Program, information processing method, and information processing apparatus

Abstract

This invention provides a program, an information processing method, and an information processing device that efficiently generate proposal information regarding welding robot systems requested by customers based on acquired customer work images. [Solution] The program causes the computer to perform the following processes: acquire customer work images taken of the welding work of a customer requesting a welding robot system; derive attribute information of the customer's welding work based on the acquired customer work images; derive similar work similar to the customer's welding work based on the derived attribute information; derive project information to which the similar work is applied based on the derived similar work; and generate proposal information regarding the welding robot system requested by the customer based on the derived project information.

Description

Technical Field

[0001] The present invention relates to a program, an information processing method, and an information processing apparatus.

Background Art

[0002] For example, Patent Document 1 mentions a support system that accepts model data of an article from a client terminal and supports setting processing conditions for the article.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the support system disclosed in Patent Document 1, no consideration is given to generating proposal information regarding the welding robot system desired by the customer based on the acquired customer work image.

[0005] The present invention has been made in view of such circumstances, and an object thereof is to provide a program or the like that can efficiently generate proposal information regarding the welding robot system desired by the customer based on the acquired customer work image.

Means for Solving the Problems

[0006] A program according to one aspect of this disclosure causes a computer to perform the following processes: acquire customer work images taken of a welding workpiece of a customer requesting a welding robot system; derive attribute information of the customer's welding workpiece based on the acquired customer work images; derive similar workpieces similar to the customer's welding workpiece based on the derived attribute information; derive case information to which the similar workpiece is applied based on the derived similar workpieces; and generate proposal information regarding the welding robot system requested by the customer based on the derived case information.

[0007] In this embodiment, the welding robot system is composed of multiple devices or equipment, such as a manipulator, control device, welding power supply, wire feeding control device, and welding torch, and the device configuration of the welding robot system is determined according to the workpiece to be welded. At the site of a customer requesting the introduction or upgrade of a welding robot system, the customer's workpiece is imaged by an information terminal, such as a smartphone, used by a person in charge of the manufacturer or seller of the welding robot system. The control unit of the information processing device acquires the imaged image (customer workpiece image) from the information terminal. Based on the acquired customer workpiece image, the control unit of the information processing device derives attribute information of the customer's welding workpiece included in the customer workpiece image. This attribute information may include at least one of the following: workpiece shape, size, welding location, welding method, welding conditions, welding position, and material. Based on the derived attribute information (attribute information of the customer's welding workpiece), the control unit of the information processing device derives similar workpieces that have attribute information similar to or approximate to the derived attribute information, and identifies the case information to which the derived similar workpiece is applied. The case information includes various sales and technical information used in past sales activities or sales performance, and may consist of sales performance data stored in sales support systems such as SFA (Sales Force Automation) and CRM (Customer Relationship Management). The number of similar workpieces derived may be one or more, and the control unit of the information processing device may derive case information to which each of these multiple similar workpieces applies. Based on this derived case information, the control unit of the information processing device generates and outputs proposal information regarding the welding robot system requested by the customer. In generating this proposal information, the control unit of the information processing device may input the derived case information into, for example, a large-scale language model (LLM).In this case, the control unit of the information processing device may generate a prompt as an instruction to the Large-Scale Language Model (LLM), for example, instructing it to generate proposal information regarding the welding robot system requested by the customer using the derived case information, and inputting this prompt into the LLM. In this way, since proposal information regarding the welding robot system requested by the customer is generated based on welding work images taken at the customer's site, etc., and using case information accumulated from past sales activities, etc., even a person with relatively little experience can efficiently generate the proposal information.

[0008] A program according to one aspect of this disclosure derives attribute information of a welded work included in a customer work image by inputting the acquired customer work image into a welding work model that has been trained to output attribute information of a welded work when an image including a welded work is input.

[0009] In this embodiment, the control unit of the information processing device inputs the acquired customer work image into a welding work model, and derives attribute information of the welding work included in the customer work image from the welding work model. The welding work model is trained to output attribute information of the welding work when an image containing a welding work is input, and the actual file of the welding work model is stored, for example, in the storage unit of the information processing device. By using the welding work model in this way, attribute information of the welding work included in the customer work image can be efficiently derived.

[0010] A program according to one aspect of this disclosure derives similar workpieces by performing a search on a workpiece attribute information database in which multiple welded workpieces and the attribute information of each of the multiple welded workpieces are stored in association with each other, wherein the attribute information includes at least one piece of information from among the workpiece shape, size, and welding location of the welded workpiece, and the derived attribute information is used to derive similar workpieces.

[0011] In this embodiment, the attribute information output by the welding work model includes at least one piece of information from among the work shape, size, welding location, welding method, welding conditions, and welding position of the welding work, and may include all of this information. The storage unit of the information processing device stores a work attribute information database in which each of the multiple welding workpieces and the attribute information of each of these multiple welding workpieces are associated and stored. The data stored in the work attribute information database (attribute information of the welding workpieces) may be normalized or structured data using, for example, data such as various forms, documents, and photographs associated with the history or results of past sales projects. The control unit of the information processing device extracts similar workpieces to the customer's welding workpiece by performing a search process on the information stored in the management items of the work attribute information database for each item of various information such as work shape, size, welding location, welding method, welding conditions, and welding position included in the attribute information (attribute information of the welding workpiece included in the customer work image), which is the output result from the welding work model. The control unit of the information processing device, when performing the search process, is not limited to searching for similar workpieces that exactly match the attribute information of the customer's welded workpiece. For example, by specifying the similarity of the workpiece shape in the attribute information, the range specification for the size, the degree of agreement in the welding conditions, etc., it may extract (derive) welded workpieces as similar workpieces that have attribute information with a similarity to the attribute information of the customer's welded workpiece that is equal to or greater than a predetermined value. By using a workpiece attribute information database in this way, similar workpieces with similar attribute information to the attribute information of the welded workpiece included in the customer's workpiece image can be efficiently derived.

[0012] A program according to one aspect of this disclosure includes, in which the case information includes at least one of the following: attribute information of a welding workpiece, equipment configuration of a welding robot system, and information regarding estimates and delivery records, and derives the case information by performing a search on the derived similar workpiece in a case database in which the case information is stored.

[0013] In this embodiment, the storage unit of the information processing device stores a case database in which case information is stored. The case information includes at least one of the following pieces of information: specifications of a welding workpiece, equipment configuration of a welding robot system, estimates, and delivery records, and may include all of these pieces of information. The specifications of a welding workpiece included in the case information may include, for example, the workpiece shape, size, material, welding location, welding method, welding conditions, and welding position of the welding workpiece, and may include all types of information (attribute items) in the attribute information of the welding workpiece. Therefore, the case database stores the specifications (attribute information) of a welding workpiece and technical information or sales information such as equipment configuration, estimates, and delivery records related to the welding robot system corresponding to the welding workpiece, in correspondence or association. The control unit of the information processing device performs a search process on the case database configured in this way using similar workpieces to derive case information that includes technical information or sales information such as equipment configuration, estimates, and delivery records related to the welding robot system corresponding to the similar workpiece, thus enabling efficient processing of deriving the case information.

[0014] A program according to one aspect of this disclosure obtains customer specifications requested by the customer, and derives the project information based on the obtained customer specifications and the derived similar work.

[0015] In this embodiment, the control unit of the information processing device acquires customer work images taken of the customer's welded workpiece, as well as information regarding customer specifications requested by the customer. This customer specification information includes, for example, welding methods, welding conditions, quotation conditions, welding power supply, wire feeding, or required quality, i.e., conditions for operating the welding robot system, specification limitations for equipment or facilities constituting the welding robot system, or information regarding welding surface standards and yield in the welded workpiece. Based on the customer specifications and similar workpieces, the control unit of the information processing device derives case information, and can efficiently generate proposal information regarding a suitable welding robot system adapted to the customer's welded workpiece.

[0016] A program according to one aspect of this disclosure includes, in which the customer specifications include at least one specification item of welding method, welding conditions, quotation conditions, welding power source, and wire feeding, and the project information includes information corresponding to the specification item in the customer specifications, and the program derives the project information by performing a search on a project database in which the project information is stored using the derived similar work and the acquired customer specifications.

[0017] In this embodiment, the control unit of the information processing device performs a search on the case database using the acquired customer specifications along with the attribute information of the customer work image, and extracts (derives) case information of welding work that has attribute information similar to the said attribute information and that meets the customer specifications. By taking customer specifications into account in this way and performing a search on the case database, it is possible to efficiently generate proposal information regarding suitable welding robot systems adapted to the customer's welding work.

[0018] A program relating to one aspect of this disclosure comprises a case database configured by RAG (Retrieval-Augmented Generation) and including a vector index in which the case information is vectorized.

[0019] In this embodiment, the case database stored in the memory unit of the information processing device is configured using RAG (Retrieval-Augmented Generation). The source data of the case information stored in the case database is, for example, various sales information or technical information used in past sales activities or sales performance, and it is expected that the data format will be multimodal data consisting of different formats such as forms, documents, and images. In contrast, since the case database is configured using RAG, even if the stored case information is multimodal data consisting of different formats and is a group of data that is not normalized or structured (unstructured data), a vector index can be generated by inputting the unstructured data group into the embedding model provided by RAG, and search processing can be performed using this vector index. The vector index is a vector database that stores sentence embedding vectors, which are vectorized at the sentence level or the like in the case information. Searching the vector index (vector database) in the RAG may involve extracting information that has a similarity level of or higher by performing query processing using, for example, cosine similarity, Euclidean distance, or dot product similarity measure. By building a case database using RAG in this way, efficient search processing can be performed on case information consisting of unstructured data.

[0020] A program relating to one aspect of this disclosure includes the proposed information, which includes information regarding the cost estimate and equipment configuration of a welding robot system.

[0021] In this embodiment, the control unit of the information processing device generates proposal information based on derived case information, for example, using a large-scale language model (LLM), and generates the proposal information by including information regarding the estimate and equipment configuration of the welding robot system. In this case, the large-scale language model (LLM) may be implemented on an external server (LLM server) that is connected to the information processing device via an external network such as the Internet. The control unit of the information processing device outputs an instruction (prompt) to the external server (LLM server) to instruct it to generate proposal information, and obtains the proposal information which is the output result from the external server. When including information regarding the estimate in the proposal information, the control unit of the information processing device may include in the instruction (prompt) a statement to insert the estimate items and estimated amounts included in the derived case information into a predetermined estimate template. In this way, when submitting proposal information for a suitable welding robot system to a customer requesting the introduction or upgrade of a welding robot system, by including information regarding the estimate and equipment configuration in the proposal information, strong support can be provided to the personnel of the manufacturer or seller of the welding robot system.

[0022] An information processing method according to one aspect of the present disclosure involves causing a computer to perform the following processes: acquire customer work images taken of a welding work of a customer requesting a welding robot system; derive attribute information of the customer's welding work based on the acquired customer work images; derive similar work similar to the customer's welding work based on the derived attribute information; derive case information to which the similar work is applied based on the derived similar work; and generate proposal information regarding the welding robot system requested by the customer based on the derived case information.

[0023] In this embodiment, an information processing method can be provided that generates proposed information regarding a welding robot system requested by a customer based on acquired customer work images.

[0024] An information processing apparatus according to one aspect of the present disclosure includes an acquisition unit that acquires a customer work image obtained by imaging a welding work of a customer who desires a welding robot system, an attribute information derivation unit that derives attribute information in the welding work of the customer based on the acquired customer work image, a similar work derivation unit that derives a similar work similar to the welding work of the customer based on the derived attribute information, a case information derivation unit that derives case information to which the similar work has been applied based on the derived similar work, and a proposal information generation unit that performs a process for generating proposal information regarding the welding robot system desired by the customer based on the derived case information.

[0025] In this aspect, an information processing apparatus that generates proposal information regarding a welding robot system desired by a customer based on the acquired customer work image can be provided.

Effects of the Invention

[0026] A program or the like that generates proposal information regarding a welding robot system desired by a customer based on the acquired customer work image can be provided.

Brief Description of the Drawings

[0027] [Figure 1] It is a schematic diagram for explaining the outline of a welding robot system proposal support system according to Embodiment 1. [Figure 2] It is a block diagram showing the configuration of an information processing apparatus. [Figure 3] It is a functional block diagram exemplifying functional units included in the control unit of an information processing apparatus. [Figure 4] It is an explanatory diagram regarding the generation process of a welding work model. [Figure 5] It is an explanatory diagram exemplifying a work attribute information database. [Figure 6] It is a flowchart exemplifying a processing procedure by the control unit of an information processing apparatus.

Modes for Carrying Out the Invention

[0028] (Embodiment 1) The embodiments will be described below with reference to the drawings. Figure 1 is a schematic diagram illustrating the outline of the welding robot system proposal support system S according to Embodiment 1. Figure 2 is a block diagram showing the configuration of the information processing device 1. The welding robot system proposal support system S is configured with the information processing device 1 as the main device, and the information processing device 1 acquires images of the welding workpiece W (customer workpiece images) captured by the imaging device C. The images of the welding workpiece W (customer workpiece images) are captured, for example, by an information terminal T such as a smartphone used by a person in charge of a company that manufactures or sells welding robot systems at the customer's site. Furthermore, the information processing device 1 is connected to an external server G (LLM server) on which a large-scale language model (LLM) is implemented, for example, via an external network such as the Internet. The information processing device 1 outputs a prompt to the large-scale language model implemented on the external server G, and acquires generated data from the external server G according to the instructions contained in the prompt.

[0029] The imaging device C is, for example, a CMOS camera. The imaging device C may also be, for example, a camera built into an information terminal T such as a smartphone held by a sales representative of the welding robot system, or it may be connected via the input / output I / F 14 of the information processing device 1. The information terminal T of the sales representative of the welding robot system and the information processing device 1 are communicated with each other via an external network such as the Internet or a LAN. The image of the welding workpiece W (customer workpiece image) captured by the imaging device C is transmitted from the information terminal T of the sales representative, acquired via the input / output I / F 14 of the information processing device 1, or acquired offline via a storage medium such as a memory card, etc., by the information processing device 1.

[0030] The information processing device 1 is a computer capable of various information processing and information transmission / reception, such as a server device or a personal computer. The server device includes not only a single server device but also a cloud server device or a virtual server device composed of multiple computers. The information processing device 1 includes a control unit 11, a storage unit 12, a communication unit 13, and an input / output interface 14.

[0031] The control unit 11 has one or more arithmetic processing units equipped with timing functions such as a CPU (Central Processing Unit), MPU (Micro-Processing Unit), and GPU (Graphics Processing Unit), and performs various information processing, control processing, etc. by reading and executing the program P (program product) stored in the storage unit 12.

[0032] The storage unit 12 includes volatile storage areas such as SRAM (Static Random Access Memory), DRAM (Dynamic Random Access Memory), and flash memory, as well as non-volatile storage areas such as EEPROM or hard disk. The storage unit 12 pre-stores programs P (program products) and data referenced during processing, and also stores various data, including intermediate data, generated during processing. The programs P (program products) stored in the storage unit 12 may be programs P (program products) read from a recording medium M that the information processing device 1 can read. Alternatively, programs P (program products) may be downloaded from an external computer (not shown) connected to a communication network (not shown) and stored in the storage unit 12. As will be described in detail later, the storage unit 12 stores a work attribute information database, a case database, and actual files of welding work models 101.

[0033] The communication unit 13 is a communication module or communication interface for communicating with information terminals T etc. via wired or wireless methods such as Ethernet®, for example, a narrow-area wireless communication module such as Wi-Fi® or Bluetooth®, or a wide-area wireless communication module such as 4G or 5G. The control unit 11 communicates with information terminals T such as smartphones held by sales representatives etc., or external servers G (LLM servers) via the communication unit 13, for example, through an external network such as the Internet or a LAN.

[0034] Figure 3 is a functional block diagram illustrating the functional units included in the control unit 11 of the information processing device 1. The control unit 11 of the information processing device 1 functions as an acquisition unit 111, a similar work deriving unit 112, a case information deriving unit 113, a proposal information generation unit 114, and an output unit 115 by executing a program P stored in the storage unit 12. The control unit 11 of the information processing device 1 functions as a welding work model 101 (learning model) by reading the actual file of the welding work model 101 (learning model) stored in the storage unit 12.

[0035] The acquisition unit 111 acquires an image (customer work image) including the customer's welding workpiece W, captured by an imaging device C such as a CMOS camera, and further acquires information regarding the customer's specifications. The customer work image is, for example, an image of the welding workpiece W taken at the customer's site by a person in charge of a company that manufactures or sells welding robot systems, and may be captured by an information terminal T such as a smartphone used by that person in charge. The information regarding customer specifications may be data entered into the information terminal T by the person in charge based on the results of interviews with the customer. The information regarding customer specifications includes, for example, welding methods, welding conditions, quotation conditions, welding power supply, wire feeding, or required quality, that is, conditions for operating the welding robot system, specification limitations of equipment or facilities that constitute the welding robot system, or welding surface standards and yield information for the welding workpiece W, and corresponds to the customer's required specifications. The information processing device 1 and the information terminal T are communicated together, and the acquisition unit 111 acquires the customer work image and customer specifications from the information terminal T online or offline.

[0036] The acquisition unit 111 inputs a customer work image into the welding work model 101. Based on the input customer work image, the unit outputs attribute information of the customer's welding work W contained in the customer work image, and thus functions as an attribute information derivation unit.

[0037] Figure 4 is an explanatory diagram regarding the generation process of the welding work model 101. The control unit 11 of the information processing device 1 uses training data for the welding work model 101 to train a neural network such as R-CNN, and generates a welding work model 101 (learned model) that takes an image including the welding work W as input and outputs attribute information of the welding work W. The attribute information includes, for example, the work shape, size, welding location, welding method, welding conditions, welding position, and material of the welding work W.

[0038] The training data includes problem data and answer data. The image containing the welded workpiece W corresponds to the problem data, and the attribute information of the welded workpiece W (workpiece shape, size, etc.) corresponds to the answer data. The attribute information of the welded workpiece W, which is the answer data, may be set by annotating (adding) the workpiece shape, size, welding location, welding method, welding conditions, welding position, and material of the welded workpiece W to the image containing the welded workpiece W. The dataset of problem data and answer data included in the training data for training the welded workpiece model 101 (learning model) and the dataset of input data and output data when using the welded workpiece model 101 (learning model) are synonymous, and if defined in one dataset, it will naturally apply to the other dataset as well.

[0039] The neural network (welding work model 101) trained using training data is intended to be used as a program module, which is part of artificial intelligence software. The welding work model 101 (trained model) is used in an information processing device 1, which includes a control unit 11 (CPU, etc.) and a memory unit 12, as described above. By being executed in the information processing device 1, which has such computational processing capabilities, a neural network system is formed. Specifically, the control unit 11 of the information processing device 1 performs calculations to extract feature quantities from an image containing the welding work W input to the input layer, according to commands from the welding work model 101 (trained model) stored in the memory unit 12, and outputs attribute information of the welding work W (work shape, size, etc.) from the output layer.

[0040] The learning model is constructed using, for example, R-CNN (Region Convolutional Neural Network) or YOLO, and has an input layer that accepts an image containing the welding workpiece W as input, an intermediate layer that extracts features from the image containing the welding workpiece W, and an output layer that outputs attribute information of the welding workpiece W. The input layer has multiple neurons that accept an image containing the welding workpiece W as input, and passes the input values ​​to the intermediate layer. The intermediate layer is defined using an activation function such as a ReLU function or a sigmoid function, and has multiple neurons that extract features from each input value, and passes the extracted features to the output layer. Parameters such as the weighting coefficients and bias values ​​of the activation function are optimized using backpropagation. The output layer is constructed using, for example, a fully connected layer, and outputs attribute information of the welding workpiece W based on the features output from the intermediate layer.

[0041] In this embodiment, the welding work model 101 (learning model) is said to be R-CNN, etc., but is not limited to this, and may be a learning model constructed with other machine learning algorithms such as neural networks other than R-CNN, transformers, BERT, GPT, RNN (Recurrent Neural Network), LSTM (Long-short term model), SVM (Support Vector Machine), Bayesian networks, linear regression, regression trees, multiple regression, random forests, ensembles, etc. Alternatively, the welding work model 101 (learning model) may be constructed using an artificial intelligence chatbot such as ChatGPT. In this case, ChatGPT may be fine-tuned to efficiently output attribute information of the welding work W. Alternatively, a question sentence generated using an external database such as WebDB may be input to the prompt, which is the input interface of ChatGPT, along with an image containing the welding work W.

[0042] In this embodiment, to acquire attribute information of the customer's welded workpiece W, a customer workpiece image is input to the welded workpiece model 101, but this is not limited to this. The control unit 11 of the information processing device 1 may derive attribute information such as the workpiece shape and size of the welded workpiece W by performing image analysis processing on the customer workpiece image, for example, edge detection or pattern detection. In this case, the control unit 11 of the information processing device 1 functions as an attribute information deriving unit. Alternatively, the control unit 11 (attribute information deriving unit) of the information processing device 1 may derive attribute information such as the workpiece shape and size of the welded workpiece W based on drawing data such as 2D or 3D CAD showing the customer's welded workpiece W.

[0043] The similar workpiece deriving unit 112 acquires attribute information from the welding workpiece model 101 (attribute information deriving unit). The similar workpiece deriving unit 112 derives similar workpieces by performing a search operation using the derived attribute information against the stored workpiece attribute information database, which associates welding workpieces W with the attribute information of welding workpiece W. The workpiece attribute information database is stored in the storage unit 12 of the information processing device 1.

[0044] Figure 5 is an explanatory diagram illustrating a work attribute information database. The work attribute information database includes, for example, work image, work shape, size, welding location, welding method, welding conditions, welding position, and material as management items (fields). The management item for work image stores image data of the welded workpiece W as object data. Alternatively, the management item for work image may store a file path including the folder where the image data of the welded workpiece W is stored.

[0045] Workpiece shape, size, welding location, welding method, welding conditions, welding position, and material correspond to the respective attribute items in the attribute information. The workpiece shape management item stores data related to the type, classification, or part name of the welded workpiece W included in the corresponding workpiece image. The welding location management item stores the part name indicating the welding location in the welded workpiece W included in the corresponding workpiece image. The welding conditions management item stores data related to the welding conditions of the welded workpiece W included in the corresponding workpiece image. The welding position management item stores data related to the welding position of the welded workpiece W included in the corresponding workpiece image. The material management item stores the material name of the welded workpiece W included in the corresponding workpiece image.

[0046] The similar work deriving unit 112 is not limited to searching for a welded work W that exactly matches the attribute information output by the welded work model 101 when searching the work attribute information database. For example, the similar work deriving unit 112 may set a search range by specifying the similarity of the work shape in the attribute information, the range specification for the size, the degree of agreement for the welding conditions, etc., and search for a welded work W that has attribute information whose similarity to the attribute information of the customer's welded work W is greater than or equal to a predetermined value. The similar work deriving unit 112 may perform a search process within the search range that is within the predetermined similarity, and derive one or more welded work W extracted as search results as similar work.

[0047] The case information extraction unit 113 extracts case information related to a similar work by executing a search process (query) on the case database based on information about similar work extracted by the similar work extraction unit 112, namely the work image and attribute information of the similar work, and the customer specifications from the acquisition unit 111. The case database is stored in the storage unit 12 of the information processing device 1, and stores attribute information (specifications) of the welding work W and technical information or sales information such as equipment configuration, estimates, and delivery records related to the welding robot system corresponding to the welding work W, in correspondence or association. The case database is configured, for example, by RAG (Retrieval-Augmented Generation) and includes a vector index in which the case information is vectorized.

[0048] The source data for case information stored in the case database may be, for example, sales performance data stored in sales support systems such as SFA (Sales Force Automation) and CRM (Customer Relationship Management), or document data such as various forms, documents, and photographs stored on a shared file server used by sales representatives. The control unit 11 of the information processing device 1 generates a vector index (vector database) by inputting such a variety of document data, which are not normalized or structured (unstructured data), into the embedding model provided by RAG, using, for example, an LLM framework such as LangChain. The vector database has a data structure in which, for example, a vectorized sentence embedding vector is assigned as an index to each sentence unit (token) included in the case information. The case information derivation unit 113 performs query processing on the case database (vector database) configured in RAG using, for example, cosine similarity, Euclidean distance, or dot product similarity measure, to extract case information targeting welding work W that have a similarity of a predetermined degree or higher to similar work in accordance with customer specifications.

[0049] The proposal information generation unit 114 generates proposal information that reflects the content of the case information by inputting the case information from the case information derivation unit 113 into, for example, a large-scale language model (LLM). The large-scale language model (LLM) is implemented, for example, in an external server G (LLM server), which is a different device that becomes the information processing device 1.

[0050] The proposal information generation unit 114 may generate a prompt instructing the system to generate proposal information regarding the welding robot system, taking into account the case information obtained from the case information derivation unit 113, along with the contents of the case information, and output the generated prompt to the external server G. The prompt may define the person in charge of the manufacturer or seller of the welding robot system as an item indicating its role, and then instruct the system to generate proposal information regarding the welding robot system based on the case information obtained from the case information derivation unit 113. Furthermore, the prompt may include instructions that the proposal information should be inserted into a predetermined quotation template with the quotation items and quotation amounts included in the derived case information. Furthermore, the prompt may include instructions that the proposal information should be output in the format of a system proposal document that includes information on the equipment configuration. By setting the contents of the prompt in this way, the large-scale language model (LLM) implemented on the external server G (LLM server) outputs proposal information, including a quotation and a system proposal document in a predetermined format, according to the case information derived by the case information derivation unit 113 based on the customer work image and customer specifications. The proposal information generation unit 114 generates proposal information by acquiring proposal information from an external server G (LLM server).

[0051] In this embodiment, the proposed information generation unit 114 uses a large-scale language model (LLM) implemented on an external server G (LLM server), but is not limited to this. The large-scale language model (LLM) may also be implemented on the information processing device 1. In this case, the control unit 11 of the information processing device 1 may generate a large-scale language model (private LLM) by training a neural network composed of transformers, etc., with a corpus generated using a large amount of document data acquired from, for example, the Internet. The control unit 11 of the information processing device 1 may then perform instruction tuning or fine tuning on the private LLM thus generated using various information related to the welding robot system.

[0052] The output unit 115 outputs (transmits) the proposal information (quotation, equipment configuration specifications, system proposal) generated by the proposal information generation unit 114 to an information terminal T such as a smartphone used by a person in charge of a company that manufactures or sells welding robot systems, via the communication unit 13. By obtaining this proposal information (quotation, equipment configuration specifications, system proposal), the person in charge can efficiently make highly accurate proposals to customers who request welding robot systems. The output unit 115 may also store the proposal information output this time in the work attribute information database and the case database. In this way, by adding proposal information generated in individual sales activities to the work attribute information database and the case database, the amount of case information etc. stored in these databases can be efficiently increased.

[0053] Figure 6 is a flowchart illustrating the processing procedure performed by the control unit 11 of the information processing device 1. The control unit 11 of the information processing device 1 receives operator input, for example, from a keyboard connected to the input / output I / F 14, and performs the following processing based on the received operation.

[0054] The control unit 11 of the information processing device 1 acquires customer work images and customer specifications, which are images of the welding workpiece W of a customer requesting a welding robot system (S101). The control unit 11 of the information processing device 1 acquires images of the customer's welding workpiece W (customer work image) and customer specifications (welding method, welding conditions, quotation conditions, welding power supply, wire feeding or required quality, etc., required specifications) from an information terminal T such as a smartphone used by a person in charge of a company that manufactures or sells welding robot systems, and stores them in the storage unit 12.

[0055] The control unit 11 of the information processing device 1 derives attribute information of the customer's welded work W based on the acquired customer work image (S102). The control unit 11 of the information processing device 1 inputs the acquired customer work image into the welded work model 101. The welded work model 101 outputs attribute information (work shape, size, welding location, welding method, welding conditions, welding position, and material, etc.) of the customer's welded work W contained in the customer work image based on the input customer work image. The control unit 11 of the information processing device 1 derives the attribute information by acquiring the attribute information from the welded work model 101.

[0056] The control unit 11 of the information processing device 1 derives similar workpieces based on the derived attribute information (S103). The control unit 11 of the information processing device 1 derives similar workpieces by performing a search process on the workpiece attribute information database based on the derived attribute information. The workpiece attribute information database stores attribute items such as workpiece shape, size, welding location, welding method, welding conditions, welding position, and material for each welded workpiece W, with these attribute items associated with the welded workpiece W. The control unit 11 of the information processing device 1 sets a search condition with a predetermined search range for the value or content of each attribute item included in the attribute information, and performs a search process using the search condition to derive welded workpieces W that have similar attribute information to the derived attribute information (attribute information of the customer's welded workpiece W) as similar workpieces.

[0057] The control unit 11 of the information processing device 1 derives case information based on the derived similar workpiece (S104). The control unit 11 of the information processing device 1 derives case information (various information generated in sales cases) to which the similar workpiece was applied or targeted by performing a search process on a case database, for example, configured in RAG, based on the derived information about the similar workpiece. This case information includes technical information or sales information such as equipment configuration, estimates, and delivery records related to welding robot systems.

[0058] The control unit 11 of the information processing device 1 generates proposal information based on the derived case information (S105). The control unit 11 of the information processing device 1 inputs a prompt generated by taking the derived case information into account to a large-scale language model (LLM) implemented on an external server G, for example. The large-scale language model (LLM) outputs case information including an estimate, equipment configuration specifications, and a system proposal in response to the prompt. The control unit 11 of the information processing device 1 derives the case information by acquiring the case information from the large-scale language model (LLM) implemented on the external server G.

[0059] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims, not in the sense described above, and all modifications within the sense and scope equivalent to the claims are intended.

[0060] Regarding the multiple claims described in the patent claims, they can be combined with each other regardless of the form of citation. The patent claims include multiple dependent claims that depend on multiple claims. The patent claims do not include multiple dependent claims that depend on multiple dependent claims, but multiple dependent claims that depend on multiple dependent claims may be included. [Explanation of Symbols]

[0061] S Welding robot system proposal support system, W Welding workpiece, T Information terminal, C Imaging device, G External server (LLM server) 1 Information processing device, 11 Control unit, 111 Acquisition unit, 112 Similar workpiece extraction unit, 113 Case information extraction unit, 114 Proposal information generation unit, 115 Output unit, 12 Storage unit, P Program (program product), M Recording medium, 13 Communication unit, 14 Input / Output I / F, 101 Welding workpiece model (attribute information extraction unit)

Claims

1. On the computer, We acquire customer work images, which are images of the welding workpieces of customers who request a welding robot system. Based on the acquired customer work image, attribute information of the customer's welded work is derived, Based on the attribute information derived, similar workpieces similar to the customer's welded workpiece are derived, Based on the derived similar work, the case information to which the similar work is applied is derived, Based on the derived project information, proposal information regarding the welding robot system requested by the customer is generated. A program that executes a process.

2. By inputting the acquired customer work image into a welding work model that has been trained to output attribute information of a welding work when an image containing a welding work is input, the attribute information of the welding work included in the customer work image is derived. The program according to claim 1.

3. The attribute information includes at least one piece of information from the workpiece shape, size, and welding location information of the welded workpiece. By performing a search on a work attribute information database, in which multiple welded workpieces and the attribute information of each of the multiple welded workpieces are stored in association with each other, similar workpieces are derived using the derived attribute information. The program according to claim 1.

4. The aforementioned project information includes at least one of the following: attribute information of the welding workpiece, equipment configuration of the welding robot system, and information regarding quotations and delivery records. The aforementioned case information is derived by performing a search operation on the case database in which the aforementioned case information is stored, using the derived similar work. The program according to claim 1.

5. Obtain the customer specifications requested by the aforementioned customer, Based on the acquired customer specifications and the derived similar workpieces, the project information is derived. The program according to claim 1.

6. The aforementioned customer specifications include at least one of the following specification items: welding method, welding conditions, quotation conditions, welding power source, and wire feeding. The aforementioned project information includes information corresponding to the specification items in the customer specifications, The aforementioned case information is derived by performing a search on the case database in which the aforementioned case information is stored, using the derived similar work and the acquired customer specifications. The program according to claim 5.

7. The aforementioned case database is constructed using RAG (Retrieval-Augmented Generation) and includes a vector index in which the case information is vectorized. The program according to claim 6.

8. The proposed information includes information regarding the estimate and equipment configuration of the welding robot system. The program according to any one of claims 1 to 7.

9. On the computer, We acquire customer work images, which are images of the welding workpieces of customers who request a welding robot system. Based on the acquired customer work image, attribute information of the customer's welded work is derived, Based on the attribute information derived, similar workpieces similar to the customer's welded workpiece are derived, Based on the derived similar work, the case information to which the similar work is applied is derived, Based on the derived project information, proposal information regarding the welding robot system requested by the customer is generated. An information processing method that involves having a computer perform a task.

10. An acquisition unit that acquires customer work images, which are images of the welding workpieces of customers requesting a welding robot system, An attribute information derivation unit that derives attribute information of the customer's welded work based on the acquired customer work image, A similar workpiece deriving unit derives similar workpieces that are similar to the customer's welding workpiece based on the derived attribute information, A case information derivation unit derives case information to which the similar work is applied based on the derived similar work, Based on the derived project information, a proposal information generation unit performs processing to generate proposal information regarding the welding robot system requested by the customer. An information processing device equipped with the following features.

Images