answer content; based on the promotion content, providing a promotion material production working space; using the promotion material production workspace to determine promotion start content including promotion content to be registered on the dedicated document sharing platform; and providing the promotion start content using the dedicated document sharing platform.

Description

Technical Field

[0001] The present disclosure generally relates to methods and systems for providing a dedicated document sharing platform. More specifically, some embodiments of the present disclosure relate to methods and systems for providing a dedicated document sharing platform, which are used to generate question-and-answer content for a dedicated document using a predetermined deep learning neural network and provide promotional content based on the generated question-and-answer content. Background Art

[0002] Generally, most technology information providing services using the Internet simply display a list of information provided by an operator when a user accesses a technology information providing site via the Internet, or provide simple HTML data in the form of a summary when a bookmark is clicked.

[0003] In addition, various companies, research institutes, public institutions and / or universities have accumulated data on industrial technological knowledge information and research results, but failure to properly utilize this data is a factor that hinders the country's industrial technological development.

[0004] Therefore, various companies, research institutions, public institutions and / or universities need to distribute accumulated industrial technical knowledge information and data about research results while receiving appropriate compensation, and need systems and service methods for facilitating the distribution and transaction of industrial technical knowledge information by providing reliable data to those who request industrial technical knowledge information.

[0005] Therefore, various sharing platforms for the above-mentioned specialized documents (eg, these specialized documents and / or reports) are currently being developed.

[0006] Such platforms enable sharing of various types of data, allow specialized documents to be widely disclosed and thus enable more research, and allow for more transparent research by sharing data, code, and / or drafts resulting from the research process.

[0007] Furthermore, since the research process is recorded, it can provide a head start in the event of future controversy over research ideas, and it can also recommend research works of scholars in various fields and stages according to research fields and research topics, which has the effect of causing a breakdown of citation trends where the rich get richer and the poor get poorer, with papers by scholars known in the relevant academic fields being primarily cited.

[0008] For example, such platforms include Arxiv, Open Science Framework (OSF), Dataverse, Figshare, and / or Github.

[0009] However, most of the above-mentioned conventional dedicated document sharing platforms may provide search results that simply summarize or list at least one dedicated document related to the user's desired content, and thus may require additional efforts to fully understand the content of the document from the search results.

[0010] Furthermore, conventional dedicated document sharing platforms may have limitations in that the accuracy and quality of search results are relatively low for documents that require innovation in knowledge processing and inference, such as dedicated documents regarding deep technical information (e.g., scientific and technical papers).

[0011] In addition, most private document sharing platforms can be implemented as paid services, which provide appropriate compensation to providers of private documents as the number of subscriptions to private documents on the platform increases. However, most information providers may have limited knowledge of promotion methods or content generation methods to increase the amount of sharing, and therefore find it difficult to actively promote and distribute the private documents they have written.

[0012] [Related technical literature]

[0013] [Patent Document]

[0014] Korean Patent No.10-1341948B1 Summary of the invention

[0015] According to some embodiments of the present disclosure, a method and system for providing a dedicated document sharing platform can generate question and answer content for a dedicated document using a predetermined deep learning neural network, and provide promotional content based on the generated question and answer content.

[0016] According to certain embodiments of the present disclosure, a dedicated document sharing platform provides a method and system that can generate question-answering content using a language model dedicated to a dedicated document.

[0017] Furthermore, according to some embodiments of the present disclosure, a dedicated document sharing platform providing method and system may provide promotional content using a user interface (UI), through which the generated question and answer content may be freely edited.

[0018] The embodiments of the present disclosure and the technical objectives to be achieved by the present disclosure are not limited to the above-mentioned technical objectives, and there may be other technical objectives.

[0019] According to an embodiment of the present disclosure, a method for providing a dedicated document sharing platform is a method for providing a dedicated document sharing platform through a platform application executed by at least one processor of a terminal, the method comprising: obtaining a first dedicated document; generating question and answer content for the first dedicated document based on a question and answer language model; generating promotional content based on the question and answer content, the promotional content being online promotional material for the first dedicated document; providing promotional content based on a promotional material production workspace; determining promotion start content based on the promotional material production workspace, the promotion start content being promotional content to be registered on the dedicated document sharing platform; and providing promotion start content based on the dedicated document sharing platform.

[0020] On the other hand, generating question and answer content may include inputting a first dedicated document into a question and answer language model, obtaining at least one question and answer data including question data and answer data from the question and answer language model according to the first dedicated document, and generating question and answer content based on the at least one question and answer data.

[0021] On the other hand, obtaining at least one question and answer data may include obtaining at least one question data based on a question deep learning model included in the question and answer language model, and obtaining at least one answer data corresponding to each question data in the at least one question data based on an answer deep learning model included in the question and answer language model.

[0022] On the other hand, the answer deep learning model can be a deep learning model for outputting answer data based on a multi-step inference process, wherein the multi-step inference process includes an association selection process for determining evidence paragraphs related to question data in a first special document, a basic principle generation process for obtaining at least one basic principle data related to the question data based on the evidence paragraphs, and a system composition process for performing data processing based on the basic principle data.

[0023] On the other hand, the promotional content may be content data configured based on a predetermined data volume condition.

[0024] In another aspect, determining the promotion start content may include generating promotion edited content, where the promotion edited content is promotional content edited based on the promotional material production workspace according to user input.

[0025] On the other hand, generating the promotional editorial content may include at least one of the following steps: generating the promotional editorial content based on user input for editing the composition of the question and answer content in the promotional content, and generating the promotional editorial content based on user input for editing the content of the question and answer content in the promotional content.

[0026] On the other hand, generating the promotional editorial content may include providing an editing assistance tool which is a user interface for providing at least one of a predetermined image, a table, and a formula related to the first dedicated document in a form that can be inserted into question and answer content in the promotional content.

[0027] In another aspect, generating the promoted editorial content may include providing a related document search function that automatically detects and provides at least one other specialized document related to the question and answer content in the promoted content.

[0028] In another aspect, determining the promotion start content may include determining the promotion content or the promotion editorial content as the promotion start content according to user input based on the promotion material production workspace.

[0029] In another aspect, the method of providing a dedicated document sharing platform may further include providing a dedicated document search function based on the promotion start content according to the dedicated document sharing platform.

[0030] In another aspect, the method of providing a dedicated document sharing platform may further include providing dedicated document sharing and citing functions based on the promotion start content according to the dedicated document sharing platform.

[0031] On the other hand, providing dedicated document sharing and reference functions may include at least one of the following steps: providing a promotional material identification code including an access uniform resource locator (URL) and a timestamp associated with the promotion start content, and providing a question and answer identification code including an access URL and a timestamp associated with the question and answer content in the promotion start content.

[0032] In another aspect, the method for providing a dedicated document sharing platform may further include providing a reader question recommendation function, that is, obtaining and providing question data created by a user who receives the first dedicated document based on the dedicated document sharing platform.

[0033] According to an embodiment of the present disclosure, a system for providing a dedicated document sharing platform includes: at least one memory in which a platform application is stored; and at least one processor configured to read the platform application stored in the memory and provide the dedicated document sharing platform, wherein the instructions of the platform application include instructions for performing the following steps: obtaining a first dedicated document; generating question and answer content for the first dedicated document based on a question and answer language model; generating promotional content based on the question and answer content, the promotional content being online promotional material for the first dedicated document; providing the promotional content based on a promotional material production workspace; determining the promotion start content based on the promotional material production workspace, the promotion start content being the promotional content to be registered on the dedicated document sharing platform; and providing the promotion start content based on the dedicated document sharing platform.

[0034] According to some embodiments of the present disclosure, a method and system for providing a dedicated document sharing platform may use a predetermined deep learning neural network to generate question-and-answer content for a dedicated document, and provide promotional content based on the generated question-and-answer content, thereby enabling users to automatically obtain data in a question-and-answer format that effectively includes the content of the dedicated document without making any additional efforts, and utilize promotional content based on the obtained data.

[0035] According to certain embodiments of the present disclosure, methods and systems for providing a dedicated document sharing platform can generate question-answering content using a language model dedicated to dedicated documents, thereby providing question-answering content that solves questions with illusions and ambiguous answers, which is a limitation of traditional general language models (e.g., OpenAI GPT, etc.).

[0036] Therefore, according to some embodiments of the present disclosure, methods and systems for providing a dedicated document sharing platform can use advanced inference / estimation functions implemented through a process similar to human cognitive inference to provide question-answering content that is faithful to the question, illusion control, and evidence.

[0037] In addition, according to certain embodiments of the present disclosure, the method and system for providing a dedicated document sharing platform can use a user interface (UI) to provide promotional content, through which the generated question and answer content can be freely edited, thereby enabling users to create customized promotional materials optimized for the form desired by the user with a higher degree of freedom, and actively use the customized promotional materials to perform promotional activities beyond the level of simply sharing dedicated documents.

[0038] Furthermore, according to some embodiments of the present disclosure, methods and systems for providing a dedicated document sharing platform may allow for easy generation and utilization of high-quality promotional materials optimized for user needs while including deep and practical question-and-answer content obtained from a deep learning model dedicated to dedicated documents. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a conceptual diagram of a system for providing a dedicated document sharing platform according to an embodiment of the present disclosure.

[0040] Figure 2 is a block diagram of a terminal according to an embodiment of the present disclosure.

[0041] Figure 3 is a conceptual diagram illustrating a question-answering language model according to an embodiment of the present disclosure.

[0042] Figure 4 and Figure 5 is a diagram illustrating a multi-step inference process of an answer deep learning model according to an embodiment of the present disclosure.

[0043] Figure 6 Detailed description is a flowchart illustrating a method for providing an author-side dedicated document sharing platform according to an embodiment of the present disclosure.

[0044] Figure 7 An example of question and answer content according to an embodiment of the present disclosure is shown.

[0045] Figure 8 An example of promotional content according to an embodiment of the present disclosure is shown.

[0046] Fig. 9 and Fig.10 is an exemplary diagram illustrating a method for adding question and answer data based on user input according to an embodiment of the present disclosure.

[0047] Fig.11 is a diagram illustrating a method of providing an editing assistance tool and / or a related document search function according to an embodiment of the present disclosure.

[0048] Fig.12 It is a flowchart for illustrating a method for providing a reader-specific document sharing platform according to an embodiment of the present disclosure.

[0049] Fig.13 is an exemplary diagram for illustrating a method for providing dedicated document sharing and citing functions according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0050] The present disclosure may be modified in various ways and may have various embodiments, and specific embodiments are shown in the drawings and described in detail in the specific embodiments. The effects and features of the present disclosure and methods for implementing the same will become clear with reference to the embodiments described in detail below in conjunction with the drawings. However, the present disclosure is not limited to the embodiments disclosed below and may be implemented in various forms.

[0051] Figure 1 is a conceptual diagram of a system for providing a dedicated document sharing platform according to an embodiment of the present disclosure.

[0052] refer to Figure 1 According to an embodiment of the present disclosure, the system 1000 for providing a dedicated document sharing platform can be configured to provide a dedicated document sharing platform providing service, which provides a service for generating question and answer content for a dedicated document using a predetermined deep learning neural network and providing promotional content based on the generated question and answer content.

[0053] In one embodiment, a system 1000 for providing a dedicated document sharing platform configured to provide a dedicated document sharing platform providing service may include a terminal 100 , a platform providing server 200 , and a network 300 .

[0054] Here, the terminal 100 and / or the platform providing server 200 may be connected through a network 300 .

[0055] Here, the network 300 according to the embodiment may mean a connection structure for information exchange between nodes, such as the terminal 100 and / or the platform providing server 200 .

[0056] Examples of network 300 include, but are not limited to, a 3rd Generation Partnership Project (3GPP) network, a Long Term Evolution (LTE) network, a Worldwide Interoperability for Microwave Access (WIMAX) network, the Internet, a local area network (LAN), a wireless local area network (WLAN), a wide area network (WAN), a personal area network (PAN), a Bluetooth network, a satellite broadcast network, an analog broadcast network, and / or a digital multimedia broadcasting (DMB) network.

[0057] In the following, reference will be made to Figure 1 The terminal 100 and the platform providing server 200 included in the system 1000 for providing a dedicated document sharing platform are described in detail.

[0058] -Terminal 100

[0059] The terminal 100 according to an embodiment of the present disclosure may be a predetermined computing device on which a platform application (hereinafter referred to as an application) providing a dedicated document sharing platform providing service is installed.

[0060] For example, the terminal 100 may include a mobile computing device 100 - 1 and / or a desktop computing device 100 - 2 on which applications are installed.

[0061] Here, the mobile computing device 100 - 1 may be a mobile device on which an application is installed.

[0062] For example, the mobile computing device 100 - 1 may include a smart phone, a mobile phone, a digital broadcaster, a personal digital assistant (PDA), a portable multimedia player (PMP), and / or a tablet personal computer (PC), but is not limited thereto.

[0063] Furthermore, the desktop computing device 100 - 2 may be a wired and / or wireless communication based device on which an application is installed.

[0064] For example, the desktop computing device 100 - 2 may include a personal computer, such as a stationary desktop PC, a notebook computer, and / or an ultrabook.

[0065] According to an embodiment, the terminal 100 may further include a server-type computing device that provides a dedicated document sharing platform providing a service environment.

[0066] Figure 2 is a block diagram of a terminal according to an embodiment of the present disclosure.

[0067] Reference Figure 2 , the terminal 100 may include a memory 110, a processor component 120, a communication processor 130, an interface 140, an input system 150, a sensor system 160, and a display system 170. In one embodiment, one or more of the above components of the terminal 100 may be disposed in a housing.

[0068] Specifically, the memory 110 may store an application 111 .

[0069] Here, the application 111 may include one or more of various application programs, data, and instructions for providing a dedicated document sharing platform providing service environment.

[0070] That is, the memory 110 may store instructions, data, etc. that may be used to create a dedicated document sharing platform providing a service environment.

[0071] Furthermore, the memory 110 may include a program area and a data area.

[0072] Here, the program area according to the embodiment may be linked between an operating system (OS) booting the terminal 100 and functional elements.

[0073] In addition, the data area according to the embodiment may store data generated when the terminal 100 is used or a program is executed.

[0074] Furthermore, the memory 110 may include at least one non-transitory computer-readable storage medium and a temporary computer-readable storage medium.

[0075] For example, the memory 110 may be various storage devices such as a ROM, an EPROM, a flash drive, and a hard disk drive, and may include a web storage that performs a storage function of the memory 110 on the Internet.

[0076] The processor component 120 may include at least one processor configured to execute instructions of the application 111 stored in the memory 110 to perform various operations for creating or providing a dedicated document sharing platform providing service environment.

[0077] In an implementation, the processor component 120 may control the overall operation of the components of the terminal 100 through the application 111 of the memory 110 to provide a dedicated document sharing platform to provide services.

[0078] For example, the processor component 120 may include a system on chip (SOC) suitable for the terminal 100 including a central processing unit (CPU) and / or a graphics processing unit (GPU).

[0079] In addition, the processor component 120 may execute an operating system (OS) and / or application programs stored in the memory 110 .

[0080] In addition, the processor component 120 may control each component of the terminal 100 .

[0081] In addition, the processor assembly 120 may communicate internally with each component via a system bus and may include one or more bus structures including a local bus.

[0082] In addition, the processor component 120 may include, for example, but not limited to, at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor and / or other electrical units for performing functions.

[0083] The communication processor 130 may include one or more devices configured to communicate with one or more external devices. The communication processor 130 may perform communication via a wireless or wired network.

[0084] Specifically, the communication processor 130 may communicate with the terminal 100 storing a content source for creating a dedicated document sharing platform providing service environment.

[0085] Additionally, the communication processor 130 may communicate with various user input components, such as a controller, to receive user input.

[0086] In an embodiment, the communication processor 130 may transmit or receive various types of data related to services provided by the dedicated document sharing platform to or from another terminal 100 and / or an external server.

[0087] The communication processor 130 can wirelessly send or receive data to or from at least one of a base station, another terminal 100, and any server on a mobile communication network, which is constructed by a communication device that can execute a technical standard or communication method for mobile communication (for example, Long Term Evolution (LTE), Advanced Long Term Evolution (LTE-A), 5G New Radio (NR), and WI-FI) or a short-range communication method.

[0088] The sensor system 160 may include various sensors such as an image sensor 161 , a position sensor (IMU) 163 , an audio sensor 165 , a distance sensor, a proximity sensor, and a contact sensor.

[0089] Here, the image sensor 161 may capture an image (eg, a still or moving image and / or video) of a physical space around the terminal 100 .

[0090] Specifically, the image sensor 161 may capture a physical space through a camera disposed toward the outside of the terminal 100 .

[0091] In one embodiment, the image sensor 161 may be provided at the front and / or rear of the terminal 100 to capture a physical space in a direction in which the image sensor 161 is provided.

[0092] In an embodiment, the image sensor 161 may capture and obtain various images related to the service provided by the dedicated document sharing platform (eg, images of dedicated documents, etc.).

[0093] The image sensor 161 may include an image sensor device and an image processing module.

[0094] Specifically, the image sensor 161 may process a still image or a moving image obtained by an image sensor device (eg, CMOS or CCD).

[0095] In addition, the image sensor 161 may process a still image or a moving image obtained by the image sensor device using an image processing module to extract necessary information and transmit the extracted information to the processor component 120 .

[0096] The image sensor 161 may be a camera assembly including at least one camera.

[0097] Among them, the camera component may include an ordinary camera for shooting in the visible light band, and may also include special cameras such as infrared cameras, stereo cameras, etc.

[0098] In addition, the image sensor 161 may be included in the terminal 100 and operate according to an embodiment, or may be included in an external device (eg, an external server, etc.) and communicatively connected to the terminal 100 using the above-described communication processor 130 and / or interface 140 .

[0099] The position sensor (IMU) 163 may detect at least one of movement and acceleration of the terminal 100. For example, the position sensor 163 may be configured by a combination of various position sensors such as an accelerometer, a gyroscope, and / or a magnetometer.

[0100] In addition, the position sensor (IMU) 163 may recognize spatial information about a physical space around the terminal 100 in conjunction with the position communication processor 130 such as a GPS of the communication processor 130 .

[0101] The audio sensor 165 may sense or recognize sounds around the terminal 100 .

[0102] For example, the audio sensor 165 may include a microphone configured to detect an audio input of a user using the terminal 100 .

[0103] In an embodiment, the audio sensor 165 may receive audio data from the user that is required for the dedicated document sharing platform to provide services.

[0104] The interface 140 may connect the terminal 100 to one or more other devices so that the terminal 100 can communicate with the other devices.

[0105] In particular, interface 140 may include wired and / or wireless communication means that are compatible with one or more different communication protocols.

[0106] Through the interface 140 , the terminal 100 can be connected to various input and / or output devices.

[0107] For example, the interface 140 may be connected to an audio output device such as a headphone port or a speaker to output audio.

[0108] For example, the audio output device is described as being connected through the interface 140 , but in an embodiment, the audio output device may be included in the terminal 100 .

[0109] Additionally, for example, interface 140 may be connected to input devices such as a keyboard and / or a mouse to obtain user input.

[0110] The interface 140 may include at least one of a wired / wireless headset port, an external charger port, a wired / wireless data port, a memory card port, a port for connecting a device equipped with an identification module, an audio input / output (I / O) port, a video I / O port, a headphone port, a power amplifier, an RF circuit, a transceiver, and other communication circuits.

[0111] The input system 150 may detect user input (eg, gesture, voice command, button operation, touch, or any type of input) related to providing services on the dedicated document sharing platform.

[0112] Specifically, the input system 150 may include a button, a touch sensor, an image sensor 161 that detects a user motion input, and / or an audio sensor 165 that detects a user audio input.

[0113] Furthermore, the input system 150 may be connected to an external controller through the interface 140 to receive user input.

[0114] The display system 170 may output various types of information related to the service provided by the dedicated document sharing platform as a graphic image.

[0115] In an embodiment, the display system 170 may display various user interfaces, dedicated document data, question and answer content, and / or promotional content for the dedicated document sharing platform to provide services.

[0116] Such a display system may include, for example, but not limited to, at least one of a liquid crystal display (LCD), a thin film transistor-liquid crystal display (TFT LCD), an organic light emitting diode (OLED), a flexible display, a 3D display, and / or an electronic ink display.

[0117] In addition, according to an embodiment, the display system 170 may include a display 171 which outputs an image and a touch sensor 173 which detects a user's touch input.

[0118] For example, the display 171 may be implemented as a touch screen by forming a layer structure or an integral structure with the touch sensor 173 .

[0119] Such a touch screen may be used as a user input unit providing an input interface between the terminal 100 and the user, and may provide an output interface between the terminal 100 and the user.

[0120] The terminal 100 according to an embodiment of the present disclosure may perform deep learning related to the dedicated document sharing platform providing service based on a predetermined deep learning neural network.

[0121] Here, the deep learning neural network according to the embodiment may include, for example, but not limited to, OpenAI GPT, Instruction GPT, Bidirectional LSTM (Bi-LSTM), Long Short-Term Memory Model (LSTM), Multilayer Perceptron (MLP), EfficientNet, ResNet, Autoregressive Integrated Moving Average (ARIMA), Vector Autoregression (VAR), Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), Generative Adversarial Network (GAN), DualStyleGAN, StyleGAN, Graph Convolutional Network (GCN), Convolutional Neural Network (CNN), Deep Plane Sweep Network (DPSNet), Attention Guided Network (AGN), Region with CNN Features (R-CNN), Fast R-CNN, Faster R-CNN, Mask R-CNN and / or U-Net network.

[0122] Specifically, in one embodiment, the terminal 100 may perform deep learning for providing a service configuration for a dedicated document sharing platform by operating in conjunction with at least one deep learning neural network capable of implementing a question-answering language model according to an embodiment of the present disclosure.

[0123] Here, in one embodiment, the question-answering language model may perform deep learning by receiving predetermined dedicated document (e.g., papers and / or reports) data and outputting at least one question data and answer data set having an organic flow based on the input dedicated document data.

[0124] However, the present disclosure does not limit or constrain the deep learning algorithm itself that implements the question-answering language model, and the question-answering language model according to an embodiment of the present disclosure can be implemented based on at least one known deep learning algorithm.

[0125] On the other hand, according to an embodiment, the terminal 100 may also perform one or some of functional operations performed by the platform providing server 200 which will be described below.

[0126] - The platform provides servers 200

[0127] The platform providing server 200 according to an embodiment of the present disclosure may perform a series of processes for providing a dedicated document sharing platform providing service.

[0128] Specifically, in one embodiment, the platform providing server 200 may provide a dedicated document sharing platform providing service by exchanging data required to operate a dedicated document sharing platform providing process in an external device such as the terminal 100 with the external device.

[0129] For example, the platform providing server 200 may provide an environment in which the application 111 may operate on an external device (eg, the mobile computing device 100 - 1 and / or the desktop computing device 100 - 2 in the embodiments).

[0130] To this end, the platform providing server 200 may include an application program, data, and / or instructions for the operation of the application 111 , and may transmit or receive various types of data to or from an external device.

[0131] Furthermore, in an embodiment, the platform providing server 200 may obtain a predetermined dedicated document.

[0132] Furthermore, in an embodiment, the platform providing server 200 may generate question and answer content based on the acquired dedicated document.

[0133] Here, the question-and-answer content according to the embodiment may include content that visualizes at least one question data and an answer data set obtained by the question-and-answer language model according to the embodiment of the present disclosure in a predetermined manner.

[0134] Furthermore, in an embodiment, the platform providing server 200 may generate promotional content based on the generated question and answer content.

[0135] Here, the promotional content according to the embodiment may include online promotional materials for the purpose of promoting the dedicated document.

[0136] Furthermore, in an embodiment, the platform providing server 200 may provide a promotional material production workspace based on the generated promotional content.

[0137] Here, the promotion material production workspace according to an embodiment may include a user interface for determining promotion start content.

[0138] Here, the promotion start content according to the embodiment may include promotion content that is finally registered and shared on the dedicated document sharing platform.

[0139] Furthermore, in an embodiment, the platform providing server 200 may determine promotion start content based on the provided promotion material production workspace.

[0140] In addition, in an embodiment, the platform providing server 200 may provide the determined promotion start content.

[0141] Furthermore, in an embodiment, the platform providing server 200 may perform a dedicated document search function based on a dedicated document sharing platform.

[0142] That is, the platform providing server 200 may perform a dedicated document search function for searching for at least one dedicated document registered in the dedicated document sharing platform.

[0143] Furthermore, in an embodiment, the platform providing server 200 may perform dedicated document sharing and citing functions based on a dedicated document sharing platform.

[0144] Specifically, the platform providing server 200 may perform a dedicated document sharing and referencing function for sharing and referring to promotion start content, question data, and / or answer data related to a predetermined dedicated document.

[0145] Furthermore, in an embodiment, the platform providing server 200 may perform a reader question recommendation function based on a dedicated document sharing platform.

[0146] In other words, the platform providing server 200 may perform a reader question recommendation function for generating and providing suggested question data, which is question data written by a side receiving a predetermined dedicated document (eg, a reader side in the embodiment).

[0147] Furthermore, in an embodiment, the platform providing server 200 may perform deep learning for providing a service configuration for a dedicated document sharing platform based on a predetermined deep learning neural network.

[0148] Specifically, in one embodiment, the platform providing server 200 can perform deep learning for providing service configuration for a dedicated document sharing platform by operating in conjunction with at least one deep learning neural network capable of implementing or providing a question-answering language model according to an embodiment of the present disclosure.

[0149] More specifically, in an embodiment, the platform providing server 200 may read a predetermined deep learning neural network driver configured to perform deep learning from the memory module 230 .

[0150] Then, the platform providing server 200 may perform deep learning for providing service configuration for the dedicated document sharing platform according to the read predetermined deep learning neural network system.

[0151] Here, the deep learning neural network according to the embodiment may include, for example, but not limited to, OpenAI GPT, Instruction GPT, Bidirectional LSTM (Bi-LSTM), Long Short-Term Memory Model (LSTM), Multilayer Perceptron (MLP), EfficientNet, ResNet, Autoregressive Integrated Moving Average (ARIMA), Vector Autoregression (VAR), Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), Generative Adversarial Network (GAN), DualStyleGAN, StyleGAN, Graph Convolutional Network (GCN), Convolutional Neural Network (CNN), Deep Plane Sweep Network (DPSNet), Attention Guided Network (AGN), Region with CNN Features (R-CNN), Fast R-CNN, Faster R-CNN, Mask R-CNN and / or U-Net network.

[0152] Here, according to an embodiment, the deep learning neural network may be included in the platform providing server 200 , or may be implemented as another device and / or server separate from the platform providing server 200 .

[0153] In the following description, the deep learning neural network is described as a device included and implemented in the platform providing server 200, but the present disclosure is not limited thereto.

[0154] Furthermore, in an embodiment, the platform providing server 200 may store and manage various applications, instructions and / or data for implementing the dedicated document sharing platform provision service.

[0155] In an implementation, the platform provides that the server 200 may store and manage at least one of dedicated document data, question-and-answer content, promotional content, data processing algorithms, deep learning algorithms, and / or user interfaces.

[0156] However, in an embodiment of the present disclosure, the functional operations that can be performed by the platform providing server 200 are not limited to the above-mentioned functional operations, and the platform providing server 200 may be configured to further perform other functional operations.

[0157] Reference Figure 1 In an embodiment, the platform providing server 200 may be implemented as a computing device or computer including at least one processor module 210 for data processing, at least one communication module or communicator 220 for exchanging data with an external device, and at least one memory module 230 for storing various applications, data and / or instructions for providing a dedicated document sharing platform provisioning service.

[0158] Here, the memory module 230 may store any one or more of an operating system (OS), various application programs, data, and instructions for providing a dedicated document sharing platform provision service.

[0159] Furthermore, the memory module 230 may include a program area and a data area.

[0160] Here, the program area according to the embodiment may be linked between an operating system (OS) of a boot server and a functional element.

[0161] In addition, the data area according to the embodiment may store data generated when the server 200 is used or a program is executed.

[0162] Furthermore, the memory module 230 may be various storage devices such as ROM, RAM, EPROM, flash drive, and hard disk drive, and may also be a web storage device that performs a storage function of the memory module 230 on the Internet.

[0163] In addition, the memory module 230 may be a removable storage medium included in the server 200 .

[0164] The processor module 210 may control the overall operation of each unit of the above-mentioned server 200 so as to implement the dedicated document sharing platform provision service.

[0165] For example, the processor module 210 may be a system on chip (SOC) suitable for a server including a central processing unit (CPU) and / or a graphics processing unit (GPU).

[0166] In addition, the processor module 210 may execute an operating system (OS) and / or application programs stored in the memory module 230 .

[0167] Furthermore, the processor module 210 may control each component included in the server 200 .

[0168] In addition, the processor module 210 may communicate with each component internally through a system bus, and may include one or more predetermined bus structures including a local bus.

[0169] In addition, the processor module 210 can be implemented using at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor and / or other electrical units for performing functions.

[0170] In the above description, it has been described that the platform providing server 200 according to an embodiment of the present disclosure performs the above-mentioned functional operations. Alternatively, according to various embodiments, when necessary, one or some of the functional operations performed by the platform providing server 200 may be performed by an external device (e.g., the terminal 100), and / or one or some of the functional operations performed by the external device may be performed by the platform providing server 200.

[0171] - Question Answering Language Model (QALM)

[0172] Figure 3 is a conceptual diagram illustrating a question-answering language model (QALM) according to an embodiment of the present disclosure.

[0173] Reference Figure 3 According to an embodiment of the present disclosure, a question-answering language model QALM may include a deep learning model configured to generate and provide at least one question data and answer data set QA (hereinafter referred to as question-answering data) with an organic flow based on a predetermined dedicated document SD (e.g., a paper and / or a report).

[0174] For example, the question and answer language model QALM according to an embodiment may include a deep learning model configured to receive predetermined dedicated document data and output at least one question and answer data QA having an organic flow based on the input dedicated document data.

[0175] Here, the question and answer data QA according to the embodiment may have a structure connected according to a series of organic flows based on information included in the dedicated document SD.

[0176] That is, in an embodiment, the question data and the answer data may be a more cognitive form of question and answer data QA in addition to the typical question and answer data form of an extraction method that follows the standardized structure of a dedicated document SD (e.g., an introduction, research questions, research methods, research results and / or research conclusions according to the table of contents of a general paper).

[0177] Specifically, refer to Figure 3, the question-answering language model QALM in the implementation scheme may include a question deep learning model (QDM) and an answer deep learning model (ADM).

[0178] Specifically, the question deep learning model QDM according to an embodiment may be a deep learning model configured to receive predetermined dedicated document data and output at least one of question data Q1, Q2, Q3, ..., Qn based on the input dedicated document data.

[0179] In an embodiment, the question deep learning model QDM can generate 1) question data based on information included in the dedicated document SD, 2) question data based on previously generated question data, 3) question data based on previously generated answer data, and / or 4) question data based on other information related to the information included in the dedicated document SD.

[0180] Here, in one embodiment, the question deep learning model QDM may generate one or more question data as described above, such that the question data are connected according to an organic flow.

[0181] For example, the question deep learning model QDM can generate first question data based on a predetermined dedicated document, obtain first answer data as answer data for the generated first question data, and generate second question data based on the obtained first answer data and / or first question data.

[0182] Therefore, the question deep learning model QDM can provide deeper and higher quality question data than question data generated based solely on entities and / or relations between entities disclosed in the dedicated document SD.

[0183] The answer deep learning model ADM according to an embodiment may be a deep learning model configured to receive predetermined dedicated document data and question data, and output at least one of answer data A1, A2, A3, ..., An based on the input dedicated document data and question data.

[0184] Figure 4 and Figure 5 is a diagram showing a multi-step reasoning process of an answer deep learning model ADM according to an embodiment of the present disclosure.

[0185] Specifically, refer to Figure 4 and Figure 5 , according to the implementation scheme, the answer deep learning model ADM can generate answer data according to a multi-step reasoning process.

[0186] Here, the multi-step reasoning process according to the embodiment may have a structured process for generating answer data based on predetermined dedicated document data and question data using deep learning.

[0187] The multi-step inference process can include a relevance selection process, a rationale generation process, and a system synthesis process.

[0188] More specifically, the answer deep learning model ADM according to the implementation scheme can 1) determine the paragraphs in the dedicated document SD linked to the question data (hereinafter referred to as evidence paragraphs) (association selection process).

[0189] Specifically, the answer deep learning model ADM can extract at least one evidence paragraph containing the answer and / or rationale of the question data from K (K≥1) paragraphs included in the dedicated document SD.

[0190] Therefore, compared to the conventional method of extracting short answers to predetermined question data, the answer deep learning model ADM can generate answer data based on more extended rationale data.

[0191] Furthermore, in one embodiment, the answer deep learning model ADM may 2) obtain rationale data of the question data based on the determined evidence paragraphs (rationale generation process).

[0192] Here, the rationale data according to the embodiment may include data used as a basis for generating answer data of the question data.

[0193] For example, in an embodiment, the rationale data may be a collection of various types of data used in generating answer data.

[0194] Specifically, in one embodiment, the answer deep learning model ADM can detect main answer data, which is data including a direct answer to question data based on at least one paragraph of evidence, explanatory sentence data (data that explains the direct answer in detail) and / or auxiliary information data (data including relevant background knowledge).

[0195] The answer deep learning model ADM can obtain the basic principle data as described above based on the detected data.

[0196] Furthermore, in one embodiment, the answer deep learning model ADM may 3) generate answer data based on the acquired rationale data (system synthesis process).

[0197] Specifically, in this embodiment, the answer deep learning model ADM can generate answer data by performing data processing based on the acquired basic principle data.

[0198] In one embodiment, the answer deep learning model ADM may generate answer data by performing predetermined data processing (e.g., removing redundant text) to improve the conciseness and readability of the answer based on the evidence data.

[0199] As described above, the answer deep learning model ADM in one embodiment can extract evidence paragraphs that can answer question data from a dedicated document SD, generate evidence rationales based on the content of the extracted evidence paragraphs, and provide answer data that provides a deeper answer based on the generated evidence rationales.

[0200] That is, the answer deep learning model ADM in one embodiment can be implemented as a language model specifically for dedicated documents SD, which solves the problems of hallucination and ambiguous answers, which are limitations of existing general language models (e.g., OpenAI GPT, etc.) when performing question answering tasks for dedicated documents SD.

[0201] Therefore, the system 1000 for providing a dedicated document sharing platform according to some embodiments of the present disclosure can provide or generate fairer answer data to questions, illusion control, and evidence based on high-level reasoning / estimation functions implemented as processes similar to human cognitive reasoning.

[0202] The question and answer language model QALM including the above-mentioned question deep learning model QDM and the answer deep learning model ADM can generate at least one of the question and answer data QA((Q1, A1), (Q2, A2), ..., (Qn, An)) by matching at least one of the question data Q1, Q2, Q3, ..., Qn obtained via the question deep learning model QDM with at least one of the answer data A1, A2, A3, ..., An obtained via the answer deep learning model ADM based on each question data.

[0203] Here, in one embodiment, the question and answer language model QALM may generate at least one of the question and answer data QA listed sequentially according to the organic flow.

[0204] Furthermore, in one embodiment, the question answering language model QALM may provide or output at least one of the generated question answering data QA.

[0205] -Methods to provide a dedicated document sharing platform

[0206] Hereinafter, a method for generating question-and-answer content for a dedicated document SD using a predetermined deep learning neural network by an application 111 executed by at least one processor of a terminal 100 and providing promotional content based on the generated question-and-answer content (i.e., a method for implementing a dedicated document sharing platform to provide services) according to an embodiment of the present disclosure will be described in detail with reference to the accompanying drawings.

[0207] For the purpose of illustration only, the method for providing a dedicated document (SD) sharing platform will be described by being divided into a method for providing a dedicated document (SD) sharing platform on the author side and a method for providing a dedicated document (SD) sharing platform on the reader side. However, various embodiments can be implemented in a manner in which at least some embodiments including the above-mentioned providing method are organically combined and operated.

[0208] In one embodiment of the present disclosure, at least one processor of the terminal 100 may execute at least one application 111 stored in at least one memory 110 or enable at least one application 111 to run in a background state.

[0209] In the following embodiments, an operation in which at least one processor of the terminal 100 executes an instruction of the application 111 to perform a method for providing a dedicated document (SD) sharing platform will be described as the application 111 performing the method.

[0210] [Method for providing a dedicated document sharing platform on the author side]

[0211] Figure 6 Detailed description is a flowchart illustrating a method for providing a dedicated document (SD) sharing platform on an author side according to an embodiment of the present disclosure.

[0212] Reference Figure 6 In one embodiment, an application 111 executed by at least one processor of the terminal 100 or operating in a background state may obtain a predetermined dedicated document SD (S101).

[0213] Specifically, the application 111 may obtain a predetermined dedicated document SD (eg, a paper and / or a report) based on a dedicated document (SD) sharing platform according to an embodiment of the present disclosure.

[0214] For example, the application 111 may obtain a dedicated document SD uploaded on a dedicated document (SD) sharing platform through user input.

[0215] Furthermore, in an embodiment, the application 111 may generate question and answer content based on the obtained dedicated document SD ( S103 ).

[0216] Figure 7 An example of question and answer content according to an embodiment of the present disclosure is shown.

[0217] Here, the question and answer content QAC according to one embodiment may include content that visualizes at least one piece of question and answer data QA obtained by the question and answer language model QALM according to an embodiment of the present disclosure according to a predetermined method.

[0218] Specifically, in one embodiment, the application 111 may generate the question and answer content QAC in association with the question and answer language model QALM.

[0219] More specifically, in an embodiment, the application 111 may input the dedicated document (SD) data obtained as described above into the question answering language model QALM.

[0220] Then, in one embodiment, the question-answering language model QALM can obtain at least one of the question data Q1, Q2, Q3, ..., Qn based on the input dedicated document (SD) data on the basis of the question deep learning model QDM.

[0221] In addition, in one embodiment, the question-answering language model QALM can obtain at least one of the answer data A1, A2, A3, ..., An based on the input dedicated document (SD) data and question data on the basis of the answer deep learning model ADM.

[0222] Then, in one embodiment, the question and answer language model QALM can generate at least one of the question and answer data QA((Q1, A1), (Q2, A2), ..., (Qn, An)) by matching at least one of the obtained question data Q1, Q2, Q3, ..., Qn with at least one of the answer data A1, A2, A3, ..., An corresponding to each question data.

[0223] Furthermore, in one embodiment, the question and answer language model QALM may provide at least one of the generated question and answer data QA to the application 111 .

[0224] Therefore, in an embodiment, the application 111 may obtain at least one piece of question and answer data QA about the dedicated document SD from the question and answer language model QALM.

[0225] Furthermore, in one embodiment, the application 111 may generate question and answer content QAC based on at least one of the obtained question and answer data QA.

[0226] As an implementation, the application 111 may generate the question and answer content QAC by visualizing at least one piece of the question and answer data QA as a predetermined graphic image.

[0227] Here, according to an embodiment, the application 111 may generate question and answer content QAC including a predetermined evidence paragraph identification code.

[0228] The evidence paragraph identification code according to an embodiment may include an identification code corresponding to a predetermined evidence paragraph among identification codes (eg, a serial number in a preset format, etc.) assigned to paragraphs in the dedicated document SD.

[0229] Specifically, in one embodiment, the application 111 may detect at least one evidence paragraph corresponding to each answer data in the question and answer content QAC.

[0230] Furthermore, the application 111 may extract an identification code corresponding to the detected evidence paragraph (ie, an evidence paragraph identification code).

[0231] The application 111 may generate or display the question and answer content QAC having the extracted at least one evidence paragraph identification code by matching the extracted at least one evidence paragraph identification code with each answer data corresponding thereto.

[0232] Therefore, the application 111 can clearly present the paragraphs disclosing or showing the contents underlying each answer data piece and the positions of the paragraphs in the dedicated document SD so that they can be easily checked.

[0233] Furthermore, according to an embodiment, the application 111 may generate question and answer content QAC including predetermined downward characteristic data.

[0234] Here, the downward feature data according to an embodiment may include data obtained by reducing the amount of information of predetermined core data (eg, an image, a table, and / or a formula) in the dedicated document SD.

[0235] That is, in one embodiment, the downward feature data may include data obtained by converting characteristic feature values ​​contained in predetermined core data in the dedicated document SD according to a predetermined method to reduce the amount of the core data.

[0236] Specifically, in an embodiment, when the application 111 generates the question and answer content QAC based on the question and answer data QA obtained from the question and answer language model QALM, the application 111 may determine whether the predetermined answer data includes core data.

[0237] Here, if the application 111 determines that the core data exists in the predetermined answer data, the application 111 may generate downward feature data corresponding to the core data.

[0238] In an embodiment, the application 111 may generate downward feature data by performing data processing including data type conversion, data value removal, and / or data volume compression based on the core data.

[0239] For example, if the core data is in the form of a table, the application 111 may generate drop-down feature data that modifies the table into a simple bar chart.

[0240] Furthermore, in an embodiment, the application 111 may replace the core data corresponding to the downward feature data with the generated downward feature data.

[0241] That is, the application 111 may replace the core data in the question and answer data QA obtained from the question and answer language model QALM with the downward feature data.

[0242] In an embodiment, the application 111 may generate the question and answer content QAC based on the question and answer data QA in which the core data is replaced with the downward feature data.

[0243] In this way, the application 111 can reduce the information amount of data determined as core data of the dedicated document SD so that the dedicated document SD is exposed with a reduced amount of data.

[0244] Therefore, according to certain embodiments of the present disclosure, application 111 can prevent the core content in the dedicated document SD from being overly disclosed by promotional content provided later, and can also prevent the problem of reducing subscriptions and / or references to the dedicated document SD due to promotional content.

[0245] Here, according to an embodiment, the application 111 may convert the core data into downward feature data based on a user input.

[0246] In an embodiment, the application 111 may obtain a user input for setting a conversion degree, a conversion method, and / or conversion details regarding the core data.

[0247] Furthermore, the application 111 may generate downward feature data according to the obtained user input.

[0248] That is, the application 111 may generate question and answer content QAC based on downward feature data edited in a form optimized for user needs.

[0249] Furthermore, in an embodiment, the application 111 may generate promotional content based on the generated question and answer content QAC ( S105 ).

[0250] Figure 8 An example of promotional content according to an embodiment of the present disclosure is shown.

[0251] Here, the promotion content PMC according to the embodiment may include online promotion materials for promoting a predetermined dedicated document SD.

[0252] In one embodiment, the promotional content PMC may be content data configured according to a preset data volume (eg, one page).

[0253] For example, such promotional content PMC may include summary data SMD and / or question and answer content QAC regarding a predetermined dedicated document SD.

[0254] Specifically, in this embodiment, the application 111 can obtain summary data SMD about a dedicated document SD associated with a predetermined deep learning neural network.

[0255] Here, the specific method by which the application 111 according to the embodiment of the present disclosure can obtain the summary data SMD can be implemented based on various disclosed algorithms and / or processes, and the present disclosure is not limited to or restricted to these algorithms and / or processes.

[0256] Furthermore, in an embodiment, the application 111 may generate promotional content PMC including the acquired summary data SMD and / or the above-mentioned question and answer content QAC.

[0257] In an implementation, the application 111 may provide a promotional material production workspace based on the generated promotional content PMC ( S107 ).

[0258] Here, the promotion material preparation workspace according to an embodiment may include a user interface for determining promotion start content.

[0259] The promotion start content according to the embodiment may include promotion content PMC that is finally registered and shared on the dedicated document sharing platform.

[0260] Specifically, the application 111 according to an embodiment may provide a promotional material production workspace for displaying the promotional content PMC generated as described above and selecting and / or editing the displayed promotional content PMC according to a user input to determine promotion start content.

[0261] Here, the application 111 according to an embodiment may provide a promotion material preparation workspace configured to determine promotion start content based on inputs of a plurality of users.

[0262] Therefore, when there are multiple authors of a dedicated document SD, the application 111 can support a collaborative process in which the multiple authors can edit and / or select a promotional content PMC together.

[0263] In addition, in an embodiment, the application 111 may determine promotion start content based on the provided promotion material production workspace ( S109 ).

[0264] In other words, the promotion start content according to the embodiment may include promotion content PMC which is finally registered and shared on the dedicated document sharing platform according to the user input based on the promotion material production workspace.

[0265] For example, in an embodiment, the application 111 may determine the promotional content PMC as the promotional start content when obtaining a user input for selecting (eg, approving) the promotional content PMC based on the provided promotional material production workspace.

[0266] Meanwhile, in an embodiment, when obtaining user input for editing promotional content PMC based on the provided promotional material production workspace, the application 111 may generate promotional edited content, which is promotional content PMC edited according to the obtained user input.

[0267] Specifically, in one embodiment, the application 111 may 1) obtain user input for editing the composition of the question and answer content QAC in the promotional content PMC.

[0268] In an embodiment, the application 111 may obtain a user input for selecting, deleting and / or sorting at least one question and answer data QA (ie, question data and answer data set) in the question and answer content QAC.

[0269] Furthermore, the application 111 may obtain a user input for adding at least one question and answer data QA to the question and answer content QAC.

[0270] Fig. 9 and Fig.10 is an exemplary diagram illustrating a method of adding question and answer data QA based on user input according to an embodiment of the present disclosure.

[0271] Reference Fig. 9 In an implementation, the application 111 may generate at least one creative question and answer data CD based on user input.

[0272] Here, the creative question and answer data CD according to an embodiment may include a data set including question data directly created by a user and answer data with respect to the question data.

[0273] More specifically, in one embodiment, the application 111 may obtain question data (hereinafter referred to as creative question data CQD) directly created by the user according to user input.

[0274] Furthermore, in an embodiment, the application 111 may obtain answer data (hereinafter, additional answer data) for the creative question data CQD in association with the question answering language model QALM.

[0275] Specifically, in one embodiment, the application 111 may input the creative question data CQD into the question-answering language model QALM.

[0276] Then, in one embodiment, the question-answering language model QALM can obtain answer data based on the answer deep learning model ADM, based on the input dedicated document data and the creative question data CQD as described above.

[0277] Furthermore, the question-answering language model QALM may provide the obtained answer data (ie, additional answer data) to the application 111 .

[0278] Therefore, in an embodiment, the application 111 may obtain additional answer data from the question-answering language model QALM.

[0279] In an embodiment, the application 111 may generate creative question and answer data CD including the acquired additional answer data and creative question data CQD corresponding to the additional answer data.

[0280] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated creative question and answer data CD to the question and answer content QAC.

[0281] According to an embodiment, the application 111 may obtain answer data (hereinafter referred to as creative answer data) directly created by the user according to user input.

[0282] The application 111 may generate creative question and answer data CD including the acquired creative answer data and creative question data CQD corresponding to the creative answer data.

[0283] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated creative question and answer data CD to the question and answer content QAC.

[0284] As described above, according to certain embodiments of the present disclosure, the application 111 may provide question-and-answer content QAC created with a higher degree of freedom and cognitive elements by utilizing questions directly created by users and their answers.

[0285] Reference Fig.10 , the application 111 may generate at least one suggested question and answer data SD based on the user input.

[0286] Here, the suggested question and answer data SD according to an embodiment may include a data set including additionally provided question data and answer data of the question data based on the question and answer language model QALM.

[0287] More specifically, in an embodiment, the application 111 may obtain additionally provided question data (hereinafter referred to as suggested question data (SQD)) associated with the question answering language model QALM.

[0288] In an embodiment, application 111 can obtain 1) suggested question data SQD based on information included in the dedicated document SD, 2) suggested question data SQD based on previously generated question data, 3) suggested question data SQD based on previously generated answer data, and / or 4) suggested question data SQD based on other information related to the information included in the dedicated document SD.

[0289] Furthermore, in one embodiment, the application 111 may obtain answer data (hereinafter referred to as suggested answer data) of the suggested question data SQD associated with the question-answering language model QALM.

[0290] Specifically, in an embodiment, the application 111 may input the suggested question data SQD into the question-answering language model QALM.

[0291] Then, in one embodiment, the question-answering language model QALM may obtain answer data based on the answer deep learning model ADM, based on the dedicated document data and the suggested question data SQD input as described above.

[0292] Furthermore, the question answering language model QALM may provide the acquired answer data (ie, suggested answer data) to the application 111 .

[0293] Then, in one embodiment, the application 111 may obtain suggested answer data from the question-answering language model QALM.

[0294] In an embodiment, the application 111 may generate suggested question and answer data SD, which includes the acquired suggested answer data and suggested question data SQD corresponding to the suggested answer data.

[0295] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated suggested question and answer data SD to the question and answer content QAC.

[0296] According to an embodiment, the application 111 may obtain answer data (eg, creative answer data) directly created by the user according to user input.

[0297] The application 111 may generate suggested question and answer data SD including the acquired creative answer data and suggested question data SQD corresponding to the creative answer data.

[0298] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated suggested question and answer data SD to the question and answer content QAC.

[0299] As described above, the application 111 can generate question and answer content QAC based on at least one question and answer data QA additionally suggested by the question and answer language model QALM, thereby minimizing the difficulty of creating questions for improving the dedicated document SD and providing question and answer content QAC based on more extensive various options.

[0300] Reference Fig.10 In an embodiment, the application 111 may generate at least one suggested question and answer data PD based on an input of another user.

[0301] Here, the suggested question and answer data PD according to the embodiment may include a data set including question data directly created by another user (eg, a reader in the embodiment, etc.) and answer data for the question data.

[0302] Specifically, in one embodiment, the application 111 may obtain question data (hereinafter referred to as suggested question data (PQD)) directly created by another user in association with the terminal 100 of another user.

[0303] Furthermore, in one embodiment, the application 111 may obtain answer data (hereinafter referred to as suggested answer data) for the suggested question data PQD in association with the question answering language model QALM.

[0304] Specifically, in an embodiment, the application 111 may input the suggested question data PQD into the question-answering language model QALM.

[0305] Then, in an embodiment, the question-answering language model QALM may obtain answer data based on the answer deep learning model ADM, based on the dedicated document data and the suggested question data PQD input as described above.

[0306] Furthermore, the question answering language model QALM may provide the obtained answer data (ie, suggested answer data) to the application 111 .

[0307] Therefore, in an embodiment, the application 111 may obtain suggested answer data from the question-answering language model QALM.

[0308] In an embodiment, the application 111 may generate suggested question and answer data PD including the acquired suggested answer data and suggested question data PQD corresponding to the suggested answer data.

[0309] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated suggested question and answer data PD to the question and answer content QAC.

[0310] According to an embodiment, the application 111 may obtain answer data (eg, creative answer data) directly created by the user according to user input.

[0311] The application 111 may generate suggested question and answer data PD including the acquired creative answer data and suggested question data PQD corresponding to the creative answer data.

[0312] Furthermore, in an embodiment, the application 111 may obtain a user input for adding the generated suggested question and answer data PD to the question and answer content QAC.

[0313] In this way, according to some embodiments of the present disclosure, application 111 can generate question and answer content QAC, which includes question data and answer data generated by other users, including reader users who subscribe to dedicated documents SD instead of author users who create dedicated documents SD.

[0314] Therefore, the application 111 can actively accept questions that the demander of the dedicated document may be very curious about and interested in, and provide promotional content PMC including question-and-answer content QAC reflecting these questions.

[0315] Therefore, according to some embodiments of the present disclosure, the application 111 may obtain user input for adding the above-mentioned creative question and answer data CD, suggested question and answer data SD and / or suggested question and answer data PD to the question and answer content QAC.

[0316] In one embodiment, application 111 can edit the composition of the question and answer content QAC in the promoted content PMC based on the user input obtained as described above (for example, user input for selecting at least one question and answer data QA in the question and answer content QAC, user input for deleting the question and answer content QAC, user input for ordering the question and answer content QAC, and / or user input for adding the question and answer content QAC).

[0317] Meanwhile, in one embodiment, the application 111 may obtain 2) user input for editing the content of the question and answer content QAC in the promotional content PMC.

[0318] In an embodiment, the application 111 may obtain user input for modifying (eg, correcting, deleting, and / or adding) the content of predetermined question data and / or answer data in the question and answer content QAC.

[0319] In an implementation, the application 111 may edit the content of the question-and-answer content QAC in the promotional content PMC based on the obtained user input.

[0320] According to an embodiment, the application 111 may provide an automatically modified related content update function.

[0321] Here, the automatic modification related content update function according to the embodiment may include a function of automatically detecting and modifying content (hereinafter referred to as related data) related to modification content (hereinafter referred to as modification data).

[0322] Specifically, in one embodiment, the application 111 may detect at least one data related to the predetermined modification data in the question and answer content QAC.

[0323] In one embodiment, the application 111 may detect at least one piece of related data including information (eg, text, image, table, and / or formula) having a predetermined level or more of similarity to information included in the modification data.

[0324] In an embodiment, the application 111 may collectively correct the detected related data by applying the modification history applied to the modified data.

[0325] For example, if the modification history applied to the first modification data in the dedicated document SD is changing the "main contribution" to the "key contribution", the application 111 may change all the "main contributions" existing in the dedicated document SD to "key contributions".

[0326] Therefore, in an embodiment, the application 111 may provide the following function (ie, the automatic modification related content update function): when a user modifies specific content in the question and answer content QAC, the content related to the specific content is automatically corrected according to the modification.

[0327] Therefore, according to certain embodiments of the present disclosure, the application 111 can improve the user convenience and usability of editing the question and answer content QAC, thereby improving user satisfaction.

[0328] Fig.11 is a diagram illustrating a method of providing an editing assistance tool and / or a related document search function according to an embodiment of the present disclosure.

[0329] Reference Fig.11 In an embodiment, the application 111 may provide an editing assistance tool (EAT).

[0330] Here, the editing assistance tool according to an embodiment may include a user interface configured to provide a predetermined image, table, and / or formula related to a predetermined dedicated document in a form that is easily inserted into the question and answer content QAC.

[0331] For example, editing assistance tools may include a Markdown editor, etc.

[0332] In an embodiment, the application 111 may provide an editing assistance tool (EAT) that provides predetermined images, tables, and / or formulas included in the dedicated document SD in a form that can be used for the question and answer content QAC.

[0333] That is, in this embodiment, the application 111 allows the user to easily and conveniently utilize various images, tables and / or formulas related to the dedicated document SD by using the editing assistance tool EAT when editing the question and answer content QAC.

[0334] Therefore, the application 111 may provide a question and answer content QAC editing process that may easily utilize richer data.

[0335] In addition, refer to Fig.11 In an implementation, the application 111 may provide a related document search function.

[0336] Here, the related document search function according to an embodiment may include a function of automatically detecting and providing a dedicated document SD (hereinafter referred to as a related document (RD)) related to predetermined question data and / or answer data.

[0337] Specifically, in an embodiment, the application 111 may obtain user input for selecting predetermined question data and / or answer data in the question and answer content QAC.

[0338] Furthermore, in one embodiment, the application 111 may detect at least one dedicated document SD related to question data and / or answer data (hereinafter, selected data) selected according to the obtained user input.

[0339] Here, the specific method for the application 111 to detect the related document RD according to the embodiment of the present disclosure may be implemented based on various disclosed algorithms and / or processes, and the present disclosure is not limited to or restricted to these algorithms and / or processes.

[0340] Furthermore, in an embodiment, the application 111 may provide at least one detected related document RD by matching the at least one detected related document RD with the selected data.

[0341] Therefore, when editing the question and answer content QAC, the application 111 can enable the user to easily refer to and utilize the dedicated document SD related to the question data and / or the answer data without performing a separate search.

[0342] Therefore, the application 111 can support users so that they can easily compare their dedicated documents SD with other dedicated documents SD and easily highlight the differences between them based on the additional related document information provided as described above.

[0343] As described above, in one embodiment, the application 111 configured to obtain user input for editing works and / or content of question and answer content QAC in the promotional content PMC may generate promotional editorial content according to the obtained editing input.

[0344] In addition, in an embodiment, the application 111 may determine the promotional editorial content as the promotional start content when obtaining a user input for selecting (ie, approving) the promotional editorial content based on the promotional material production workspace.

[0345] Thus, application 111 may allow users to create customized promotion start content optimized for forms in which users may have a high degree of freedom.

[0346] According to an embodiment, the application 111 may provide a question and answer content volume adjustment function.

[0347] Here, the question and answer content amount adjustment function according to an embodiment may include a function of compressing or increasing the question and answer content QAC in the determined promotion start content according to a predetermined method when the determined promotion start content does not meet a preset data amount condition (e.g., one page).

[0348] Specifically, in one embodiment, when the promotion start content exceeds a preset data amount (for example, when the amount of the promotion start content exceeds one page), the application 111 may reduce the amount of at least one answer data in the promotion start content.

[0349] In one embodiment, the application 111 may obtain the thumbnail data of at least one answer data in the promotion start content in association with a predetermined deep learning neural network.

[0350] Here, the thumbnail data according to an embodiment may include data obtained by reducing the amount of predetermined answer data according to a predetermined method (eg, aggregation).

[0351] Here, the specific method for the application 111 to obtain thumbnail data according to the embodiment of the present disclosure can be implemented on the basis of various disclosed algorithms and / or processes, and the present disclosure is not limited to or restricted to these algorithms and / or processes.

[0352] In one embodiment, the application 111 may obtain the abbreviated data of each answer data in the promotion start content based on the user input.

[0353] That is, according to an embodiment, the application 111 may obtain thumbnail data according to content directly input by the user.

[0354] Furthermore, in an embodiment, the application 111 may replace the answer data corresponding to the obtained thumbnail data with the obtained thumbnail data.

[0355] That is, the application 111 can replace the answer data in the promotion start content with the thumbnail data.

[0356] In an embodiment, upon acquiring a user input (eg, a selection input) for the thumbnail data in the promotion start content, the application 111 may provide answer data (hereinafter, original data) corresponding to the thumbnail data.

[0357] On the other hand, if the promotion start content is less than a preset data amount (eg, when the amount of the promotion start content is equal to or less than one page), the application 111 may additionally provide at least one piece of suggested question and answer data SD as described above in the embodiments.

[0358] Then, the application 111 may further include suggested question and answer data SD (hereinafter, added question and answer data) selected from at least one additionally provided suggested question and answer data SD in the question and answer content QAC in the promotion start content according to a user input.

[0359] Here, in one embodiment, the application 111 may determine whether the added question and answer data satisfies a preset data amount condition when the added question and answer data is included in the question and answer content QAC.

[0360] If the preset data volume condition is met, the application 111 may insert the added question and answer data into the question and answer content QAC within the promotion start content.

[0361] On the other hand, if the preset data volume condition is not met, the application 111 may repeatedly execute the above-mentioned question and answer content volume adjustment function.

[0362] In this way, according to some embodiments of the present disclosure, the application 111 may provide a function of easily processing the determined promotion start content into a form that satisfies a preset quantity condition.

[0363] In addition, in an embodiment, the application 111 may provide the determined promotion start content ( S111 ).

[0364] Specifically, in one implementation, the application 111 may provide promotion start content based on a dedicated document sharing platform.

[0365] For example, in one embodiment, the application 111 may register the promotion start content on the dedicated document sharing platform to share the promotion start content determined as described above and provide it to users (eg, authors and / or readers in the embodiment) using the dedicated document sharing platform.

[0366] As described above, in one embodiment, application 111 can use the question and answer language model QALM according to the embodiment of the present disclosure to generate question and answer content QAC for a predetermined dedicated document SD, provide an interface for freely editing the generated question and answer content QAC, and provide promotional content PMC based on the generated question and answer content QAC.

[0367] Therefore, application 111 can easily provide high-quality promotional materials optimized for user needs, while including deep and practical question-and-answer content obtained from a deep learning model specifically designed for dedicated documents SD.

[0368] [Method for providing a dedicated document sharing platform on the reader side]

[0369] Fig.12 is a flowchart illustrating a method for providing a dedicated document sharing platform on a reader side according to an embodiment of the present disclosure.

[0370] Reference Fig.12 , the application 111 may provide a dedicated document search function based on a dedicated document sharing platform (S201).

[0371] Specifically, in one embodiment, the application 111 may provide a dedicated document search function for searching for at least one dedicated document SD registered in the dedicated document sharing platform (hereinafter referred to as a registered document).

[0372] For example, in an embodiment, the application 111 may provide a dedicated document search function based on promotion start content corresponding to each registered document.

[0373] Specifically, in one implementation, the application 111 may obtain at least one search keyword based on user input.

[0374] In addition, the application 111 may detect at least one promotion start content including the obtained search keyword.

[0375] That is, the application 111 may detect at least one promotion start content based on the summary data SMD and / or the question and answer content QAC including the predetermined search keyword.

[0376] In addition, the application 111 may provide at least one detected promotion start content through a dedicated document sharing platform.

[0377] In other words, the application 111 may provide at least one promotion start content corresponding to the search keyword as a search result according to the user input.

[0378] In this manner, the application 111 can improve the search speed and / or performance by performing a dedicated document search process using content (eg, promotion start content) that effectively and specifically suggests and expresses the content of the dedicated document SD.

[0379] Furthermore, the application 111 may provide a dedicated document search service by providing search results in the form of promotion start content to allow the user to more easily and efficiently identify a dedicated document SD related to content desired by the user.

[0380] Furthermore, in an embodiment, the application 111 may provide a dedicated document sharing and citing function based on a dedicated document sharing platform ( S203 ).

[0381] Fig.13is an exemplary diagram for illustrating a method for providing dedicated document sharing and citing functions according to an embodiment of the present disclosure.

[0382] Reference Fig.13 , the application 111 may provide a dedicated document sharing and referencing function for sharing and referring to promotion start content corresponding to a predetermined registered document.

[0383] For example, in an embodiment, the application 111 may allocate an identification code corresponding to each promotion start content (hereinafter referred to as a promotional material identification code (PIC)).

[0384] Here, the PIC according to an embodiment may include a uniform resource locator (URL) through which predetermined promotion start content can be accessed and / or a timestamp specifying a reference time.

[0385] Furthermore, the application 111 according to the embodiment may provide a sharing and referencing function for a predetermined promotion start content based on the PIC as described above.

[0386] In an embodiment, the application 111 may obtain user input for requesting to share and / or reference a predetermined promotion start content.

[0387] Then, the application 111 may detect a PIC that matches the corresponding promotion start content.

[0388] The application 111 may output and provide the detected PIC in a predetermined manner (eg, in the form of a draft post that can be published on a specific service).

[0389] Furthermore, in one embodiment, the application 111 may provide a dedicated document sharing and referencing function for sharing and referring to the question and answer data QA included in the question and answer content QAC in the predetermined promotion start content.

[0390] Specifically, in one embodiment, the application 111 may allocate an identification code corresponding to each question and answer data QA (hereinafter referred to as a question and answer identification code (DIC)).

[0391] Here, the DIC according to the embodiment may include a URL through which promotion start content including corresponding question and answer data QA can be accessed and / or a timestamp specifying a reference time.

[0392] Furthermore, in an embodiment, the application 111 may provide sharing and reference functions for the predetermined question and answer data QA based on the DIC as described above.

[0393] In an embodiment, the application 111 may obtain a user input for requesting to share and / or refer to the predetermined question and answer data QA.

[0394] Then, the application 111 may detect the DIC matching the question and answer data QA.

[0395] Then, the application 111 may output and provide the detected DIC in a predetermined manner (eg, in the form of a draft post that can be published on a specific service).

[0396] Here, in an embodiment, the application 111 may provide a symbol format (eg, a reference symbol) suggested when the PIC and / or the DIC is referenced as a reference in a pop-up window.

[0397] For example, the application 111 may provide reference symbols based on the APA version, the MLA version, or the Bibtex version as a pop-up window.

[0398] In this way, according to some embodiments of the present disclosure, application 111 can provide a technical configuration that can share and quote promotion start content for a dedicated document SD registered on a dedicated document sharing platform, or share and quote question and answer data QA in the question and answer content QAC included in the promotion start content.

[0399] Therefore, the application 111 can increase the accessibility to registered documents on the dedicated document sharing platform, and can also further enhance the impact of author insights related to the registered documents.

[0400] According to an embodiment, when certain promotion start content and / or question and answer data QA is cited based on PIC and / or DIC, the application 111 may provide information (eg, citation credit) for notifying the author who wrote the relevant registration document of the citation.

[0401] Therefore, the application 111 can enable the author to ascertain the status of sharing and reference of his / her private document SD in real time.

[0402] Furthermore, in an embodiment, the application 111 may provide a reader question recommendation function based on a dedicated document sharing platform ( S205 ).

[0403] For example, in an embodiment, the application 111 may provide a reader question recommendation function for obtaining and providing suggested question data PQD for a predetermined registered document.

[0404] More specifically, in one embodiment, the application 111 may provide a user interface (hereinafter referred to as a suggested question input interface), through which the suggested question data PQD as described above may be obtained and provided based on a dedicated document sharing platform.

[0405] In other words, the application 111 may provide a suggested question input interface through which suggested question data PQD may be obtained and provided, the suggested question data PQD being question data directly created by a user other than the author user who wrote the registered document (eg, a reader in an embodiment).

[0406] Furthermore, the application 111 may obtain suggested question data PQD for a predetermined registered document based on the input of the user (reader in the embodiment) based on the provided suggested question input interface.

[0407] Furthermore, the application 111 may provide the obtained suggested question data PQD to the terminal 100 of the author who wrote the relevant registered document.

[0408] Therefore, the application 111 can generate question and answer content QAC based on the above-mentioned suggested question and answer data PD.

[0409] Therefore, according to some embodiments of the present disclosure, the application 111 can implement promotion start content, which proactively reflects the issues that readers are curious about and interested in regarding a specific registration document.

[0410] Here, the application 111 may set the disclosure scope of the suggested question data PQD based on input of a user of the suggested question input interface (reader in the embodiment) and / or input of a user of the promotional material creation workspace in the embodiment (author in the embodiment).

[0411] In an embodiment, application 111 can set the disclosure scope based on user input for setting "full disclosure" or "author disclosure", where "full disclosure" discloses the suggested question data PQD to all users using the dedicated document disclosure platform, and "author disclosure" discloses the suggested question data PQD only to the author who wrote the relevant dedicated document SD.

[0412] Furthermore, in one embodiment, the application 111 may provide suggested question data PQD to all users or authors according to a set disclosure scope.

[0413] Therefore, the application 111 can more effectively protect the intellectual property rights of readers and / or authors by selectively determining the disclosure scope of question data suggested by readers.

[0414] As described above, according to certain embodiments of the present disclosure, the method and system for providing a dedicated document sharing platform can use a predetermined deep learning neural network to generate question and answer content QAC for a dedicated document SD, and provide promotional content PMC based on the generated question and answer content QAC, thereby enabling users to automatically obtain data in a question and answer format that efficiently includes the content of the dedicated document SD without separate efforts, and utilize the promotional content PMC based on the obtained data.

[0415] According to some embodiments of the present disclosure, the method and system for providing a dedicated document sharing platform can use a language model dedicated to a dedicated document SD to generate question-answering content QAC, thereby providing question-answering content QAC that solves hallucination and ambiguous answer problems, which are limitations of existing general language models (e.g., OpenAI GPT, etc.).

[0416] Therefore, the method and system for providing a dedicated document sharing platform according to certain embodiments of the present disclosure can provide question-answering content QAC that is reasonable for questions, illusion control, and evidence based on advanced inference / estimation functions implemented through a process similar to human cognitive inference.

[0417] In addition, according to certain embodiments of the present disclosure, the method and system for providing a dedicated document sharing platform can provide promotional content PMC based on a user interface UI, through which the generated question and answer content QAC can be freely edited, thereby supporting users to create customized promotional materials optimized for forms in which users want to have a higher degree of freedom beyond the level of simply sharing dedicated documents SD, and actively use the customized promotional materials to perform promotional activities.

[0418] Therefore, the methods and systems for providing a dedicated document sharing platform according to some embodiments of the present disclosure can easily generate and utilize high-quality promotional materials optimized for user needs, while including deep and practical question-and-answer content obtained from a deep learning model dedicated to dedicated documents SD.

[0419] The embodiments of the present disclosure described above can be implemented in the form of program instructions, which can be executed by various computer components and recorded on a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc., alone or in combination. The program instructions recorded on the computer-readable recording medium may be program instructions specially designed and configured for the present disclosure or program instructions known and available to technicians in the field of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floppy disks, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using a compiler, etc. Hardware devices can be changed into one or more software modules to perform processing according to the present disclosure, and vice versa.

[0420] Cross-references to related applications

[0421] This application claims priority to and the benefit of Korean Patent Application No. 10-2023-0152908, filed on November 7, 2023, which is hereby incorporated by reference in its entirety.

Claims

1. A computer-implemented method, the computer-implemented method comprising the following steps: Obtain dedicated documentation; generating question-answer content for the dedicated document using a question-answer language model; generating promotional content including online promotional materials for the dedicated document based on the question and answer content for the dedicated document; Providing a workspace for producing promotional materials based on the promotional content; determining promotion start content including promotion content to be registered on a dedicated document sharing platform using the promotion material production workspace; and The promotion start content is provided using the dedicated document sharing platform.

2. The computer-implemented method of claim 1 , wherein: The step of generating the question and answer content comprises the following steps: Inputting the dedicated document into the question-answering language model; Outputting, by the question-answering language model, at least one question-answering data including question data and answer data in response to the input first dedicated document; and The question and answer content is generated based on the at least one question and answer data.

3. The computer-implemented method of claim 2, wherein: The step of outputting the at least one question-answer data by the question-answer language model comprises the following steps: Outputting at least one question data using a question deep learning model included in the question-answering language model; and Use the answer deep learning model included in the question-answering language model to output at least one answer data corresponding to each of the at least one question data.

4. The computer-implemented method of claim 3, wherein: The answer deep learning model includes a deep learning model for outputting the answer data by performing a multi-step inference process, the multi-step inference process including the following steps: Identifying the paragraph of evidence in the dedicated document that is relevant to the data in question; Obtaining at least one piece of basic principle data related to the problem data based on the evidence paragraph; and Data processing is performed based on the basic principle data.

5. The computer-implemented method of claim 1 , wherein: The promotional content includes content data configured based on a predetermined data volume condition.

6. The computer-implemented method of claim 1 , wherein: The step of determining the promotion start content includes the step of generating promotion edited content using the promotion material production workspace, wherein the promotion edited content is promotion content edited according to user input.

7. The computer-implemented method of claim 6, wherein: The step of generating the promotional editorial content includes one or both of the following steps: generating the promotional editorial content based on user input for editing the composition of the question and answer content in the promotional content, or generating the promotional editorial content based on user input for editing the content of the question and answer content in the promotional content.

8. The computer-implemented method of claim 6, wherein: The step of generating the promotional editorial content includes the step of providing an editing assistance tool, which includes a user interface for providing one or more of predetermined images, tables or formulas related to the dedicated document in a form configured to be included in the question and answer content in the promotional content.

9. The computer-implemented method of claim 6, wherein: The step of generating the promoted editorial content includes the step of performing a related document search for automatically detecting at least one other specialized document related to the question-and-answer content in the promoted content.

10. The computer-implemented method of claim 6, wherein: The step of determining the promotion start content includes the step of: using the promotion material production workspace, selecting one of the promotion content or the promotion editorial content as the promotion start content based on user input.

11. The computer-implemented method of claim 1 , further comprising the steps of: A dedicated document search is performed based on the promoted start content using the dedicated document sharing platform.

12. The computer-implemented method of claim 1 , further comprising the steps of: Dedicated document sharing and quoting are performed based on the promotion start content using the dedicated document sharing platform.

13. The computer-implemented method of claim 12, wherein: The step of executing the dedicated document sharing and referencing includes one or both of the following steps: providing a promotional material identification code including an access URL and a timestamp associated with the promotion start content, or providing a question and answer identification code including an access URL and a timestamp associated with the question and answer content in the promotion start content, wherein the access URL is an access uniform resource locator.

14. The computer-implemented method of claim 1, further comprising the step of performing a reader question recommendation for providing question data created by a user who receives the dedicated document using the dedicated document sharing platform.

Images