Drawing creation support system

Abstract

To provide an information processing system capable of obtaining a drawing demanded by a user.SOLUTION: A drawing creation support system is constructed in which: a drawing is expanded, a hatching pattern is applied, and the hatching pattern is selected using a generative AI model; a magnification at which intervals of hatching lines become the optimum width is calculated by the generative AI model; and the creation of the drawing for application is assisted.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] One aspect of the present invention relates to an information processing system and an information processing method.

[0002] Note that one embodiment of the present invention is not limited to the above technical field. The technical field of one embodiment of the invention disclosed in this specification relates to an object, a method, or a manufacturing method. Alternatively, one embodiment of the present invention relates to a process, a machine, a manufacture, or a composition of matter. Therefore, more specifically, examples of the technical field of one embodiment of the present invention disclosed in this specification include a semiconductor device, a display device, a light-emitting device, a power storage device, a memory device, a driving method thereof, or a manufacturing method thereof. [Background technology]

[0003] In recent years, the development of language models using artificial neural networks has been actively pursued, with large-scale language models (LLMs) attracting particular attention. A large-scale language model is a natural language processing model trained using a large amount of data. A large-scale language model can realize, for example, a dialogue model that responds to user instructions. Non-Patent Document 1 discloses GPT-4 (Generative Pre-trained Transformer 4) (registered trademark) as a large-scale language model, and ChatGPT as a dialogue model. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Summary of ChatGPT / GPT-4 Research and Perspective Towards the Future of Large Language Models, Yiheng Liu et al. (Submitted on 4 Apr 2023, [online], Internet<URL:https: / / arxiv.org / abs / 2304.01852> Summary of the Invention [Problem to be solved by the invention]

[0005] When manufacturing an industrial product, it is necessary to determine the product layout design and manufacturing process, and adjust the process conditions to manufacture the product. In order to protect the industrial product, it is necessary to prepare the necessary drawings corresponding to the industrial product and file a patent application in a short time.

[0006] If the structure of an industrial product is relatively simple, it often does not take much time to prepare the application drawings. For example, you can take a photograph of the top or side of the industrial product with a camera and prepare the application drawings based on that photograph. You can also prepare the application drawings by referring to the design drawings based on the drafting method used to manufacture the industrial product.

[0007] However, in the case of precision equipment, such as industrial products with fine layouts like IC chips, it is possible to take a photograph of the exterior using a special microscope (electron microscope), but this is insufficient as a drawing for an application, and it takes time to create a corresponding drawing using CAD (Computer Aided Design) software.Furthermore, it is also possible to prepare a photograph by cutting or processing the industrial product into a thin sample and then taking a cross-sectional photograph, but it also takes time to create a corresponding drawing using CAD software.

[0008] In addition, it is common for drafters to use CAD software to create drawings for application documents, enlarging portions of the drawings to make them easier to understand or to emphasize the distinctive features of industrial products. Drawings for application documents are generally schematic, and are often created using CAD software based on the experience or subjective opinion of the drafter.

[0009] In addition, circuit diagrams for industrial products are created by designers using design support software, and verification is carried out using the design data. This design support software is often different from the CAD software used by the person creating the drawings for the application.

[0010] Thus, conventionally, different software programs have generally been used for drawings for application and drawings for design.

[0011] In addition, in some foreign countries, drawings that are faithful to the actual product are preferred for application drawings rather than drawings that have been edited to highlight certain parts.

[0012] Therefore, one of the objectives of one embodiment of the present invention is to enable a drawing creator to create drawings using a generative AI (Artificial Intelligence) model by effectively utilizing design data created by a designer in the manufacture of industrial products with fine layouts. Note that drawings obtained using the generative AI model are created based on drawings that are faithful to the actual product. [Means for solving the problem]

[0013] When manufacturing industrial products with fine layouts, drawing data that is faithful to the design values ​​used by the circuit designer is prepared, and portions of that drawing data are enlarged and used in the drawings for application. In order to clearly define areas or boundaries in the drawings for application, the same hatching pattern is used for the same material. Furthermore, hatching patterns must be selected appropriately, as enlargement magnifications of less than 1 can make it difficult to distinguish small areas.

[0014] Examples of industrial products with fine layouts include displays with fine pixel layouts that are difficult to see with the naked eye, sensors with fine image sensor layouts, and CPUs with semiconductor circuits made up of fine wiring and multiple semiconductor elements.

[0015] One aspect of the present invention provides a drawing creation support system that uses a generative AI model (also called a generative model) to enlarge drawings, add hatching patterns, and select hatching patterns to assist in the creation of drawings for application.

[0016] The configuration of the invention disclosed in this specification is a drawing creation support system that selects a portion of three-dimensional drawing data including semiconductor circuit design data based on a user's specified operation, and outputs a two-dimensional cross-sectional view of that portion enlarged at the same magnification, the two-dimensional cross-sectional view having multiple hatched areas, inputs the two-dimensional cross-sectional view into a trained generative AI model that outputs a magnification corresponding to the hatched lines of the input drawing, and when the spacing between the hatched lines in the hatched area is narrow, causes the generative AI model to calculate a magnification that will result in an optimal width for the spacing between the hatched lines.

[0017] The system also creates drawings that fit within the application drawing frame. Another aspect of the invention is a drawing creation support system that selects a portion of three-dimensional drawing data containing semiconductor circuit design data based on a first specification operation, and outputs a two-dimensional cross-sectional view of the selected portion enlarged at the same magnification. The two-dimensional cross-sectional view has multiple hatched areas, and the system inputs the two-dimensional cross-sectional view into a trained generative AI model that outputs a magnification corresponding to the hatched lines in the input drawing. If the spacing between the hatched lines in the hatched area is narrow, the system causes the generative AI model to calculate a magnification that optimizes the spacing between the hatched lines. The system then displays the new two-dimensional drawing data on a display device, and saves the new two-dimensional drawing data based on a second specification operation. The drawing frame (also called the drawing area) on A4 paper is 170mm wide x 255mm high for Japanese applications and 170mm wide x 262mm high for PCT applications. Additionally, the top, bottom, left and right margins are specified as 25mm, 10mm, 25mm and 15mm.

[0018] In each of the above configurations, the first position and the second position of the three-dimensional drawing data have distance information corresponding to the actual dimensions. That is, an arbitrary coordinate (first position) in the three-dimensional drawing data and an arbitrary coordinate (second position) in another location correspond to the design data dimensions. A finished product manufactured according to the design data dimensions corresponds to the actual dimensions.

[0019] In each of the above configurations, the three-dimensional drawing data is data that can be used to check the operation using a circuit verification tool. However, instead of the three-dimensional drawing data, multiple two-dimensional drawing data that are cross sections of the three-dimensional drawing data may be used.

[0020] In each of the above configurations, the 2D cross-sections are vector images. Vector images are composed of mathematical formulas, lines, and curves (using fixed points on a grid), so they can be resized infinitely larger (or smaller) without losing resolution. Vector images retain their image quality even when resized, making them easy to edit. Vector images can distinguish areas by using hatching for color or texture, or by applying a hatching texture to each area of ​​the image. [Effects of the Invention]

[0021] Since the design data prepared by the designer of an industrial product with a detailed layout can be used to create the drawings for application, it is possible to prepare drawings for application that reflect actual dimensions, which allows for efficient application preparation. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 is an example of a flowchart illustrating an information processing method. [Figure 2] FIG. 2 is an example of a flowchart illustrating the information processing method. [Figure 3] FIG. 3 is a diagram illustrating information to be input to the information processing system. [Figure 4] FIG. 4 is a diagram illustrating the configuration of the information processing system. [Figure 5] FIG. 5 is a block diagram illustrating the configuration of the information processing system. DETAILED DESCRIPTION OF THE INVENTION

[0023] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. However, the present invention is not limited to the following description, and it will be readily understood by those skilled in the art that various modifications can be made to the embodiments and details. Furthermore, the present invention should not be interpreted as being limited to the description of the embodiments shown below.

[0024] In this specification, the words "first" and "second" are used for the convenience of understanding the technical content or to identify each component. Therefore, the words "first" and "second" do not limit the number of each component. Furthermore, the words "first" and "second" do not limit the order of each component. Furthermore, the words "first" and "second" or identifying symbols used in this specification may not match the words or identifying symbols in the claims.

[0025] (Embodiment 1) In this embodiment, an information processing method according to one embodiment of the present invention will be described with reference to FIGS.

[0026] FIG. 1 is a diagram illustrating an information processing method according to one embodiment of the present invention. The information processing system according to one embodiment of the present invention includes components 10, 20, and 30. Component 10 may be, for example, a desktop computer that allows a user to input information or view information on a display unit. The display unit may be, for example, a liquid crystal display device, a light-emitting device (e.g., a light-emitting device having a light-emitting element such as an organic EL element in each pixel), an electrophoretic display device, a digital micromirror device (DMD), a plasma display panel (PDP), or a field emission display (FED). Component 20 may be, for example, a workstation, a server computer, or a supercomputer. Component 30 may be, for example, a large computer such as a server computer or a supercomputer. Component 30 may be larger in scale and have higher computing power than component 20. Component 30 may also have the function of performing processing using a generative AI model incorporating a graph neural network (GNN). Using a generative AI model incorporating a GNN enables effective learning, inference, and other processing of image data, such as vector-formatted patent drawings. Representative models using GNNs include, for example, GCN (Graph Convolutional Network) and GraphSAGE3. Raster format images (raster images) can also be used. If raster images are used, the component 30 does not necessarily require a GNN. While a raster image is represented by a collection of multiple "pixels," a vector image is represented by multiple "points" and the "lines" connecting them.

[0027] First, two-dimensional drawing data (two-dimensional cross-sectional view of a semiconductor substrate) prepared in advance is used. Hatching patterns are used in certain parts of the two-dimensional drawing data for clarity. The two-dimensional drawing data is also a vector image. A vector image can be rephrased as an image in vector format. For efficiency reasons, it is preferable to create two-dimensional drawing data by converting data (such as three-dimensional drawings) used in design support software (such as Process Explorer) for manufacturing semiconductor devices including semiconductor circuits.

[0028] The data used in the design support software (such as three-dimensional drawing data) is data that faithfully reproduces the actual dimensions of the product to be manufactured in advance, so it contains distance information corresponding to the actual dimensions, and it is preferable that it can be used directly for verification processing with semiconductor circuit verification software (operation confirmation using commercially available circuit verification tools).

[0029] In the component 10, the user inputs information IN1, which is a prepared vector image (A), into the drawing creation support system through a first specifying operation (S1: Step 1). The drawing creation support system applies hatching to each area in the vector image (A), which is the input two-dimensional drawing data.

[0030] Information IN1 includes image data displaying a list of hatching patterns. Fig. 3 shows image data displaying a list of 12 types of hatching patterns (hatching pattern A to hatching pattern L) as an example of information included in information IN1. Note that this is just an example and is not limited to 12 types, and there may be fewer than 12 types or 13 or more types. A hatching pattern is made up of one or more hatching lines, and the multiple hatching lines are arranged at equal intervals.

[0031] In addition, a vector image (A) can be obtained that shows the area B that the user wants to enlarge, framed by a dotted line. In this embodiment, an example is shown in which two figures are arranged on one drawing, and one vector image (A) is prepared and arranged by the user. The other figure corresponds to the area B (the area selected by the user) that the user wants to enlarge in the vector image (A).

[0032] Next, the component 20 of the drawing creation support system enlarges area B included in the vector image (A) using a vector image editing application provided in the drawing creation support system. A plurality of pieces of information IN2, each conditioned on an enlargement magnification for area B in the vector image (A), are generated, specifically, vector images (B1) to (Bn (n is an integer of 2 or greater)) obtained by enlarging area B. The enlargement magnification and the vector images (B1) to (Bn) are associated in advance. An enlargement magnification of 1x is called a constant magnification, and the enlargement magnification includes a constant magnification. When the enlargement magnification is less than 1x, it can be said to be a reduction, but is included in the enlargement magnification.

[0033] Next, the drawing creation support system sends the vector images (B1) through (Bn) and instruction PT1 as a prompt to component 30 (S2: step 2). For example, instruction PT1 might be, "Please tell me an image in which the hatching in each area of ​​the attached image can be distinguished." Component 30 has the function of processing the input instruction using a generative AI model. The generative AI model is a trained generative AI model that outputs a magnification corresponding to the spacing of the hatching lines in the input drawing. In this configuration, instruction PT1 does not use the statement, "Please tell me the hatching magnification so that the hatching in each area of ​​the attached image can be distinguished." By using this statement in instruction PT1, hatching converted at an appropriate magnification is provided to the user. However, some users manage hatching by linking it to a color and a name, and if the number of hatchings increases infinitely with repeated use of this system, this management becomes difficult.

[0034] "Each hatch is distinguishable" refers to the ability to distinguish a hatch pattern from other hatch patterns in a given area. Some hatch patterns are considered indistinguishable if the area is smaller than the spacing between the hatch lines, or if the boundaries of the hatch lines are parallel to the hatch lines. Furthermore, because there are limits to the line thicknesses permitted in patent drawings, hatch patterns are judged as being distinguishable when they exceed these limits. For example, the human eye cannot recognize patterns printed at a resolution of approximately 300 dpi (dots per inch). For hatch patterns with spacing between hatch lines greater than 300 dpi (which is indistinguishable from the human eye), when the surrounding area of ​​the drawing is filled in with black, the hatch pattern in one area can be used in the drawing. If it is distinguishable, it can be used in the drawing.

[0035] Next, the component 30 returns a response to the instruction statement PT1 to the component 20 (S3: Step 3).

[0036] Next, the drawing creation support system obtains the identifiable image output from the generative AI model of component 30 to component 20, and obtains information IN3, specifically the minimum value of the enlargement ratio, from the correspondence. In this way, the optimal drawing magnification is determined (S4: Step 4).

[0037] Note that some patent offices in other countries may specify a resolution. For example, the maximum resolution for drawings in PCT electronic filing software is 400 dpi. Meanwhile, the U.S. Patent and Trademark Office and the Chinese Patent Office specify a maximum resolution of 300 dpi. Therefore, it is recommended that you add a directive to calculate the minimum magnification factor, taking into account the resolution of the patent office of the country where the user intends to file.

[0038] Next, the drawing creation support system arranges the vector image (A) and a new vector image (B) enlarged by a magnification equal to or greater than the minimum magnification, i.e., an optimal magnification, within the range of a pre-specified format, and displays the image on the display unit of the display device to propose to the user. The user can confirm the proposed drawing (image) as an answer on the display unit of component 10, and can save the proposed drawing (image) based on a second specifying operation (S5: step 5). The saving location may be component 10 or component 20.

[0039] By using the above procedure, the user can quickly prepare the drawing for the application they desire, specifically a single drawing in which vector image (A) and vector image (B) of enlarged area B are arranged side by side.

[0040] This embodiment mode can be freely combined with other embodiment modes.

[0041] (Embodiment 2) In this embodiment, an information processing method according to one embodiment of the present invention will be described with reference to FIG.

[0042] This embodiment is the same as the first embodiment up to step 3 (S3), and therefore a detailed description thereof will be omitted.

[0043] Furthermore, with regard to the symbols used in the following description, the contents explained in embodiment 1 can be applied to symbols that are the same as those used in embodiment 1. Therefore, detailed explanations of the meanings and definitions of symbols may be omitted below.

[0044] This embodiment shows an example of arranging multiple cross-sectional structural diagrams (two-dimensional cross-sectional diagrams) within one drawing. Specifically, within the range of a pre-specified format, a total of three figures are arranged: vector image (A), vector image (B) enlarged by a magnification equal to or greater than the minimum value, and vector image (C). Vector image (B) corresponds to an enlarged view of area B in vector image (A), and vector image (C) corresponds to an enlarged view of area C in vector image (A).

[0045] Similar to the generation of images (B1) through (Bn) in the first embodiment, vector images (C1) through (Cn) are generated, and instruction PT2 is set to "Please advise how to arrange the attached vector image (A), one of vector images (Bm) through (Bn), and one of vector images (Cm) through (Cn) so that they fit as evenly as possible within the specified format. If they do not fit, please add them in the same format. If they do fit, please select vector images (Bm) through (Bn) and vector images (Cm) through (Cn) that are as large as possible." Instruction PT2 and each vector image are sent as prompts to the generative AI model of component 30 (S6: Step 6). Note that 1 <m<nとする。

[0046] The generative AI model included in component 30 preferably includes a GNN. A GNN is a graph neural network that is suitable for solving problems related to shapes, such as vector images. For example, a GNN may be able to grasp the positional relationship between a hatched area (also called a hatched area) and adjacent areas, effectively solving the problem.

[0047] As a result, the drawing creation support system obtains information IN3 using the generative AI model to arrange vector image (A), vector image (B) obtained by enlarging area B, and vector image (C) obtained by enlarging area C within the format. Information IN3 may, for example, be the enlargement factor and layout coordinates. Information IN3 is returned to component 20 in response to instruction statement PT2 (S7: Step 7).

[0048] The component 20 creates a drawing (a new two-dimensional cross-sectional drawing) for application based on the layout information contained in the received information IN3 (S8: Step 8).

[0049] The drawing creation support system proposes a drawing (a new two-dimensional cross-sectional drawing) for application to the user via the component 10. This proposed drawing is a drawing that fits within a predetermined frame for the drawing for application. The proposed drawing for application is displayed on the display unit of the component 10, and the user confirms it (S9: step 9). After confirming it, the user can also save the proposed drawing (image) based on a second specifying operation. The location where the drawing (image) is saved may be the component 10 or the component 20.

[0050] By following the above procedure, the user can quickly prepare the drawings for the application they desire, specifically a single drawing in which vector image (A) is arranged with vector image (B) and vector image (C).

[0051] Although an example of three vector images in one drawing has been shown, this is not particularly limited, and application is also possible to the case where four or more vector images are arranged in one drawing.

[0052] Also, after creating vector images (Bm) to (Bn) and vector images (Cm) to (Cn), you can obtain the FFT spatial frequency of each hatching and enter in the prompt as supplementary data whether the diameter of each area is n times the wavelength.

[0053] This embodiment mode can be freely combined with Embodiment Mode 1.

[0054] (Embodiment 3) An information processing system according to one embodiment of the present invention will be described below with reference to FIGS.

[0055] <Example of information processing system configuration> 4 shows an example of the configuration of each component of an information processing system according to one embodiment of the present invention and a network connecting these components. It also shows an example of the flow of data exchanged between each component of the information processing system.

[0056] FIG. 5 shows a more detailed example of the flow of data exchanged between the components of the information processing system shown in FIG.

[0057] In Figures 4 and 5, the components are classified by function and shown as independent components or blocks, but in reality it is difficult to completely separate the components by function, and one component may be involved in multiple functions.

[0058] 4, an information processing system according to one embodiment of the present invention includes a component 10, a component 20, and a component 30. In the information processing system, predetermined data is exchanged between the components via a network 50.

[0059] Hereinafter, functions of an information processing system according to an embodiment of the present invention will be described for each component constituting the information processing system with reference to Figures 4 and 5. Note that some of the descriptions regarding data exchange between the components may be repeated.

[0060] <<Component 10 configuration example>> The component 10 can accept data input by a user. The component 10 has at least an environment in which a user can use an image editing application (e.g., Vectorworks (registered trademark)) in advance. Furthermore, the component 10 can present data output by the component 20 to the user by displaying it on a display unit.

[0061] For example, dedicated application software or a web browser may be operated. A user can access the information processing system via either of these, thereby receiving services using the information processing system according to one embodiment of the present invention.

[0062] Component 10 has a function of accepting information input by a user (information IN1 shown in FIG. 4) in process T1 indicated by an arrow in FIG. 5, and transferring it to component 20. Information IN1 is vector image data.

[0063] In the following, we will explain a case in which an information processing system of one embodiment of the present invention is used to create a drawing in which a vector image and a vector image that is an enlarged portion of that vector image are arranged side by side at an appropriate magnification within a single drawing.However, this is not limited to this, and the information processing system can be used in a variety of cases.

[0064] Information IN1 is vector image data that includes a lot of design information, such as wiring width (line thickness), layout information including three-dimensional data, hatching pattern information (line type of hatching lines, spacing between hatching lines, etc.), and material information.

[0065] Information IN2 is a plurality of vector image data prepared by component 20, with enlargement magnifications set as conditions for regions in the vector image. Because component 20 only enlarges vector images, it does not need to have an environment in which an image editing application (e.g., Vectorworks) can be used, and an image editing application different from component 10 can also be used.

[0066] Information IN3 is text data of the minimum value of the enlargement magnification calculated from the plurality of vector image data input as information IN2.

[0067] Furthermore, the component 10 has a function of receiving information generated by the component 30 (information IN3 shown in FIG. 4) and providing it to the user in process T4 indicated by an arrow in FIG. 5. Details of the information IN3 have been described above.

[0068] Furthermore, the component 10 has a function of receiving information (information OUT shown in FIG. 4) generated by the component 30 in process T6 indicated by an arrow in FIG. 5. Details of the information OUT will be described below in the section "Configuration Example of Component 20."

[0069] <<Component 20 configuration example>> The component 20 preferably has a function as a parallel computer. By using it as a parallel computer, it is possible to perform large-scale calculations required for learning and inference in artificial intelligence (AI), for example.

[0070] Furthermore, the component 20 can perform processing using a natural language processing model that uses an AI model.

[0071] For example, processing can be performed using natural language processing models such as BERT (Bidirectional Encoder Representations from Transformers), T5 (Text-to-Text Transfer Transformer), GPT-3 (registered trademark), GPT-3.5 (registered trademark), GPT-4 (registered trademark), LaMDA (Language Model for Dialogue Applications), PaLM (Pathways Language Model), and Llama2.

[0072] Component 20 has a function of receiving information IN1 passed from component 10 in process T1 indicated by an arrow in FIG.

[0073] Component 20 also has the function of passing instruction statement PT1 or instruction statement PT2 to component 30. Instruction statement PT1 or instruction statement PT2 is a statement to be passed to component 30, including an instruction content using information IN1 passed from component 10.

[0074] The instruction PT1 or PT2 describes in natural language the specific processing to be performed by the component 30. An example of the instruction PT1 is "Please provide an image in which the hatching in each area of ​​the attached image can be identified."

[0075] The component 20 may be configured to have a function of automatically generating the directive PT1 in accordance with the contents of the information IN1 and information IN2.

[0076] An example of instruction PT2 is, "Please tell me how to arrange the attached image (A), one of the images (Bm)-(Bn), and one of the images (Cm)-(Cn) so that they fit as evenly as possible within the specified format. If they do not fit, please add them in the same format. If they do fit, please select images (Bm)-(Bn) and images (Cm)-(Cn) that are as large as possible." Note that one instruction can be divided into a first instruction and a second instruction. The second instruction is a statement written as an additional condition accompanying the first instruction. Note that there can be two or more additional conditions, and the number of statements written in instruction PT2 can be increased depending on the number of conditions desired by the user.

[0077] Furthermore, in process T2 indicated by an arrow in FIG. 5, component 20 has the function of transferring instruction statement PT1 and the image data to component 30 and causing component 30 to execute the instructions described in instruction statement PT1.

[0078] Component 20 also has the function of transferring directive statement PT2 to component 30 in process T5 indicated by an arrow in FIG. 5, and causing component 30 to execute the instructions described in directive statement PT2.

[0079] The information OUT output from component 30 shown in FIG. 4 indicates that it contains information that may be the cause of difficulty in processing in component 30, specifically, insufficient information, or inability to obtain calculation results.

[0080] 4, the component 20 can also have a function to receive the information OUT generated by the component 30. Furthermore, the component 20 can also have a function to generate a script that is suitable for an image editing application such as the above-mentioned Vectorworks based on the information OUT and pass the script to the component 10.

[0081] <<Configuration example of 30 components>> For example, a large computer such as a server computer or a supercomputer can be used as the component 30. Note that the component 30 is larger in scale than the component 20 and has higher calculation capabilities.

[0082] Furthermore, the component 30 preferably has a function as a parallel computer. By using it as a parallel computer, it is possible to perform large-scale calculations necessary for learning and inference of an AI model, for example.

[0083] The component 30 also processes images and text, such as hatching, etc. The component 30 can be a multimodal model or a base model that handles images and text.

[0084] For example, CLIP, DALL-E, Flamingo, United-IO, Gato, Imagen, and Parti are available as base models for processing images and text.

[0085] Furthermore, component 30 is configured to process raster images using an AI model when it receives them. Furthermore, when it receives vector images, it may have a function to process them using an AI model incorporating a GNN. By using an AI model incorporating a GNN, it is possible to perform learning, inference, etc. on image data (vector format) such as drawings for application. Representative models using GNN include, for example, GCN and GraphSAGE3.

[0086] Note that a person who provides a service using an information processing system according to one embodiment of the present invention does not necessarily have to own the component 30. For example, a service provider can use part of a service provided by another business operator as the component 30.

[0087] In this way, the component 30 can execute the instructions of the instruction statement PT1 or PT2 by performing processing using a model that handles images and text.

[0088] Component 30 also has the function of passing information OUT to component 10 in process T6 indicated by an arrow in FIG.

[0089] Component 30 has a function of performing processing using a large-scale language model (different from the large-scale language model that component 20 may have). The large-scale language model has already learned a data set. As a result, as described above, component 30 can determine whether information IN1 and information IN2 are sufficient as input content based on information IN1 and information IN2, directives PT1 and PT2 passed from component 20.

[0090] The information processing system according to one embodiment of the present invention has the above-described various functions, and therefore, regardless of the user's sense, skill level, etc., the user can simply provide the necessary information to the information processing system to easily create attractive drawings for application (e.g., drawings in which each area is clearly identified with an appropriate hatching pattern). Therefore, a novel information processing system with excellent convenience can be provided.

[0091] An information processing system according to one aspect of the present invention includes an information processing device that performs the functions of the above-described components.

[0092] By configuring an information processing system according to one embodiment of the present invention using a plurality of information processing devices, the load related to information processing can be distributed. [Explanation of symbols]

[0093] 10 Components 20 Components 30 Components 50 Network

Claims

1. For three-dimensional drawing data including semiconductor circuit design data, A part is selected based on the user's specified operation, and a two-dimensional cross-sectional view of that part is output at the same magnification. the two-dimensional cross-sectional view has a plurality of hatched regions; A drawing creation support system in which the two-dimensional cross-sectional view is input into a trained generative AI model that outputs a magnification corresponding to the hatching lines of the input drawing, and when the spacing between the hatching lines in the hatched area is narrow, the generative AI model calculates a magnification that will result in an optimal width for the spacing between the hatching lines.

2. For three-dimensional drawing data including semiconductor circuit design data, A part is selected based on the first specifying operation, and the part is enlarged at the same magnification to produce drawing data for a two-dimensional cross-sectional view. the two-dimensional cross-sectional view has a plurality of hatched regions; A drawing creation support system that inputs the two-dimensional cross-sectional view into a trained generative AI model that outputs a magnification corresponding to the hatching lines of the input drawing, and if the spacing between the hatching lines in the hatched area is narrow, has the generative AI model calculate a magnification that will make the spacing between the hatching lines optimal, displays new two-dimensional drawing data sized to fit within the frame of the drawing for application on the display unit of a display device, and saves the new two-dimensional drawing data based on a second specification operation.

3. 3. A drawing creation support system according to claim 1, wherein the first position and the second position of the three-dimensional drawing data have distance information corresponding to actual dimensions.

4. 3. A drawing creation support system according to claim 1, wherein the three-dimensional drawing data is data that can be used to check operation using a circuit verification tool.

5. 3. The drawing creation support system according to claim 1, wherein the two-dimensional cross-sectional view is a vector image.

Images