User-assembled artificial intelligence service platform apparatus through artificial intelligence pipelining

Abstract

The present invention relates to a user-assembled artificial intelligence service platform apparatus through artificial intelligence pipelining, the apparatus comprising: an artificial intelligence block curation unit that curates a plurality of artificial intelligence blocks; an artificial intelligence block selection unit that selects at least first and second artificial intelligence blocks from among the plurality of artificial intelligence blocks; an artificial intelligence pipeline generation unit that generates artificial intelligence pipelining by connecting an output port in the first artificial intelligence block to an input port in the second artificial intelligence block; and a user-assembled artificial intelligence model generation unit that generates a user-assembled artificial intelligence model through the artificial intelligence pipelining.

Description

User-assembled AI service platform device through AI pipelining

[0001] The present invention relates to artificial intelligence service platform technology, and more specifically, to a user-assembly type artificial intelligence service platform device through artificial intelligence pipelining that can reduce the burden of artificial intelligence production by easily connecting neural networks between artificial intelligence models using no-code blocks to create artificial intelligence.

[0002] Artificial intelligence (AI) development technology encompasses the process of designing, developing, and training AI systems by combining various sub-technologies and disciplines. AI has largely evolved around machine learning and deep learning; however, it also utilizes diverse technologies such as data processing and preparation, computer vision, natural language processing (NLP), reinforcement learning, and AI model training and optimization, as well as AI tools provided by cloud platforms like AWS and Google Cloud. Through this, it is possible to create intelligent systems that learn from data and solve problems without human intervention. However, the current AI development industry market faces a shortage of skilled researchers and developers, as well as issues regarding development time and cost burdens.

[0003] Meanwhile, AI services provide APIs and services that allow for the easy utilization of pre-trained models. Examples include Google's TensorFlow, OpenAI's GPT API, and Microsoft's Azure AI. Model serving, which involves deploying trained models to actual service environments, is one of the final stages of AI development; by integrating trained models into web services or applications, they can respond to user requests in real time.

[0004] Therefore, there is a need for technology that can support the development of artificial intelligence more simply and efficiently based on such AI model serving.

[0005] Prior Art: Korean Registered Patent No. 10-2284539 (July 27, 2021)

[0006] One embodiment of the present invention aims to provide a user-assembly type artificial intelligence service platform device through artificial intelligence pipelining, which can reduce the burden of artificial intelligence production by easily connecting neural networks between artificial intelligence models using no-code blocks to create artificial intelligence.

[0007] One embodiment of the present invention aims to provide a user-assembled artificial intelligence service platform device through artificial intelligence pipelining that can lower the difficulty of use by serving a context-appropriate artificial intelligence model workflow using a natural language-based chatbot and create and distribute new artificial intelligence models by combining artificial intelligence models.

[0008] One embodiment of the present invention aims to provide a user-assembled artificial intelligence service platform device through artificial intelligence pipelining that can provide a community where users can easily design artificial intelligence workflows through dragging and share the final completed artificial intelligence model.

[0009] Among the embodiments, a user-assembled artificial intelligence service platform device through artificial intelligence pipelining comprises: an artificial intelligence block curation unit that curates a plurality of artificial intelligence blocks; an artificial intelligence block selection unit that selects at least first and second artificial intelligence blocks among the plurality of artificial intelligence blocks; an artificial intelligence pipeline generation unit that generates artificial intelligence pipelining by connecting an output port in the first artificial intelligence block to an input port in the second artificial intelligence block; and a user-assembled artificial intelligence model generation unit that generates a user-assembled artificial intelligence model through the artificial intelligence pipelining.

[0010] The above artificial intelligence block curation unit provides the plurality of artificial intelligence blocks for sale on Marketplace, and each of the plurality of artificial intelligence blocks may correspond to a user-created or user-ordered artificial intelligence model.

[0011] The above artificial intelligence block selection unit analyzes user-defined tasks to recommend a plurality of candidate artificial intelligence blocks composed of heterogeneous artificial intelligence services that perform different functions, and can select at least the first and second artificial intelligence blocks through user control.

[0012] The artificial intelligence pipeline generation unit can determine the connection suitability between the output port and the input port during the connection process and generate a data compatibility interface between the output port and the input port to perform data transfer between the first and second artificial intelligence blocks.

[0013] The artificial intelligence pipeline generation unit above can determine the first artificial intelligence block as a dedicated chatbot-based artificial intelligence head capable of receiving user input, or generate a public chatbot-based artificial intelligence head by connecting a public chatbot to the input port of the first artificial intelligence block.

[0014] The above user-assembled artificial intelligence model generation unit can verify whether there is an error in a specific artificial intelligence block constituting the above user-assembled artificial intelligence model by verifying a user input-based artificial intelligence service through the above user-assembled artificial intelligence model.

[0015] The disclosed technology may have the following effects. However, this does not mean that a specific embodiment must include all of the following effects or only the following effects; therefore, the scope of the rights of the disclosed technology should not be understood as being limited by this.

[0016] A user-assembled artificial intelligence service platform device using artificial intelligence pipelining according to one embodiment of the present invention can reduce the burden of artificial intelligence production by easily connecting neural networks between artificial intelligence models with no-code blocks to create artificial intelligence.

[0017] A user-assembled artificial intelligence service platform device through artificial intelligence pipelining according to one embodiment of the present invention can lower the difficulty of use by serving a situation-appropriate artificial intelligence model workflow using a natural language-based chatbot and can create and distribute new artificial intelligence models by combining artificial intelligence models.

[0018] A user-assembled artificial intelligence service platform device through artificial intelligence pipelining according to one embodiment of the present invention can provide a community where users can easily design an artificial intelligence workflow by dragging and share the final completed artificial intelligence model.

[0019] FIG. 1 is a drawing illustrating a user-assembled artificial intelligence service platform system according to the present invention.

[0020] Figure 2 is a diagram illustrating the system configuration of the artificial intelligence service platform device of Figure 1.

[0021] Figure 3 is a diagram illustrating the configuration of the artificial intelligence service platform device of Figure 1.

[0022] FIG. 4 is a flowchart illustrating the process of a user-assembled artificial intelligence service through artificial intelligence pipelining of an artificial intelligence service platform device according to the present invention.

[0023] FIG. 5 is a drawing illustrating an embodiment of a virtual reality experience scenario for a metal type production experience service device according to the present invention.

[0024] FIG. 6 is a diagram showing an example of user-assembled artificial intelligence model generation of an artificial intelligence service platform device according to the present invention.

[0025] The description of the present invention is merely an example for structural or functional explanation, and therefore the scope of the present invention should not be interpreted as being limited by the examples described in the text. That is, since the examples are subject to various modifications and may take various forms, the scope of the present invention should be understood to include equivalents capable of realizing the technical concept. Furthermore, the objectives or effects presented in the present invention do not imply that a specific example must include all of them or only such effects; therefore, the scope of the present invention should not be understood as being limited by them.

[0026] Meanwhile, the meaning of the terms described in this application should be understood as follows.

[0027] Terms such as "first," "second," etc., are intended to distinguish one component from another, and the scope of rights shall not be limited by these terms. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0028] When it is stated that one component is "connected" to another component, it should be understood that it may be directly connected to that other component, or that there may be other components in between. Conversely, when it is stated that one component is "directly connected" to another component, it should be understood that there are no other components in between. Meanwhile, other expressions describing the relationships between components, such as "between" and "exactly between," or "adjacent to" and "directly adjacent to," should be interpreted in the same way.

[0029] A singular expression should be understood to include a plural expression unless the context clearly indicates otherwise, and terms such as "include" or "have" are intended to specify the existence of the implemented features, numbers, steps, actions, components, parts, or combinations thereof, and should be understood not to preclude the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0030] In each step, identifiers (e.g., a, b, c, etc.) are used for convenience of explanation and do not describe the order of the steps; the steps may occur differently from the specified order unless a specific order is clearly indicated in the context. That is, the steps may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.

[0031] The present invention may be implemented as computer-readable code on a computer-readable recording medium, and the computer-readable recording medium includes all types of recording devices in which data that can be read by a computer system is stored. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. Additionally, the computer-readable recording medium may be distributed across networked computer systems, so that computer-readable code can be stored and executed in a distributed manner.

[0032] Unless otherwise defined, all terms used herein have the same meaning as generally understood by those skilled in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having meanings consistent with the context of the relevant technology and should not be interpreted as having an ideal or overly formal meaning unless explicitly defined in this application.

[0033]

[0034] FIG. 1 is a drawing illustrating a user-assembled artificial intelligence service platform system according to the present invention.

[0035] Referring to FIG. 1, a user-assembled artificial intelligence service platform system (100) may include a user terminal (110), an artificial intelligence service platform device (130), and a database (150).

[0036] The user terminal (110) can be connected to the artificial intelligence service platform device (130) via a network, can receive user-assembled artificial intelligence services provided by the artificial intelligence service platform device (130), and may correspond to a computing terminal operated by a user.

[0037] The user terminal (110) may be composed of a single or multiple units, and if composed of multiple units, it may include a first user terminal, a second user terminal, ..., the nth (n is a natural number) user terminal. For example, the user terminal (110) may be implemented as a smartphone, laptop, or computer capable of operating in connection with the artificial intelligence service platform device (130), but is not necessarily limited thereto and may be implemented as various devices including a tablet PC, etc.

[0038] Additionally, the user terminal (110) may correspond to a target customer terminal such as a company, developer, or marketer that requires the introduction of an artificial intelligence workflow itself. The user terminal (110) may correspond to a target user terminal such as a front-end developer or a group of developers who wish to generate revenue through the design of an artificial intelligence workflow.

[0039] The artificial intelligence service platform device (130) may be implemented as a server corresponding to a computer or program that performs user-assembled artificial intelligence services through artificial intelligence pipelining according to the present invention. The artificial intelligence service platform device (130) may be connected to a user terminal (110) via a wired network or a wireless network such as Bluetooth, WiFi, or LTE, and may transmit and receive data with the user terminal (110) through the wired or wireless network.

[0040] In one embodiment, the artificial intelligence service platform device (130) may be implemented as a cloud server and may provide user-assembled artificial intelligence services to a user terminal (110) through the cloud service. In one embodiment, when the artificial intelligence service platform device (130) is implemented as a cloud server, it may serve an artificial intelligence workflow modularized in the form of no-code blocks to the user terminal (110). In one embodiment, the artificial intelligence service platform device (130) may be implemented as a server corresponding to a computer or program that provides connections between various artificial intelligence blocks according to user dragging regarding artificial intelligence blocks during the process of serving the artificial intelligence workflow to the user terminal (110).

[0041] Additionally, the artificial intelligence service platform device (130) can receive user-defined tasks from the user terminal (110) and recommend artificial intelligence blocks. For example, the artificial intelligence service platform device (130) can receive a user's task order from the user terminal (110) via a chatbot and recommend multiple artificial intelligence blocks related to the task to the user so that the user can select them.

[0042] The database (150) may correspond to a storage device that stores various information required during the operation of the artificial intelligence service platform device (130). For example, the database (150) may store various types of no-code block-type artificial intelligence models created by a developer, or may store information regarding models and learning algorithms for creating user-assembled artificial intelligence models, but is not necessarily limited thereto, and may store information collected or processed in various forms during the process in which the artificial intelligence service platform device (130) performs a user-assembled artificial intelligence service method through artificial intelligence pipelining according to the present invention.

[0043] In addition, in FIG. 1, the database (150) is depicted as a device independent of the artificial intelligence service platform device (130), but it is not necessarily limited thereto and can be implemented as a logical storage device included in the artificial intelligence service platform device (130).

[0044]

[0045] Figure 2 is a diagram illustrating the system configuration of the artificial intelligence service platform device of Figure 1.

[0046] Referring to FIG. 2, the artificial intelligence service platform device (130) may include a processor (210), memory (230), user input / output unit (250), and network input / output unit (270).

[0047] The processor (210) can execute a user-assembled artificial intelligence service procedure according to an embodiment of the present invention, manage memory (230) that is read or written during this process, and schedule the synchronization time between volatile memory and non-volatile memory in memory (230). The processor (210) can control the overall operation of the artificial intelligence service platform device (130) and is electrically connected to memory (230), user input / output unit (250), and network input / output unit (270) to control the data flow between them. The processor (210) can be implemented as a CPU (Central Processing Unit) or GPU (Graphics Processing Unit) of the artificial intelligence service platform device (130).

[0048] The memory (230) may include an auxiliary storage device implemented as non-volatile memory such as an SSD (Solid State Disk) or HDD (Hard Disk Drive) and used to store all data required for the artificial intelligence service platform device (130), and may include a main memory implemented as volatile memory such as RAM (Random Access Memory). Additionally, the memory (230) may store a set of instructions that execute the user-assembled artificial intelligence service method according to the present invention by being executed by an electrically connected processor (210).

[0049] The user input / output unit (250) includes an environment for receiving user input and an environment for outputting specific information to the user, and may include an input device including an adapter such as a touch pad, touch screen, virtual keyboard, or pointing device, and an output device including an adapter such as a monitor or touch screen. In one embodiment, the user input / output unit (250) may correspond to a computing device connected via remote access, and in such case, the artificial intelligence service platform device (130) may be performed as an independent server.

[0050] The network input / output unit (270) provides a communication environment for connecting to a user terminal through a network and may include an adapter for communication such as a LAN (Local Area Network), MAN (Metropolitan Area Network), WAN (Wide Area Network), and VAN (Value Added Network). Additionally, the network input / output unit (270) may be implemented to provide short-range communication functions such as WiFi and Bluetooth, or wireless communication functions of 4G or higher, for wireless transmission of learning data.

[0051]

[0052] Figure 3 is a diagram illustrating the configuration of the artificial intelligence service platform device of Figure 1.

[0053] Referring to FIG. 3, the artificial intelligence service platform device (130) may include an artificial intelligence block curation unit (310), an artificial intelligence block selection unit (330), an artificial intelligence pipeline generation unit (350), a user-assembled artificial intelligence model generation unit (370), and a control unit (390).

[0054] The artificial intelligence service platform device (130) does not need to include all of the above functional configurations simultaneously, and depending on each embodiment, some of the above configurations may be omitted, or some or all of the above configurations may be selectively included. Additionally, the artificial intelligence service platform device (130) may be implemented as an independent module that selectively includes some of the above configurations, and may perform a user-assembled artificial intelligence service method through artificial intelligence pipelining according to the present invention through interoperability between each module. The operation of each configuration is described in detail below.

[0055] The AI ​​block curation unit (310) can curate multiple AI blocks. In one embodiment, the AI ​​block curation unit (310) can provide multiple AI blocks for sale on Marketplace. Here, each AI block may correspond to a user-created or user-ordered AI model. A user-created AI model may correspond to an AI model custom-made by the user, and a user-ordered AI model may correspond to an existing AI model developed by external developers or companies. The user can create a customized AI model by selecting, purchasing, or using blocks required on Marketplace and combining multiple blocks.

[0056] The AI ​​block curation unit (310) can be implemented by including an AI block library. The AI ​​block library stores various AI blocks and classifies the AI ​​blocks by function, use case, or performance indicator to help users easily find the blocks they need. Each AI block includes an AI model that can be used in various fields such as text processing, image recognition, speech analysis, recommendation systems, and natural language processing, and is modularized so that it can operate independently. For example, the AI ​​block library can classify AI blocks into text analysis, image processing, speech recognition, etc., and can assign tags to each block to facilitate searching and filtering. In addition, the AI ​​block library can record performance improvements or bug fixes for each version of the AI ​​block to provide users with updated blocks. The AI ​​block curation unit (310) can store metadata for each AI block so that users can easily view it. Block metadata may include the block's name, description, function, performance indicator, supported data types, requirements, etc. Block performance metrics can support users in selecting the optimal block that meets their needs by managing performance indicators such as block accuracy, processing speed, and resource usage.

[0057] In one embodiment, the AI ​​block curation unit (310) can recommend block combinations suitable for the user's order tasks (e.g., project goals). When the user inputs requirements, the AI ​​block curation unit (310) can analyze blocks and suggest optimal combinations. Additionally, the AI ​​block curation unit (310) can provide AI block recommendations based on the user's requirements and past usage patterns. For example, a user who frequently processes a specific type of data can automatically receive recommendations for related blocks. Furthermore, the AI ​​block curation unit (310) can automatically curate necessary blocks according to goals or tasks set by the user. When the user inputs desired functions (e.g., text summarization, image classification, etc.), the AI ​​block curation unit (310) can suggest AI blocks suitable for them.

[0058] The artificial intelligence block selection unit (330) can select at least first and second artificial intelligence blocks among a plurality of artificial intelligence blocks. In one embodiment, the artificial intelligence block selection unit (330) analyzes a user-defined task and recommends a plurality of candidate artificial intelligence blocks composed of heterogeneous artificial intelligence services that perform different functions, and can select at least first and second artificial intelligence blocks through user control. That is, the artificial intelligence block selection unit (330) can recommend artificial intelligence blocks suitable for the task required by the user and support intuitive selection.

[0059] The AI ​​block selection unit (330) can receive a user-defined task from the user terminal (110), perform analysis, and identify the functions of the AI ​​service required for the task. When the user-defined task is input in natural language, the AI ​​block selection unit (330) can perform natural language processing (NLP)-based analysis. For example, if a user inputs a request for 'automatic email generation in an integrated email management service,' the AI ​​block selection unit (330) can identify that the functions of the automatic email generation AI service are required. The AI ​​block selection unit (330) can break down the user-defined task by function and recommend the AI ​​blocks required to perform each function.

[0060] In one embodiment, the AI ​​block selection unit (330) can perform a matching between the functions required in a user-defined task and the AI ​​blocks stored on the platform to recommend the block most suitable for the user's required functions as a candidate AI block. Additionally, the AI ​​block selection unit (330) may recommend candidate AI blocks based on the performance and characteristics of the blocks, such as performance indicators, data types, processing speed, and resource usage. For example, if a user requires real-time processing, the AI ​​block selection unit (330) may prioritize recommending a block with a fast processing speed as a candidate AI block. The AI ​​block selection unit (330) may also provide a personalized experience by analyzing the user's past AI block usage history and, if there are blocks that were frequently used when performing similar tasks, prioritizing recommending those blocks as candidate AI blocks.

[0061] In one embodiment, the AI ​​block selection unit (330) may provide a user interface (UI) for selecting an AI block to a user terminal (110). The AI ​​block selection unit (330) may provide an intuitive UI that allows the user to visually compare and select a plurality of candidate AI blocks. For example, the AI ​​block selection unit (330) may provide a dashboard-type interface that compares the performance indicators and functions of each block. In one embodiment, the AI ​​block selection unit (330) may restrict accessible blocks or manage access rights to specific blocks according to the authority given to the user during the user's block selection process. The user may receive recommendations for blocks based on the requirements of the task and select them through visual tools or search and filtering.

[0062] The AI ​​pipeline generation unit (350) can generate an AI pipeline by connecting an output port in a first AI block to an input port in a second AI block. Each AI block has an input / output port and can exchange data through this port. The AI ​​pipeline generation unit (350) can connect the output port of the first AI block and the input port of the second AI block to ensure a continuous data flow. In one embodiment, the AI ​​pipeline generation unit (350) can visually configure the pipeline by connecting AI blocks using a drag-and-drop method. That is, after selecting the necessary AI blocks, the user can easily assemble the blocks by adding them to the workflow using an intuitive drag-and-drop method. If the first and second AI blocks selected by the user need to be interconnected, the AI ​​pipeline generation unit (350) can establish a connection between the blocks and generate an optimized pipeline to ensure compatibility and functional connection between the blocks without user intervention.

[0063] Additionally, the AI ​​pipeline generation unit (350) can determine the suitability of the connection between the output port and the input port during the connection process and create a data compatibility interface between the output port and the input port to perform data transfer between the first and second AI blocks. In one embodiment, the AI ​​pipeline generation unit (350) can determine the suitability of the connection between the input and output ports through the compatibility of data formats between the blocks when generating the pipeline. For example, if the first AI block outputs video data and the second AI block receives voice data as input, this connection is impossible, so a warning message can be provided and an option to modify can be presented to the user. On the other hand, if the video output is connected to the input of the video analysis block, an appropriate connection can be made. If the data formats between the blocks are different, the AI ​​pipeline generation unit (350) can perform Auto Data Transformation to convert the data into an appropriate format and integrate the data of different formats. Here, the AI ​​pipeline generation unit (350) can visually display the lines connecting each block to clearly show which direction the data flows, thereby allowing the user to easily manage the complex pipeline.

[0064] Additionally, the AI ​​pipeline generation unit (350) may determine the first AI block as a dedicated chatbot-based AI head capable of receiving user input, or may generate a shared chatbot-based AI head by connecting a shared chatbot to the input port of the first AI block. In one embodiment, the AI ​​pipeline generation unit (350) may determine the user's dedicated chatbot as the first AI block. The dedicated chatbot has a dedicated input port for receiving user input and can be adjusted to perform specific answers or tasks corresponding to user requests based on various natural language processing (NLP) models in a user-defined manner. The input port of the dedicated chatbot receives natural language data entered by the user, and the output port transmits the response generated by the chatbot to the next block of the pipeline. The dedicated chatbot-based AI head allows the user to control the chatbot and set the conversation flow through a customized interface. In one embodiment, the AI ​​pipeline generation unit (350) may receive multiple user inputs by connecting a basic AI chatbot shared with various users to the input port of the first AI block. For example, public chatbots include commercial chatbots such as OpenAI’s GPT series-based chatbots and Google’s Dialogflow. The input port of the public chatbot can receive various inputs from multiple users, and the output port can be connected to output the response generated by the chatbot to be delivered to the user or to be further processed in the next stage of the pipeline. The artificial intelligence pipeline generation unit (350) can preprocess user input data through a dedicated or public chatbot-based artificial intelligence head and convert it into a format that other blocks in the pipeline can understand. For example, the preprocessing work may include basic natural language processing (NLP), such as text normalization, tokenization, and stop word removal.The artificial intelligence pipeline generation unit (350) can receive user input from various channels (e.g., websites, mobile apps, voice interfaces, etc.) through a dedicated or public chatbot-based artificial intelligence head, and can process the data input from each channel in an integrated manner so that the user can access the chatbot on multiple devices or in various environments. Users can easily interact with the chatbot through natural language conversation and use the artificial intelligence service without understanding complex artificial intelligence models or algorithms. In one embodiment, the artificial intelligence pipeline generation unit (350) can flexibly configure pipelines suitable for each purpose by selectively using a dedicated chatbot and a public chatbot.

[0065] The user-assembled artificial intelligence model generation unit (370) can generate a user-assembled artificial intelligence model through artificial intelligence pipelining. In one embodiment, the user-assembled artificial intelligence model generation unit (370) can verify whether there is an error in a specific artificial intelligence block constituting the user-assembled artificial intelligence model by checking a user input-based artificial intelligence service through the user-assembled artificial intelligence model.

[0066] The user-assembled artificial intelligence model generation unit (370) can generate a single integrated artificial intelligence model by combining at least first and second artificial intelligence blocks selected by the user in a pipelining manner, and the blocks can be connected to interact by exchanging data. Here, each artificial intelligence block includes an artificial intelligence service of a specific function during a user-defined task and is connected to operate organically with one another to assemble an artificial intelligence model suitable for the user-defined task. The user-assembled artificial intelligence model generation unit (370) can verify errors that may occur in each block of the artificial intelligence model assembled by the user and, in this process, detect problems such as data transfer between blocks, inconsistencies in input data formats, and insufficient accuracy of output. In one embodiment, the user-assembled artificial intelligence model generation unit (370) can perform tests to see if each block operates as expected, recognize incorrect processing or abnormal output for specific data types to detect errors per block, and provide notifications to the user to correct them. For example, if a problem is detected in the connection between blocks, the user-assembled artificial intelligence model generation unit (370) can visually display it to notify the user and present recommendations for problem solving to support the user in resolving the problem.

[0067] Additionally, the user-assembled artificial intelligence model generation unit (370) can distribute the user-assembled artificial intelligence model within the platform.

[0068] The control unit (390) controls the overall operation of the artificial intelligence service platform device (130) and can manage the control flow or data flow between the artificial intelligence block curation unit (310), the artificial intelligence block selection unit (330), the artificial intelligence pipeline generation unit (350), and the user-assembled artificial intelligence model generation unit (370).

[0069]

[0070] FIG. 4 is a flowchart illustrating the process of a user-assembled artificial intelligence service through artificial intelligence pipelining of an artificial intelligence service platform device according to the present invention.

[0071] Referring to FIG. 4, the artificial intelligence service platform device (130) can curate a plurality of artificial intelligence blocks through the artificial intelligence block curation unit (310) (step S410). The artificial intelligence service platform device (130) can select at least the first and second artificial intelligence blocks among the plurality of artificial intelligence blocks through the artificial intelligence block selection unit (330) (step S430).

[0072] Additionally, the artificial intelligence service platform device (130) can generate an artificial intelligence pipeline by connecting an output port in the first artificial intelligence block to an input port in the second artificial intelligence block through the artificial intelligence pipeline generation unit (350) (step S450). The artificial intelligence service platform device (130) can generate a user-assembled artificial intelligence model through artificial intelligence pipelining via the user-assembled artificial intelligence model generation unit (370) (step S470).

[0073]

[0074] FIG. 5 is a diagram illustrating an embodiment of an artificial intelligence workflow for an artificial intelligence service platform device according to the present invention.

[0075] Referring to FIG. 5, the artificial intelligence service platform device (130) can provide a home screen such as FIG. 5 (a) to a user terminal (110). On the home screen of FIG. 5 (a), the user can receive recommendations for the artificial intelligence model workflow they want through a Large Language Model (LM) and can design it semi-automatically (①). The user can easily and conveniently order their work from the head LLM through a chatbot.

[0076] The artificial intelligence service platform device (130) can be implemented to include a repository where artificial intelligence pipelines can be registered and shared. Upon entering the repository, one can see nodes connected to various artificial intelligence neural networks, as shown in Fig. 5 (b) (②). Here, an artificial intelligence model workflow modularized in the form of no-code blocks can be served to the user. The artificial intelligence service platform device (130) can create artificial intelligence pipelines using a simple drag-and-drop interface that does not require coding. The user can test the artificial intelligence model workflow through a chatbot (③).

[0077]

[0078] FIG. 6 is a drawing showing an example of user-assembled artificial intelligence model generation of an artificial intelligence service platform device according to the present invention, and is an example of the production of an artificial intelligence service for detecting defective manufactured parts.

[0079] In FIG. 6, the artificial intelligence service platform device (130) receives user input through a chatbot-based artificial intelligence head (610). The user can place a work order for an artificial intelligence service task of “defective manufacturing part detection” through the head LLM. The artificial intelligence service platform device (130) can complete the workflow by creating an artificial intelligence pipelining, which is created by setting criteria suitable for the ordered user-defined task, segmenting the task into an AI work structure for each function, selecting artificial intelligence blocks containing the artificial intelligence model of the corresponding function by dragging, and connecting the input / output ports of the artificial intelligence blocks in functional order. Here, for a user-defined task regarding the detection of defective manufacturing parts, the Image Recognition AI block, Image Detection AI block, Defective Part Text Generation AI block, and Object Tracking AI block are easily and quickly dragged to connect the output port of the Image Recognition AI block to the input port of the Image Detection AI block, the output port of the Image Detection AI block to the input port of the Defective Part Text Generation AI block, and finally the output port of the Defective Part Text Generation AI block to the input port of the Object Tracking AI block to create a user-assembled AI model for detecting defective manufacturing parts.

[0080] Users can quickly create their own AI pipelines by forking AI pipelines created by other users stored in the repository, thereby reducing development time and allowing them to work more efficiently.

[0081]

[0082] The user-assembled artificial intelligence service platform device through artificial intelligence pipelining according to the present invention can reduce the time and cost of developing artificial intelligence services by more than 70%, and has the advantage of being easily developed even by unskilled workers without the need for additional hiring due to a shortage of developers and backend developers.

[0083] In addition, it can be utilized in various fields as middleware for creating AI workflows in areas requiring the adoption of artificial intelligence.

[0084]

[0085] Although the present invention has been described above with reference to preferred embodiments, those skilled in the art will understand that various modifications and changes can be made to the invention without departing from the spirit and scope of the invention as described in the following claims.

Claims

1. An AI block curation unit that curates multiple AI blocks; An AI block selection unit that selects at least first and second AI blocks among the plurality of AI blocks above; An AI pipeline generation unit that generates an AI pipeline by connecting an output port in the first AI block to an input port in the second AI block; and A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, comprising a user-assembled artificial intelligence model generation unit that generates a user-assembled artificial intelligence model through the above artificial intelligence pipelining.

2. In paragraph 1, the artificial intelligence block curation unit A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, characterized by providing the plurality of artificial intelligence blocks for market place sales, wherein each of the plurality of artificial intelligence blocks corresponds to a user-created or user-ordered artificial intelligence model.

3. In paragraph 1, the artificial intelligence block selection unit A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, characterized by analyzing user-defined tasks to recommend multiple candidate artificial intelligence blocks composed of heterogeneous artificial intelligence services performing different functions, and selecting at least first and second artificial intelligence blocks through user control.

4. In paragraph 1, the artificial intelligence pipeline generation unit A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, characterized by determining the connection suitability between the output port and the input port during the above connection process and creating a data compatibility interface between the output port and the input port to perform data transfer between the first and second artificial intelligence blocks.

5. In paragraph 4, the artificial intelligence pipeline generation unit A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, characterized by determining the first artificial intelligence block as a dedicated chatbot-based artificial intelligence head capable of receiving user input, or connecting a public chatbot to the input port of the first artificial intelligence block to generate a public chatbot-based artificial intelligence head.

6. In paragraph 1, the user-assembled artificial intelligence model generation unit A user-assembled artificial intelligence service platform device through artificial intelligence pipelining, characterized by verifying user input-based artificial intelligence services through the user-assembled artificial intelligence model and verifying whether there is an error in a specific artificial intelligence block constituting the user-assembled artificial intelligence model.

Images