Block code service device and method

Abstract

This system provides an artificial intelligence model generated by selecting block codes that correspond to user commands and concatenating the selected block codes. [Solution] The present invention discloses a block code service device. The block code service device according to the present invention includes an instruction analysis unit that generates at least one embedding vector based on a user's instruction, a database management unit that manages a database for a plurality of block codes configured in block code format, a block selection unit that selects at least one block code from the plurality of block codes that corresponds to at least one embedding vector, and a block connection unit that connects the selected block codes to generate an artificial intelligence model.

Description

Technical Field

[0005] ,

[0001] The present invention relates to a block code service apparatus and method, and more particularly, to a block code service apparatus and method for selecting a block code corresponding to a user's command, concatenating the selected block codes, and generating an artificial intelligence model.

Background Art

[0002] This invention is the result of an application supported by the '2025 Global Startup Commercialization Support Project' of Gyeonggi-do and the Gyeonggi-do Economic and Scientific Promotion Institute.

[0003] In modern software development environments, there is an increasing need for systems that automatically generate and recommend code in response to user commands. In particular, the block coding method has attracted attention in order to help users with little coding experience or non-experts easily implement software functions. Block coding visually arranges pre-defined codes in block form so that codes can be intuitively constructed without knowledge of complex grammar.

[0004] Although block coding reduces the difficulty of code development, there is still the inconvenience that users must directly select and arrange each block code. In particular, in large-scale operations, the process of selecting and combining block codes one by one is inefficient, and it can be difficult to find the optimal block code that meets specific needs. Therefore, there is a need for a technology that selects the block codes required by the user and concatenates the selected block codes so that the combination of the selected block codes becomes executable.

Summary of the Invention

Problems to be Solved by the Invention

[0005] The objective of the present invention is to provide an artificial intelligence model generated by selecting block codes corresponding to user commands and concatenating the selected block codes.

[0006] The problems that this invention aims to solve are not limited to those described above, and any other problems not mentioned should be clearly understood by an ordinary person from the following description. [Means for solving the problem]

[0007] A block code service device according to one embodiment of the present invention may include an instruction analysis unit that generates at least one embedding vector based on a user's instruction, a database management unit that manages a database for a plurality of block codes configured in block code format, a block selection unit that selects at least one block code from the plurality of block codes that corresponds to at least one embedding vector, and a block connection unit that concatenates the selected block codes to generate an artificial intelligence model.

[0008] A block code service method according to one embodiment of the present invention may include the steps of: generating at least one embedding vector based on a user's command; generating a database for a plurality of block codes configured in block code format; selecting at least one block code from the plurality of block codes that corresponds to at least one embedding vector; and concatenating the selected block codes to generate an artificial intelligence model. [Effects of the Invention]

[0009] The block code service device and method according to the present invention can provide an artificial intelligence model generated by selecting block codes corresponding to user commands and concatenating the selected block codes.

[0010] Furthermore, the block code service device and method according to the present invention can provide error information to the user's terminal if there are errors in the generated artificial intelligence model.

[0011] The effects of the present invention are not limited to those described above, and any other effects not mentioned should be clearly understood by an ordinary person from the following description. [Brief explanation of the drawing]

[0012] [Figure 1] This is a diagram showing a block code service system according to one embodiment of the present invention. [Figure 2] This is a drawing showing a block code service device according to one embodiment of the present invention. [Figure 3] This is a diagram illustrating a block code service method according to one embodiment of the present invention. [Figure 4] This is a diagram showing a block database according to one embodiment of the present invention. [Figure 5] This is a diagram showing a vector database according to one embodiment of the present invention. [Figure 6] This is a diagram showing the selection of block codes corresponding to user commands according to one embodiment of the present invention. [Figure 7] This is a diagram showing an interface for receiving user commands by one embodiment of the present invention. [Figure 8] This is a diagram showing an artificial intelligence model generated by concatenating block codes according to one embodiment of the present invention. [Modes for carrying out the invention]

[0013] The following describes exemplary embodiments of the present invention in detail with reference to the attached drawings. However, the present invention is not limited or restricted by these exemplary embodiments. Unless otherwise defined, all terms used herein (including technical and scientific terms) are used in a way that can be commonly understood by a person skilled in the art to which the present invention pertains, although this may change depending on the intent of the art practitioners in that art, precedents, the emergence of new technologies, etc.

[0014] Furthermore, terms defined in commonly used dictionaries should not be interpreted ideally or excessively unless explicitly and specifically defined otherwise. In certain cases, terms may be arbitrarily selected by the applicant, and in such cases, their meaning will be described in detail in the relevant explanatory section. Therefore, terms used in this invention must not be simple terminology, but rather defined based on the meaning of the term and the overall content of this invention.

[0015] Whenever a given part of this specification 'contains' a given component, this does not mean that other components are excluded, but rather that other components may be included, unless otherwise specified. Furthermore, the singular form used in this specification also includes the plural form unless otherwise specified. Also, the expression “at least one of a, b, and / or c” used throughout this specification may include 'a alone', 'b alone', 'c alone', 'a and b', 'a and c', 'b and c', or 'all of a, b, and c'.

[0016] On the other hand, terms such as “first and / or second” used herein can be used to describe various components, but this is used solely for the purpose of distinguishing one component from others, and not to limit it to the component referred to by that term. For example, without departing from the scope of the present invention, the first component may be referred to as the second component, and similarly, the second component may also be referred to as the first component.

[0017] In addition, terms such as “… section” and “… module” described in this specification mean units that process at least one function or operation, which can be implemented in hardware or software, or can be implemented by a combination of hardware and software. Also, embodiments of the present invention can be shown in functional block configurations and various processing stages in this specification. Such functional block codes can be implemented in various numbers of hardware or / and software configurations that execute specific functions. For example, embodiments of the present invention can adopt direct circuit configurations such as a memory, a processor, a logic, a look-up table, etc. that can execute various functions under the control of one or more microprocessors or other control devices.

[0018] Hereinafter, various embodiments of the present invention will be described in detail with reference to the accompanying drawings. In describing the embodiments, technical content that is widely known in the technical field to which the present invention belongs and is not directly related to the present invention will be omitted. This is to more clearly convey the gist of the present invention without obscuring it by omitting unnecessary explanations. For the same reason, some components in the accompanying drawings are exaggerated, omitted, or schematically illustrated. Also, the sizes of each component do not fully reflect the actual sizes. In this specification, the same reference numerals throughout the text can refer to the same or corresponding components.

[0019] FIG. 1 is a drawing showing a block code service system according to an embodiment of the present invention.

[0020] Referring to FIG. 1, a block code service system 100 according to an embodiment of the present invention can include a block code service device 110, a user terminal 120, a database 130, and a network 140.

[0021] The block code service device 110 and the user terminal 120 can communicate with each other through the network 140. The block code service device 110 and the user terminal 120 can communicate with each other through various networks 140 such as a wired network or a wireless network such as Bluetooth (registered trademark), WiFi, LTE, etc.

[0022] The user terminal 120 can receive an artificial intelligence model composed of block codes from the block code service device 110 as a response to the user's command. The user terminal 120 can correspond to computing terminals such as a computer, a smartphone, a tablet, etc. operated by the user. Also, the user terminal 120 can correspond to the terminals of target customers such as enterprises, developers, marketers, etc. who need block code services.

[0023] The user terminal 120 can transmit the user's command to the block code service device 110 through the network 140. Exemplarily, the user's command can be a command requesting the generation of an artificial intelligence model required by the user such as "Generate a model that can analyze CSV files".

[0024] The block code service device 110 can select block codes corresponding to the user's command received from the user terminal 120 in the database 130, and can generate an artificial intelligence model by concatenating the selected block codes so that the combination of the selected block codes is executable. The service of providing the artificial intelligence model generated by selecting and concatenating the block codes corresponding to the user's command in the present invention to the user terminal 120 can be called a block code service.

[0025] More specifically, the block code service device 110 embeds the user's command into a vector space, as shown in Figure 6, and then selects the block code S that is closest to the embedded vector C by distance d. In this invention, embedding can be defined as converting data composed of text into a vector form and arranging it in a multidimensional space. Furthermore, the vector space can be defined as a mathematical space where vectors representing the characteristics of the block code, such as the title, description, and tags, are arranged, as shown in Figures 5 and 6.

[0026] For example, the block code closest to the vector into which the user's command is embedded may be the block code responsible for reading the CSV file. The block code service device 110 can then further select any additionally required block codes based on the selected block code. For example, if a block code for reading the CSV file is selected, it can further select block codes for preprocessing the data contained in the CSV file, and block codes for analyzing and visualizing the data contained in the CSV file.

[0027] The block code service device 110 can concatenate selected block codes, taking into account the role of each selected block code. For example, if data preprocessing is required before analyzing CSV file data, the block code service device 110 can generate an artificial intelligence model that corresponds to the user's instructions by concatenating block codes in the following order: block code for reading the CSV file - block code for preprocessing the data contained in the CSV file - block code for analyzing and visualizing the data contained in the CSV file. However, this is merely an example, and the block code service device 110 can generate an artificial intelligence model that corresponds to the user's instructions by selecting and concatenating block codes in various ways.

[0028] In some embodiments, the block code service device 110 can be implemented as a cloud server and can provide block code services to the user terminal 120 through the cloud service.

[0029] Database 130 may include a block database 131 and a vector database 132.

[0030] The block database 131 can store data containing information that indicates the characteristics of a block, such as title, description, and tags, in block codes read via the API. For example, the block database 131 can store data containing title, description, and tags in block codes included in the Langchain library, a type of Python library.

[0031] The vector database 132 can store vectors that embed each block code stored in the block database 131. For example, each block code can be converted into a vector form according to its title, description, and tags and represented in a vector space, and the vector database 132 can store data representing each block code in a vector space.

[0032] As described above, the block code service device 110 according to the present invention can select block codes corresponding to user commands and provide an artificial intelligence model generated by concatenating the selected block codes to the user terminal 120.

[0033] Figure 2 is a diagram showing a block code service device according to one embodiment of the present invention.

[0034] Figure 2 can be explained with reference to Figure 1. Referring to Figure 2, a block code service device 210 according to one embodiment of the present invention may include an instruction analysis unit 211, a database management unit 212, a block selection unit 213, a block linking unit 214, and a control unit 215.

[0035] The instruction analysis unit 211 can receive user instructions from the user terminal 120 and embed the user instructions into a vector space to generate an embedding vector. In some embodiments, the block code service device 210 may further include an interface providing unit (not shown) that provides a user interface for receiving user instructions as the user terminal 120, as will be described later with reference to Figure 7.

[0036] The database management unit 212 may be configured to manage the database 130, such as by creating and deleting databases 130. For example, the database management unit 212 can read the Langchain block library via an API and store data including titles, descriptions, and tags in the block code contained in the Langchain library in the block database 131.

[0037] Furthermore, the database management unit 212 can store vectors in the vector database 132 that embed each of the block codes stored in the block database 131. However, this is merely an example, and the database management unit 212 can read various types of block libraries and store block codes in the block database 131. Also, the database management unit 212 may delete unnecessary block codes from the database 130.

[0038] The block selection unit 213 can select a block code from the block codes stored in the block database 131 that corresponds to a user instruction. For example, the block selection unit 213 can select at least one block code that corresponds to the embedding vector in which the instruction analysis unit 211 has embedded the user instruction. In some embodiments, the block selection unit 213 can select a block code that is closest in vector space to the embedding vector in which the user instruction has been embedded. Subsequently, the block selection unit 213 can further select additionally required block codes based on the selected block code.

[0039] The block linking unit 214 can generate an artificial intelligence model by linking the block codes selected by the block selection unit 213 in an executable manner. For example, as described above with reference to Figure 1, if data preprocessing is required before analyzing CSV file data, the block code service device 210 can generate an artificial intelligence model that responds to user commands by linking the block codes in the order of: block codes for reading the CSV file - block codes for preprocessing the data contained in the CSV file - block codes for analyzing and visualizing the data contained in the CSV file.

[0040] The control unit 215 can control the overall operation of the block code service device 210 and can manage the control flow or data flow between the instruction analysis unit 211, the database management unit 212, the block selection unit 213, and the block linking unit 214.

[0041] In some embodiments, the block code service device 210 may further include an error detection unit (not shown). The error detection unit (not shown) can detect errors in the artificial intelligence model generated by the block linking unit 214 and can provide error information to the user terminal 120 so that the user can correct the errors in the artificial intelligence model.

[0042] Figure 3 is a diagram illustrating a block code service method according to one embodiment of the present invention.

[0043] Figure 3 can be explained by referring to Figures 1 and 2 mentioned above.

[0044] Referring to Figure 3, a block code service method (S100) according to one embodiment of the present invention may include the steps of: generating at least one embedding vector based on a user's command (S110); generating a database for a plurality of block codes configured in block code format (S120); selecting at least one block code from the plurality of block codes that corresponds to at least one embedding vector (S130); and concatenating the selected block codes to generate an artificial intelligence model (S140).

[0045] Figure 3 illustrates that steps S110 to S140 are carried out sequentially, but is not limited to this. Some steps may be merged and carried out simultaneously, some steps may be omitted, or new steps may be added.

[0046] In step S110, at least one embedding vector can be generated based on the user's instructions. For example, referring to Figure 1, the user's instructions such as "Generate a model that can analyze a CSV file" can be provided from the user terminal 120, and the user's instructions can be embedded in the vector space to generate an embedding vector. The embedding vector can then serve as a criterion for selecting and concatenating block codes that conform to the user's instructions.

[0047] In step S120, a database can be generated for multiple block codes that are configured in block code format.

[0048] In some embodiments, step S120 may further include the steps of generating a block database containing multiple block codes and generating a vector database by embedding multiple block codes.

[0049] In some embodiments, each of the multiple block codes may include at least one of the block's title, description, and tags.

[0050] In step S130, it is possible to select at least one block code from among multiple block codes that corresponds to at least one embedding vector.

[0051] In some embodiments, a block code that is closest in vector space to the embedding vector generated in step S110 can be selected. Subsequently, additional block codes that are additionally required to make the artificial intelligence model executable can be selected based on the selected block code.

[0052] In step S140, an artificial intelligence model can be generated by concatenating the selected block codes. For example, an artificial intelligence model can be generated by concatenating the selected block codes, taking into account the role of each block code selected in step S130. For example, an artificial intelligence model can be generated by concatenating block codes in the following order: block code for reading a CSV file - block code for preprocessing the data contained in the CSV file - block code for analyzing and visualizing the data contained in the CSV file.

[0053] In some embodiments, the block code service method (S100) may further include the steps of detecting errors in the artificial intelligence model and providing error information to the user terminal. The block code service method (S100) may provide the error information of the artificial intelligence model to the user terminal 120 so that the user can correct the errors in the artificial intelligence model.

[0054] In some embodiments, the block code service method (S100) may further include a step of providing an interface for receiving user commands.

[0055] Figure 4 is a diagram showing a block database according to one embodiment of the present invention.

[0056] Figure 4 can be explained with reference to Figures 1 through 3. Referring to Figure 4, the block database 231 can contain various block codes, and each block code can contain information that indicates the block code, such as title, description, and tags.

[0057] Figure 4 illustrates a block database 231 containing six block codes (block codes A through F) as an example. However, this is merely illustrative, and the number of block codes included in the block database 231 and the information contained within each block code can be determined differently.

[0058] In some embodiments, the database management unit 212 can read the Langchain block library via an API and store the block code contained in the Langchain library in the block database 231 with a title, description, and tags.

[0059] Figure 5 is a diagram showing a vector database according to one embodiment of the present invention.

[0060] Figure 5 can be explained by referring to Figures 1 to 4 mentioned above. Referring to Figure 5, the embedded block code B can be placed on the vector space. For example, each of the embedded block code B shown in Figure 5 can be shown by referring to Figure 4 to illustrate the case where the aforementioned block code (block code A to block code F) is embedded on the vector space. Furthermore, each of the axes (X axis, Y axis, Z axis) of the vector space shown in Figure 5 can, for example, correspond to the title, description, tags, etc. of the block code.

[0061] Figure 6 is a diagram illustrating the selection of block codes corresponding to user commands according to one embodiment of the present invention.

[0062] Figure 6 can be explained with reference to Figures 1 to 5 described above. Referring to Figure 6, the instruction analysis unit 211 can generate an embedding vector C by embedding the user's instruction into the vector space. Subsequently, the block selection unit 213 can, as an example, select a block code S that is closest to the embedding vector C in the vector space, at a distance d. The block selection unit 213 can also further select additional block codes B that are necessary for the artificial intelligence model to execute, based on the selected block code S.

[0063] Figure 7 is a diagram illustrating an interface for receiving user commands according to one embodiment of the present invention.

[0064] Figure 7 can be explained with reference to Figures 1 to 6 described above. Referring to Figure 7, the block code service device 110 can provide an interface for receiving user commands to the user terminal 120, as shown in Figure 7. The user can input the commands they need into the command input window CMD of the user terminal 120, and the input user commands can be provided to the block code service device 110 via the network 140. The block code service device 110 can use the database 130 to generate an artificial intelligence model corresponding to the user's commands and provide it to the user terminal 120 via the network 140.

[0065] Figure 8 is a diagram showing an artificial intelligence model generated by concatenating block codes according to one embodiment of the present invention.

[0066] Figure 8 can be explained with reference to Figures 1 to 7 described above. Referring to Figure 8, the artificial intelligence model AM may include a first block code BL1, a second block code BL2, a third block code BL3, and a fourth block code BL4. However, this is merely an illustrative example for the sake of simplicity, and the technical concept of the present invention is not limited thereto.

[0067] The first block code BL1 may be the block code closest to the embedding vector generated by exemplifying the embedding of user instructions in a vector space.

[0068] For example, in order to execute the first block code BL1, it may require a chat model, which is a model trained on various training data to understand the user's questions and provide relevant answers; memory to store previous conversation history and key information with the user to maintain the context of the conversation; and prompt templates to maintain the tone and style of the AI ​​model AM's answers.

[0069] The block code service device 110 can generate an artificial intelligence model AM by concatenating the first block code BL1 with the second block code BL2, which performs a chat model function, the third block code BL3, which performs a memory function, and the fourth block code BL4, which performs a prompt template function, in order to execute the first block code BL1.

[0070] On the other hand, embodiments disclosed herein can be embodied in the form of a recording medium that stores computer-executable instruction words. The instruction words can be stored in the form of program code, which, when executed by a processor, can generate a program module to perform the operations of the disclosed embodiments. The recording medium can be embodied in a computer-readable recording medium. A computer-readable recording medium can include all types of recording media that store computer-readable instruction words. Examples include ROM, RAM, magnetic tape, magnetic disk, flash memory, optical data storage devices, and the like. The computer-readable recording medium can be distributed and executed by a network of connected computer systems.

[0071] The above-described content is a specific example for carrying out the present invention. The present invention includes not only the embodiments described above, but also embodiments that can be simply redesigned or easily modified. Furthermore, the present invention includes techniques that can be easily modified and implemented using the embodiments described above. Therefore, the scope of the present invention should not be limited to the embodiments described above, but should be defined not only by the claims described later, but also by claims equivalent to those of the present invention. [Explanation of Symbols]

[0072] 110 Block Code Service Device 120 user terminals 130 Databases 131 Block DB 132 Vector DB 140 Networks 211 Instruction Analysis Department 212 Database Management Department 213 Block Selection Section 214 Block connection section 215 Control Unit

Claims

1. An instruction analysis unit that generates at least one embedding vector based on the user's instruction, A database management unit that manages a database for multiple codes composed in block code format, A block selection unit that selects at least one block code from the plurality of block codes that corresponds to at least one embedding vector, A block code service device including a block connection unit that concatenates selected block codes to generate an artificial intelligence model.

2. The block code service device according to claim 1, further comprising an error detection unit that detects errors in the artificial intelligence model and provides error information to a user terminal.

3. The block code service device according to claim 1, further comprising an interface providing unit that provides an interface for receiving the aforementioned user commands.

4. The block code service device according to claim 1, wherein the database includes a block database containing the plurality of block codes and a vector database generated by embedding the plurality of block codes.

5. The block code service device according to claim 1, wherein each of the plurality of block codes includes at least one of the block title, description, and tag.

6. A step of generating at least one embedding vector based on user instructions, The stage of generating a database for multiple block codes, which are configured in block code format, The steps include selecting at least one block code from the plurality of block codes that corresponds to at least one embedding vector, A block code service method comprising the steps of: concatenating selected block codes to generate an artificial intelligence model.

7. The step of detecting errors in the artificial intelligence model, The block code service method according to claim 6, further comprising the step of providing error information to a user terminal.

8. The block code service method according to claim 6, further comprising the step of providing an interface for receiving the user's commands.

9. The step of generating the aforementioned database is: The step of generating a block database containing the aforementioned multiple block codes, The block code service method according to claim 6, further comprising the step of generating a vector database by embedding the plurality of block codes.

10. The block code service method according to claim 6, wherein each of the plurality of block codes includes at least one of the block title, description, and tags.

Images