Computer program, information processing method, and information processing device

Abstract

The purpose of the present invention is to provide a computer program, an information processing method, and an information processing device that can be expected to support a user who uses a substrate processing device.　A computer program according to the present embodiment causes a computer to execute processing of: acquiring trouble information related to trouble of a substrate processing device and log information output by the substrate processing device that has performed substrate processing; generating countermeasure information related to countermeasures against the trouble by a language model using a knowledge database that stores information related to the substrate processing device on the basis of the acquired trouble information and log information; and outputting the generated countermeasure information.

Description

Computer Program, Information Processing Method, and Information Processing Apparatus

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

[0002] In Patent Document 1, a management system for a semiconductor manufacturing apparatus aiming to automatically eliminate troubles has been proposed. This management system identifies a trouble device that causes defects in semiconductor integrated circuits from a plurality of semiconductor manufacturing apparatuses, performs principal component analysis on the EES data of the trouble device, obtains a principal component space that distinguishes between the normal state and the trouble state, and determines a linear state equation that defines the time evolution of the internal state of the trouble device represented by the principal components. The management system examines its controllability based on the linear state equation. When it has controllability, after acquiring the EES data of the trouble device, it calculates the internal state as a principal component vector and determines whether this internal state belongs to a region representing the normal state within the principal component space. When deviating from the region representing the normal state, the management system controls the input based on the linear state equation so that the internal state returns to the normal state.

[0003] Japanese Unexamined Patent Application Publication No. 2011 - 54804

[0004] The present disclosure provides a computer program, an information processing method, and an information processing apparatus that can be expected to assist a user who uses a substrate processing apparatus.

[0005] A computer program according to an embodiment causes a computer to execute a process of acquiring trouble information regarding troubles of a substrate processing apparatus and log information output by the substrate processing apparatus that has performed substrate processing, generating countermeasure information regarding countermeasures to the troubles using a language model that uses a knowledge database storing information regarding the substrate processing apparatus based on the acquired trouble information and log information, and outputting the generated countermeasure information.

[0006] According to the present disclosure, it can be expected to assist a user who uses a substrate processing apparatus.

[0007] This is a schematic diagram illustrating the overview of the troubleshooting solution suggestion function of the information processing system according to this embodiment. This is a schematic diagram illustrating the overview of the operation support function of the information processing system according to this embodiment. This is a block diagram illustrating an example configuration of the information processing device according to this embodiment. This is a flowchart illustrating an example of the procedure for troubleshooting solution generation processing performed by the information processing device according to this embodiment. This is a flowchart illustrating an example of the procedure for troubleshooting solution generation processing performed by the information processing device according to this embodiment. This is a schematic diagram illustrating an example of a prompt that instructs the language model to generate troubleshooting solutions. This is a schematic diagram showing an example of a display of troubleshooting solutions. This is a schematic diagram illustrating an example of a prompt that instructs the language model to rank troubleshooting solutions. This is a schematic diagram showing an example of a display of ranked troubleshooting solutions. This is a schematic diagram illustrating an example of a report input screen. This is a schematic diagram showing an example of a report input screen. This is a flowchart illustrating an example of the procedure for recipe or cost function generation processing performed by the information processing device according to this embodiment. This is a schematic diagram showing an example of a display of a recipe generated by the information processing device.

[0008] Specific examples of information processing systems according to the embodiments of this disclosure will be described below with reference to the drawings. However, this disclosure is not limited to these examples and is intended to include all changes within the meaning and scope of the claims as indicated by the claims.

[0009] <System Overview> Figure 1 is a schematic diagram illustrating the overview of the troubleshooting solution suggestion function of the information processing system according to this embodiment. The information processing system according to this embodiment is configured to include an information processing device 1 and a substrate processing device 3, etc. The substrate processing device 3 is a device such as a process chamber that performs substrate processing such as etching on a substrate such as a semiconductor wafer, and performs substrate processing based on setting information, so-called recipes, for performing substrate processing that have been prepared in advance. The recipe for substrate processing is created by a user using the information processing device 1, for example, and is provided from the information processing device 1 to the substrate processing device 3.

[0010] In this embodiment, the setting information provided from the information processing device 1 to the substrate processing device 3 is referred to as a recipe, but it is not limited to this. For example, a cost function may be provided from the information processing device 1 to the substrate processing device 3 as setting information. The cost function is a function that represents the conditions that should be considered when the substrate processing device 3 operates, and it is a function that calculates the cost value by inputting a numerical value based on the result of the substrate processing or its predicted value. A recipe is a fixed parameter provided to the substrate processing device 3 in order to perform substrate processing, but a cost function is setting information that can be varied according to the quality of the result. For example, if the state of the chamber of the substrate processing device 3 changes over time, the substrate processing device 3 is controlled by maintaining the same quality of result by following the change over time using the cost function. Conventionally, it is common for the user to provide the substrate processing device 3 with fixed recipe parameters, and the substrate processing is controlled based on these recipe parameters. However, the substrate processing device 3 can control the substrate processing using variable parameters, such as a cost function, which adjusts the substrate processing itself based on the quality or finish of the product, rather than fixed parameters. For example, when the recipe setting is x0, a cost function that returns (x - x0)² may be adopted. Furthermore, the cost function may be defined using numerical values ​​related to the quality of the finished product or productivity that do not directly correspond to the recipe setting, and the substrate processing device 3 can control each parameter to realize this.

[0011] The substrate processing device 3 performs substrate processing based on a recipe (or cost function) provided by the information processing device 1. The substrate processing device 3 is equipped with one or more sensors (not shown) that measure, for example, the temperature or pressure inside the process chamber, and outputs log information to the information processing device 1, including time-series data of measurements taken by the sensors during the substrate processing process. Furthermore, if a problem or abnormality occurs during the substrate processing process, the substrate processing device 3 outputs trouble information related to the trouble to the information processing device 1. Note that the trouble information may be included in the log information.

[0012] In this embodiment, the information processing device 1 directly acquires trouble information and log information from the substrate processing device 3, but it is not limited to this. The information processing device 1 may indirectly acquire trouble information and log information, for example, through another device or system that manages the substrate processing device 3.

[0013] The information processing device 1 may also acquire trouble information by receiving input from users, such as operators, who operate the circuit board processing device 3.

[0014] The information processing device 1 creates a recipe (or cost function) based on user input, for example, and performs the desired substrate processing by providing the created recipe to the substrate processing device 3. The information processing device 1 also displays and provides to the user information such as the results of the substrate processing performed by the substrate processing device 3, and stores and accumulates the recipe provided to the substrate processing device 3 and log information obtained from the substrate processing device 3 in a database in association with each other.

[0015] The information processing device 1 according to this embodiment has a function that, when a problem occurs during substrate processing performed by the substrate processing device 3, proposes countermeasures to the user based on log information and problem information obtained from the substrate processing device 3. In order to realize this countermeasure proposal function, the information processing device 1 according to this embodiment has a knowledge DB (database) 5 that stores and accumulates information such as knowledge, experience, and know-how related to troubleshooting to date, and a language model 7 that has been generated in advance by machine learning.

[0016] Knowledge DB 5 stores information such as trouble information and log information obtained from the board processing device 3, countermeasures taken for this trouble, and the results of these countermeasures (whether the trouble was resolved or not, etc.) in an associated manner. Furthermore, Knowledge DB 5 may store information obtained not only from one board processing device 3, but also from, for example, another board processing device 3 of the same model, or information about multiple similar board processing devices 3, or information about devices other than board processing devices 3.

[0017] The language model 7 is a learning model known as an LLM (Large Language Model). The language model 7 may employ learning models such as a Transformer, BERT (Bidirectional Encoder Representations from Transformers), or GPT (Generative Pre-trained Transformer) equipped with an attention mechanism in a large-scale neural network. The language model 7 used by the information processing device 1 may be a widely available general-purpose language model, or it may be a general-purpose language model that has been fine-tuned using information related to substrate processing, or it may be a model in which the parameters of some layers have been fine-tuned using techniques such as LoRA (Low Rank Adaptation). The language model 7 used by the information processing device 1 may also be a language model specialized for the semiconductor field or a model smaller than an LLM. In this embodiment, the language model 7 can, for example, input and output information in natural language with the user, and can also, for example, extract necessary information from the knowledge DB 5 according to given conditions.

[0018] The information processing device 1 extracts information related to the trouble that occurred in the circuit board processing device 3 from the knowledge DB 5 using the language model 7, based on trouble information and log information obtained from the circuit board processing device 3, as well as requests and trouble information provided by the user in natural language. Based on the extracted information, it generates information related to countermeasures for the trouble. The information processing device 1 generates multiple countermeasures for the trouble using the language model 7, ranks the multiple countermeasures as candidates, and provides them to the user. At this time, the information processing device 1 also accepts ranking conditions from the user in natural language, and by providing the received natural language ranking conditions to the language model 7, it can cause the language model 7 to rank the multiple candidate trouble countermeasures.

[0019] The user, presented with multiple ranked countermeasures by the information processing device 1, selects one of the countermeasures and instructs the information processing device 1 to execute the selected countermeasure. The information processing device 1 stores, for example, commands, programs, or recipes for processing according to the countermeasure in a database, and executes the countermeasure by retrieving the commands, etc. for implementing the user's selected countermeasure from the database and providing them to the board processing device 3. The information processing device 1 may also execute the countermeasure by generating a program or recipe. Furthermore, the information processing device 1 may not be limited to the board processing device 3, but may instruct another device, such as a maintenance robot, to execute the countermeasure as needed.

[0020] The information processing device 1 may accept the user's selection of multiple countermeasures from among the multiple countermeasures presented, and may also accept countermeasures from the user that are different from those presented. Furthermore, in the case of countermeasures that the information processing device 1 cannot implement, such as replacing or repairing parts of the circuit board processing device 3, the user may implement the countermeasures such as replacing parts, and the user will input information such as the content and results of the implemented countermeasures into the information processing device 1.

[0021] The information processing device 1, which has implemented countermeasures for a problem, acquires log information output by the board processing device 3 as a result of the implementation of the countermeasures, or information input by the user who judged whether or not the countermeasures were feasible, as information related to the results of the countermeasures. The information processing device 1 generates a report using the language model 7 that summarizes the trouble information and log information obtained from the board processing device 3 or the user when a problem occurs, the content of the countermeasures implemented for this problem, and the results of the implemented countermeasures, and displays the generated report to the user. The information processing device 1 also stores the trouble information, log information, content of the countermeasures, results of the countermeasures, and reports in the knowledge DB 5, associating the information, data related to trouble response, and other information, and stores it in the knowledge DB 5.

[0022] Figure 2 is a schematic diagram illustrating the overview of the operation support function of the information processing system according to this embodiment. As described above, in order to perform the desired substrate processing in the substrate processing device 3, the user needs to create a recipe (or cost function) in the information processing device 1 in which the conditions for substrate processing are set in advance. The information processing device 1 according to this embodiment uses the language model 7 to receive instructions for operations related to substrate processing from the user in natural language, and generates a recipe (or cost function) based on the received instructions, thereby supporting the user's operation of the substrate processing device 3.

[0023] For example, when a user needs to perform some kind of substrate processing on the substrate processing apparatus 3 as a countermeasure to the trouble shown in Figure 1, they can use the operation support function to generate a recipe (or cost function). Furthermore, the user can use the operation support function for various purposes unrelated to troubleshooting, such as generating recipes for semiconductor product manufacturing or generating recipes for experimental substrate processing for data collection.

[0024] In the operation support function, the knowledge DB 5 of the information processing device 1 stores, for example, recipes (or cost functions) for board processing previously performed by the board processing device 3, and the results of the board processing (log information, etc.) in association with each other. The information processing device 1 according to this embodiment uses a language model 7 to decompose the board processing procedure to be performed by the board processing device 3, the procedure for creating a board processing recipe, or the calculation of a cost function for each board processing procedure into multiple subtasks based on the operation instructions in natural language input by the user. The information processing device 1 uses the language model 7 to extract recipes corresponding to each subtask from the knowledge DB 5, and also extracts information such as constraints that generally need to be imposed on the board processing device 3 from the knowledge DB 5. The information processing device 1 executes each subtask in order, taking the constraints into consideration, and generates a recipe corresponding to the user's operation instructions.

[0025] Furthermore, the information processing device 1 according to this embodiment includes a simulator that simulates the substrate processing performed by the substrate processing device 3. The information processing device 1 predicts the results of the substrate processing according to the generated recipe by simulating the substrate processing of the substrate processing device 3 using the simulator based on the generated recipe (or cost function). At this time, the information processing device 1 determines whether the substrate processing according to the generated recipe (or cost function) satisfies the constraints and how much cost is required, etc., based on the above-mentioned constraints and cost function, and thereby determines whether there is a risk in the substrate processing according to the generated recipe (or cost function).

[0026] The information processing device 1 displays information such as the results of the substrate processing obtained from the simulation and the presence or absence of risks, and provides this information to the user, and accepts the user's choice of whether or not to perform the substrate processing according to the generated recipe (or cost function). If the user chooses to perform the processing, the information processing device 1 transmits the generated recipe to the substrate processing device 3 and commands the substrate processing device 3 to perform the substrate processing based on this recipe, thereby causing the substrate processing device 3 to perform the desired substrate processing. If the user chooses not to perform the processing, the information processing device 1 accepts additional conditions regarding the substrate processing to be performed from the user via natural language input, for example, and regenerates the recipe (or cost function).

[0027] As described above, in the information processing system according to this embodiment, the information processing device 1 uses the knowledge DB 5 and the language model 7 to propose countermeasures for problems and generate board processing recipes (or cost functions), thereby enabling the exchange of information with the user via natural language, and is expected to support the user using the board processing device. In this embodiment, the information processing device 1 is configured to perform both proposing countermeasures for problems and generating board processing recipes (or cost functions), but it is not limited to this configuration, and the information processing device 1 may be configured to perform only one of either proposing countermeasures for problems or generating board processing recipes (or cost functions).

[0028] <Device Configuration> Figure 3 is a block diagram showing an example configuration of the information processing device 1 according to this embodiment. The information processing device 1 according to this embodiment can be realized by installing a predetermined application program on a general-purpose information processing device such as a personal computer or a server computer. However, the information processing device 1 may be a dedicated information processing device that controls the board processing device 3. The information processing device 1 according to this embodiment is configured to include a processing unit 11, a storage unit 12, a communication unit 13, a display unit 14, and an operation unit 15, etc. In this embodiment, the explanation will be given assuming that processing is performed by one information processing device 1, but the processing of the information processing device 1 may be distributed among multiple devices.

[0029] The processing unit 11 is composed of a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), a GPU (Graphics Processing Unit), or a quantum processor, as well as ROM (Read Only Memory) and RAM (Random Access Memory). The processing unit 11 reads and executes a program 12a stored in the memory unit 12, performing various processes such as generating countermeasures for problems that occur in the substrate processing device 3, and generating a recipe (or cost function) for substrate processing to be performed by the substrate processing device 3.

[0030] The storage unit 12 is configured using a large-capacity storage device such as a hard disk or an SSD (Solid State Drive). The storage unit 12 stores various programs executed by the processing unit 11, and various data necessary for the processing of the processing unit 11. In this embodiment, the storage unit 12 stores the program 12a executed by the processing unit 11. The storage unit 12 is also provided with a learning model storage unit 12b that stores information about learned learning models used in the processing performed by the information processing device 1. The storage unit 12 is also provided with a knowledge database 5 that stores information such as knowledge, experience, and know-how regarding troubleshooting the substrate processing device 3 and recipes for substrate processing performed by the substrate processing device 3.

[0031] In this embodiment, the program (computer program, program product) 12a is provided in a form recorded on a recording medium 99 such as a memory card or optical disc, and the information processing device 1 reads the program 12a from the recording medium 99 and stores it in the storage unit 12. However, the program 12a may also be written to the storage unit 12 during the manufacturing stage of the information processing device 1, for example. Alternatively, the program 12a may be distributed by a remote server device or the like and acquired by the information processing device 1 via communication. For example, the program 12a may be read from the recording medium 99 by a writing device and written to the storage unit 12 of the information processing device 1. The program 12a may be provided by distribution via a network, or it may be provided in a form recorded on the recording medium 99.

[0032] The learning model storage unit 12b stores information about a learning model that has been pre-machine-learned. The information about the learning model may include, for example, information indicating the configuration of the learning model and information such as the values ​​of internal parameters determined by machine learning. In this embodiment, the learning model storage unit 12b stores information about a language model 7 for inputting and outputting information in natural language. In addition to the language model 7, the learning model storage unit 12b may also store information about various other learning models, such as an image generation model, a classification model, or a regression model. In this embodiment, the information processing device 1 stores information about these learning models and performs processing such as prediction using these learning models, but it is not limited to this. The information about these learning models may be stored in a device other than the information processing device 1, and this device may perform processing such as prediction using the learning models, and the information processing device 1 may obtain the processing results from this device. Furthermore, the machine learning processing of these learning models may be performed by the information processing device 1 or by a device other than the information processing device 1.

[0033] Knowledge DB 5 is a database that stores information such as technical knowledge, experience, and know-how regarding the PCB processing device 3 and other devices of the same model or similar models. Knowledge DB 5 stores trouble case information, for example, by associating information such as troubles that have occurred in the PCB processing device 3 in the past, log information obtained from the PCB processing device 3 when a trouble occurred, countermeasures taken in response to the trouble, and the results of these countermeasures. Knowledge DB 5 also stores PCB processing case information, for example, by associating information such as recipes (or cost functions) for past PCB processing performed by the PCB processing device 3, various measurement values ​​measured by sensors during the PCB processing, and the results of the PCB processing. Knowledge DB 5 may also store manual information such as the operation manual or maintenance manual for the PCB processing device 3. Knowledge DB 5 may also store historical information such as inquiries and answers exchanged between users of the PCB processing device 3 and engineers of the company that manufactures or sells the PCB processing device 3. Knowledge DB 5 may store any other information not mentioned above.

[0034] The Knowledge DB 5 according to this embodiment stores the above-mentioned various information in a format that the language model 7 can handle, such as text information. The Knowledge DB 5 may store the text information of the above-mentioned various information together with information such as the date and time the text information was acquired and the feature quantities of the text information. In addition to text information, the Knowledge DB 5 may also store information such as still images, moving images, or audio.

[0035] In this embodiment, the information processing device 1 is configured to include a knowledge database 5, but it is not limited to this configuration. The knowledge database 5 may be provided by a device other than the information processing device 1, in which case the information processing device 1 communicates with the device providing the knowledge database 5 and performs operations such as reading information from the knowledge database 5 and writing information to the knowledge database 5.

[0036] The communication unit 13 transmits and receives data to and from the board processing device 3 via a wired or wireless network. In this embodiment, the information processing device 1 can obtain data such as log information and trouble information from the board processing device 3 by having the communication unit 13 communicate with the board processing device 3, and can provide the board processing device 3 with conditions and commands for performing board processing. In this embodiment, data is exchanged between the board processing device 3 and the information processing device 1 via communication, but this is not limited to this, and data may be exchanged via a recording medium such as a memory card. Furthermore, the communication unit 13 may communicate with a device other than the board processing device 3.

[0037] The display unit 14 is configured using a liquid crystal display or the like, and displays various images and characters based on the processing of the processing unit 11. The operation unit 15 receives user operations and notifies the processing unit 11 of the received operations. For example, the operation unit 15 receives user operations via an input device such as a mechanical button or a touch panel provided on the surface of the display unit 14. Alternatively, the operation unit 15 may be an input device such as a mouse and a keyboard, and these input devices may be configured to be detachable from the information processing device 1.

[0038] The memory unit 12 may be an external storage device connected to the information processing device 1. The information processing device 1 may be a multicomputer comprising multiple computers, or it may be a virtual machine virtually constructed by software. Furthermore, the information processing device 1 is not limited to the above configuration, and may not include, for example, a display unit 14 and an operation unit 15.

[0039] Furthermore, in the information processing device 1 according to this embodiment, the processing unit 11 reads and executes the program 12a stored in the storage unit 12, thereby realizing the information collection unit 11a, the countermeasure generation unit 11b, the recipe generation unit 11c, the simulation processing unit 11d, the control processing unit 11e, and the display processing unit 11f, etc., as software-based functional units in the processing unit 11. In this figure, the functional units of the processing unit 11 that relate to troubleshooting and recipe generation of the substrate processing device 3 are shown, and functional units related to other processing are omitted from the illustration.

[0040] The information gathering unit 11a performs processing to collect various information to be stored in the knowledge database 5. For example, by communicating with the substrate processing device 3 via the communication unit 13, the information gathering unit 11a acquires various information such as measurement values ​​from sensors, information on the results of substrate processing, log information and trouble information associated with substrate processing, and stores it in the knowledge database 5. Also, for example, the information gathering unit 11a acquires information on recipes (or cost functions) created by the user by setting the conditions for substrate processing in order to perform substrate processing on the substrate processing device 3, and stores it in the knowledge database 5.

[0041] Furthermore, the information gathering unit 11a displays on the display unit 14, for example, a message prompting the user to input information and a screen for inputting information, for information that needs to be stored in the knowledge DB 5 from the user, and acquires the information entered by the user via the operation unit 15 and stores it in the knowledge DB 5. In this embodiment, for example, if the user replaces a part in response to a problem with the substrate processing device 3, the information gathering unit 11a asks the user for information about the replaced part, information about the location where the part was replaced, and whether or not the problem was resolved by replacing the part, and stores the information entered by the user in the knowledge DB 5 as a report on troubleshooting.

[0042] The countermeasure generation unit 11b uses the language model 7 stored in the learning model storage unit 12b to generate countermeasures for problems that occur in the substrate processing device 3 and present them to the user. When a problem occurs, the countermeasure generation unit 11b analyzes the numerical values ​​of the log information according to the nature of the problem, and extracts similar information from the knowledge DB 5, based on the log information and problem information acquired from the substrate processing device 3 and the problem information from the user. Based on these analysis and extraction results, the log information and problem information acquired from the substrate processing device 3, and the problem information from the user, the countermeasure generation unit 11b generates a prompt for the language model 7 to generate countermeasures for the problem. The countermeasure generation unit 11b inputs the generated prompt to the language model 7 and acquires the information output by the language model 7 in response, thereby generating information related to countermeasures for the problem. The countermeasure generation unit 11b displays the information related to the countermeasures generated by the language model 7 on the display unit 14.

[0043] In this embodiment, the countermeasure generation unit 11b generates multiple pieces of information regarding countermeasures for problems that occur in the substrate processing device 3, and ranks and displays the generated pieces of information regarding countermeasures according to predetermined criteria. In this embodiment, the user can consider whether the ranked pieces of countermeasures displayed are appropriate, input ranking conditions in natural language, and have the information processing device 1 redisplay the countermeasures. The countermeasure generation unit 11b generates a prompt based on the ranking conditions in natural language input by the user, inputs the generated prompt to the language model 7, retrieves the information regarding countermeasures output by the language model 7, and redisplays it. However, the input of ranking conditions by the user is not limited to natural language; various methods may be used, such as selection from multiple presented candidates. When the input of conditions is accepted by selection of candidates, the countermeasure generation unit 11b can be expected to facilitate the user's selection of candidates by, for example, hierarchically presenting questions to the user and candidate answers to these questions.

[0044] Based on the information about the countermeasures for the troubles presented by the information processing apparatus 1, the user determines the countermeasures to be implemented. The user selects any one of the countermeasures displayed in a ranked manner, and causes the information processing apparatus 1 to implement the selected countermeasure, or the user implements the countermeasure. For example, when the countermeasure is to cause the substrate processing apparatus 3 to perform substrate processing under a predetermined condition, the information processing apparatus 1 gives the predetermined condition and the command to perform the substrate processing to the substrate processing apparatus 3, thereby causing the substrate processing apparatus 3 to perform the substrate processing as a countermeasure for the trouble. The information processing apparatus 1 acquires information regarding the result of the substrate processing performed as a countermeasure from the substrate processing apparatus 3, and stores information such as the content of the performed substrate processing and the acquired result of the substrate processing in the knowledge DB 5.

[0045] Also, for example, when the countermeasure is not something that the information processing apparatus 1 can perform, such as component replacement or maintenance, the user or an engineer of the substrate processing apparatus 3 performs component replacement or maintenance, etc., and inputs information such as the content of the performed component replacement or maintenance, etc., and the result of the component replacement or maintenance, etc., to the information processing apparatus 1. The information processing apparatus 1 stores the input information.

[0046] The recipe generation unit 11c uses the language model 7 stored in the learning model storage unit 12b to perform a process of generating a recipe for causing the substrate processing apparatus 3 to perform substrate processing or a cost function in the substrate processing based on the operation instruction of the substrate processing input by the user in natural language. However, the input of the operation instruction by the user is not limited to being in natural language, and various methods such as selection from a plurality of presented candidates or input of a numerical value may be adopted.

[0047] For example, the recipe generation unit 11c uses the language model 7 to decompose the operation instruction input by the user into a plurality of subtasks, and generates code of a program for generating a recipe or a cost function for each subtask. By executing the code generated by the language model 7, the recipe generation unit 11c can generate a recipe or a cost function. At this time, the recipe generation unit 11c also uses the language model 7 to generate constraint conditions and the like related to the substrate processing of the substrate processing apparatus 3. Note that the language model 7 may appropriately extract a constraint condition or the like from the constraint conditions and the like stored in the knowledge DB 5 instead of generating them. The constraint condition is a condition that prohibits performing substrate processing, such as an upper limit value and a lower limit value of a set value related to substrate processing. The cost function is a function that calculates the cost value by inputting a numerical value based on the result of substrate processing.

[0048] The simulation processing unit 11d uses a simulator that reproduces the operation of the substrate processing apparatus 3 prepared in advance to simulate the substrate processing by the recipe or the cost function generated by the recipe generation unit 11c, and performs processing to analyze the risk of the recipe or the cost function based on the simulation result. The simulation processing unit 11d determines whether or not there is a risk of substrate processing based on the recipe or the cost function by determining, for example, whether the set value included in the recipe satisfies the constraint condition, or whether the simulation result of the substrate processing based on the recipe or the cost function satisfies the constraint condition, based on the constraint condition generated by the recipe generation unit 11c. The simulation processing unit 11d displays the simulation result and the determination result of the presence or absence of risk based thereon on the display unit 14.

[0049] Based on the displayed simulation results and risk assessment results, the user decides whether or not to perform the board processing based on the generated recipe or cost function. When the information processing device 1 receives an instruction from the user to perform the board processing based on the recipe or cost function, it transmits the recipe or cost function to the board processing device 3 and instructs it to perform the board processing, thereby performing the board processing based on the recipe or cost function. If the user decides not to perform the board processing based on the recipe or cost function, the user can instruct the information processing device 1 to modify the recipe or cost function. The information processing device 1 accepts input such as modification conditions in natural language, for example, and regenerates the recipe or cost function using the language model 7.

[0050] The control processing unit 11e controls the operation of the board processing device 3 by outputting control commands. The control processing unit 11e sends a recipe or cost function generated using the language model 7, or a recipe or cost function generated in advance by the user, to the board processing device 3, and gives the board processing device 3 an instruction to perform board processing based on this recipe or cost function, thereby causing the board processing device 3 to perform board processing based on the recipe or cost function. The control processing unit 11e can also cause the board processing device 3 to perform various operations, such as performing predetermined board processing as a countermeasure against trouble, performing calibration of each part of the board processing device 3, or performing a self-check to detect fault locations of the board processing device 3.

[0051] The display processing unit 11f performs the process of displaying various characters and images on the display unit 14. For example, the display processing unit 11f displays various information collected by the information collection unit 11a or information stored in the knowledge DB 5. In this case, the display processing unit 11f may display numerical information using graphs or tables. In this embodiment, the display processing unit 11f also displays information related to multiple countermeasures generated by the countermeasure generation unit 11b in a ranked order. The display processing unit 11f also displays information such as recipes or cost functions generated by the recipe generation unit 11c, and verification results by the simulation processing unit 11d. The display processing unit 11f may also display a screen, such as a chat screen, for receiving generation conditions from the user in natural language during the generation of countermeasures and recipes. The display processing unit 11f may also display various other information on the display unit 14.

[0052] Furthermore, the input and output of information between the information processing device 1 and the user in natural language may be performed by voice rather than by text. In this case, the information processing device 1 may acquire the user's voice message via a device such as a microphone and output a voice message based on the output of a large-scale language model via a device such as a speaker. The information processing device 1 may also perform mutual conversion between voice data and text data. In addition, the information processing device 1 may, for example, receive a message from the user by voice and display the message to the user in text, or it may, for example, receive a message from the user in text and output the message to the user by voice.

[0053] <Troubleshooting Measure Generation Process> Figures 4 and 5 are flowcharts showing an example of the procedure for troubleshooting measure generation processing performed by the information processing device 1 according to this embodiment. The information collection unit 11a of the processing unit 11 of the information processing device 1 according to this embodiment acquires information related to the trouble when some kind of trouble occurs in the substrate processing device 3 (step S1). The information acquired by the information collection unit 11a at this time may include, for example, log information and trouble information obtained from the substrate processing device 3, and may also include, for example, information input in natural language such as the user's judgment of the trouble.

[0054] The countermeasure generation unit 11b of the processing unit 11 analyzes the log information included in the information acquired in step S1 (step S2). At this time, the countermeasure generation unit 11b can analyze which sensor is related to the trouble that occurred in the substrate processing device 3, for example, based on a value calculated by a predetermined calculation formula on the time-series data of sensor measurements included in the device log, or the result of comparison with a predetermined threshold. Alternatively, the countermeasure generation unit 11b can analyze the cause of the trouble in the substrate processing device 3 based on the log information, for example, by using a learning model that has been pre-trained to classify the causes of trouble in response to the input of log information.

[0055] Furthermore, the countermeasure generation unit 11b extracts information similar to the log information and trouble information acquired in step S1 from the information stored in the knowledge DB 5 as similar cases (step S3). At this time, the countermeasure generation unit 11b uses, for example, a learning model of an encoder that has been pre-machine-trained to convert log information and trouble information into features, and calculates the similarity between the features of the log information and trouble information acquired in step S1 and the features of the information stored in the knowledge DB 5. A predetermined number of pieces of information can then be extracted from the knowledge DB 5 in order of the highest similarity.

[0056] The countermeasure generation unit 11b generates a prompt instructing the generation of a countermeasure for the problem based on the log information and trouble information acquired in step S1, the results of the log information analysis performed in step S2, and the information of similar cases extracted in step S3, and inputs it to the language model 7 (step S4). At this time, the countermeasure generation unit 11b can generate a prompt by appropriately applying the information obtained in steps S1 to S3 to a pre-stored prompt template, for example. Figure 6 is a schematic diagram illustrating an example of a prompt instructing the language model 7 to generate a countermeasure for the problem. In this example, the prompt generated by the countermeasure generation unit 11b is as follows.

[0057] You are a specialist in semiconductor manufacturing equipment model XX. Following the conditions below, refer to the knowledge database and list similar trouble cases and their corresponding solutions. #Trouble Information: ...Recently, E / R and C / D have been consistently low. #Equipment log analysis results using rule list...Log analysis results file f #Numerical analysis results of log information...Numerical analysis results file g #Natural language similarity calculation results...Similarity analysis results file h #Log Information: Image data.jpeg, Recipe No. XX1, Temperature data, Gas flow rate, RF sensor data, Maintenance log #Knowledge Database: Trouble Case DB, Model XX Manual, Customer XX History DB, Semiconductor Substrate Database #Output Format: ... (Note: Figure 6 shows a table as the output format, but this is not the only option; for example, the output format may be specified using comma-separated values.)

[0058] The countermeasure generation unit 11b inputs the prompt generated in step S4 and various information specified in the prompt, such as a file name, to the language model 7. The language model 7, having received this information, generates and outputs one or more countermeasures for the problem while referring to information stored in an appropriate database included in the knowledge DB 5. The countermeasure generation unit 11b acquires information regarding the countermeasures output by the language model 7 in response to the information input in step S4 (step S5).

[0059] The countermeasure generation unit 11b displays information regarding countermeasures for the trouble of the substrate processing device 3 on the display unit 14 based on the information acquired in step S5 (step S6). Figure 7 is a schematic diagram showing an example of a countermeasure display. In this example, the information processing device 1 displays the title string "Trouble Countermeasures" and below it displays information on multiple countermeasures in a table format. The trouble countermeasures table includes items such as "Countermeasure Method," "Solvability," "Cost," "Time," "Estimated Factors," "Risk," "Reference Cases," and "Recurrence Probability." These items are specified as output format in the prompt shown in Figure 6, and the language model 7 generates trouble countermeasures that include information corresponding to these items.

[0060] For example, in Figure 7, the information processing device 1 presents the user with multiple countermeasures, such as "replace part B" and "execute recipe C," as "solution methods." For the "replace part B" solution, the "resolvability" is indicated as "excellent," the "cost" as "*** yen," the "time" as "3 hours," the "estimated cause" as "part malfunction," the "risk" as "none," and the "probability of recurrence" as "20%." For the "execute recipe C" solution, the "resolvability" is indicated as "acceptable: cause isolation only," the "cost" as "*** yen," the "time" as "15 minutes," the "estimated cause" as "optimal recipe changes over time," the "risk" as "device load," and the "probability of recurrence" as "100%." ​​For each countermeasure, a link to a troubleshooting case stored in the knowledge database 5 is provided as a "reference case."

[0061] The user can select a countermeasure to implement from the multiple countermeasures displayed on the display unit 14 of the information processing device 1, or specify conditions for ranking the displayed countermeasures. In this embodiment, the information processing device 1 can also accept ranking conditions from the user via natural language input. After displaying the countermeasures in step S6, the countermeasure generation unit 11b determines whether or not it has received input from the user for ranking conditions of the countermeasures (step S7). If it has not received input for ranking conditions (S7: NO), the countermeasure generation unit 11b proceeds to step S11.

[0062] When the system receives input for ranking conditions (S7: YES), the countermeasure generation unit 11b generates a prompt instructing the language model 7 to rank the countermeasures for the trouble based on the received ranking conditions and inputs it (step S8). Figure 8 is a schematic diagram illustrating an example of a prompt instructing the language model 7 to rank the countermeasures for the trouble. In this example, the prompt generated by the countermeasure generation unit 11b is as follows, and can be generated by adding ranking conditions to the prompt shown in Figure 6. In this example, the system receives input from the user in natural language as a ranking condition, stating, "I want a permanent solution, not just a temporary fix. Countermeasure XX has already been implemented, but its effect was only temporary."

[0063] You are a specialist in semiconductor manufacturing equipment model XX. Following the conditions below, refer to the knowledge database and list similar trouble cases and their corresponding solutions. #Trouble Information: ...Recently, E / R and C / D have been consistently low. #Equipment log analysis results using rule list...Log analysis results file f #Numerical analysis results of log information...Numerical analysis results file g #Natural language similarity calculation results...Similarity analysis results file h #Log Information: Image data.jpeg, Recipe No. XX1, Temperature data, Gas flow rate, RF sensor data, Maintenance log #Ranking Criteria: "We want a permanent solution, not just a temporary fix. Countermeasure XX has been implemented, but its effect was only temporary." #Knowledge Database: Trouble Case DB, Model XX Manual, Customer XX History DB, Semiconductor Substrate Database #Output Format: ... (Note: Figure 8 shows a table as the output format, but this is not the only option; for example, the output format may be specified using comma separation.)

[0064] The countermeasure generation unit 11b inputs the prompt generated in step S8 and various information specified in the prompt, such as a file name, to the language model 7. The language model 7, having received this information, generates one or more countermeasures for the problem while referring to information stored in an appropriate database included in the knowledge DB 5, and outputs them in order of priority according to the specified conditions. The countermeasure generation unit 11b acquires the information of the prioritized countermeasures output by the language model 7 in response to the information input in step S8 (step S9).

[0065] The countermeasure generation unit 11b displays ranked countermeasure information on the display unit 14 based on the information acquired in step S9 (step S10). Figure 9 is a schematic diagram showing one example of the display of ranked countermeasures. In this example, the information processing device 1 displays the title string "Troubleshooting Countermeasures" and below it displays information on multiple countermeasures in a table format. The table of troubleshooting countermeasures shown in this example has a "Rank" item added to the table shown in Figure 7, and the countermeasures are displayed in order from the highest rank. In this example, the countermeasure "Execute Recipe C" with a high "Rank" is displayed higher, and the countermeasure "Replace Part B" with a low "Rank" is displayed lower, with multiple countermeasures being ranked and displayed accordingly.

[0066] After displaying troubleshooting solutions, the troubleshooting solution generation unit 11b accepts a selection of troubleshooting solutions from the user (step S11). Troubleshooting solutions may include those performed by the information processing device 1 and the circuit board processing device 3, such as "execute recipe C," and those performed by the user or an engineer, such as "replace part B." The troubleshooting solution generation unit 11b determines whether the troubleshooting solution selected in step S11 is one to be performed by the information processing device 1 and the circuit board processing device 3 (step S12).

[0067] If the user selects a countermeasure to be implemented by the information processing device 1 and the board processing device 3 (S12: YES), the control processing unit 11e of the processing unit 11 sends an operation command corresponding to the selected countermeasure to the board processing device 3 (step S13), causing the countermeasure to be implemented. After the countermeasure is implemented, the information collection unit 11a of the processing unit 11 acquires information from the board processing device 3 regarding the results of the countermeasure implemented by the board processing device 3, such as log information output by the board processing device 3 (step S14). The information processing device 1 may acquire the results of the countermeasure from another device, such as a device that measures the operation of the board processing device 3, instead of acquiring them from the board processing device 3, or it may acquire whether or not the problem has been resolved by implementing the countermeasure based on input from the user, etc. The information collection unit 11a stores the information regarding the countermeasure implemented in step S13 and the information regarding the results of the countermeasure acquired in step S14 in the knowledge DB 5 in association (step S15), and terminates the process.

[0068] If no countermeasure is selected to be implemented by the information processing device 1 and the substrate processing device 3 (S12: No), that is, if a countermeasure to be implemented by the user is selected, the processing unit 11 of the information processing device 1 waits for the user to implement the countermeasure to the problem, and after the countermeasure is implemented, it receives input from the user regarding the implemented countermeasure and its results (step S16). The information collection unit 11a stores the information regarding the countermeasure and its results received in step S16 in the knowledge DB 5 (step S17), and then terminates the process.

[0069] The information processing device 1 according to this embodiment may, for example, request the user to input information regarding the countermeasures taken for a problem with the substrate processing device 3 and the results of implementing these countermeasures in steps S16 and S14 of the flowchart described above. This information entered by the user is stored in the knowledge DB 5 as a report and will be reflected in the generation of subsequent troubleshooting measures. Figures 10 and 11 are schematic diagrams showing an example of a report input screen.

[0070] The report input screen shown in Figure 10 is a screen for receiving information input when emergency measures are taken in response to a problem with the substrate processing device 3. In this embodiment, the information processing device 1 displays the report input screen shown in this figure in step S16 of the flowchart above, for example, and receives the necessary information input from the user. The report input screen in this example is a screen for receiving information about emergency measures taken in response to a problem with the substrate processing device 3, and the title string "Emergency Measures Report" is displayed at the top of the screen.

[0071] The emergency response report input screen categorizes the various pieces of information that can be entered into four groups: "Report / Approval," "Classification," "Processing Details," and "Requirements." The "Report / Approval" group includes items such as "Application," "Supervisor Approval," and "Notification Recipient upon Completion," and accepts input of information such as the applicant, application date and time, approver, approval date and time, and notification recipient. The "Classification" group includes items such as "Major Category" and "Keywords." In this diagram, "Parts" is entered as the "Major Category," and "CM" is entered as the "Keyword."

[0072] The "Action Details" group includes items such as "Date of Emergency Action" and "Details of Emergency Action," and accepts input of specific details of the countermeasures taken for the trouble with the circuit board processing device 3. In this figure, the date "2023 / 10 / 11," when the emergency action was taken, is entered as "Date of Emergency Action," and the content "BB occurred when AA was performed. CC was replaced and the problem has been resolved." is entered in natural language as "Details of Emergency Action."

[0073] The "Requirements" group includes items such as "Investigation Request," "Recipient," "Requirements," "Report Required / Not Required," and "Response Language," and accepts input of requirements for the manufacturer, distributor, or department in charge of the PCB processing device 3 regarding problems with the PCB processing device 3. In this diagram, "Required" is entered for "Investigation Request," the name of the manufacturer, such as "DDDD," is entered for "Recipient," "Requirements" is entered in natural language such as "I will return the defective product to DDDD, so please investigate," "Report Required / Not Required," and "English" is entered for "Response Language." If "Required" is entered for "Investigation Request," the information processing device 1 can send an investigation request to the manufacturer, etc. entered in "Recipient," for example, by email.

[0074] The report input screen shown in Figure 11 is a screen for receiving information input when permanent measures are taken to address a problem with the substrate processing device 3. In this embodiment, the information processing device 1 displays the report input screen shown in this figure in step S16 of the flowchart described above, for example, and receives the necessary information input from the user. The report input screen in this example is a screen for receiving information about permanent measures taken to address a problem with the substrate processing device 3, and the title string "Permanent Measures Report" is displayed at the top of the screen.

[0075] The report input screen for the permanent action report categorizes the various pieces of information that can be entered into three groups: "Report / Approval," "Permanent Action," and "Attachments." The "Report / Approval" group includes items such as "Application," "Investigation," "Approval," and "Notification Recipient upon Completion," and accepts input of information such as the applicant, application date and time, investigator, approver, approval date and time, and notification recipient.

[0076] The "Permanent Action" group includes fields such as "Cause," "Countermeasure," "Necessity of Design Change," "Issuance Number," and "Comments," and accepts input of specific details of permanent countermeasures taken for the trouble with the substrate processing device 3. In this figure, the "Cause" and "Countermeasure" fields are displayed in a table format. The "Cause" field is entered in natural language as "Event: BB occurred when AA was performed. Cause: BB occurred because MM was NN, causing PP to QQ." The "Countermeasure" field is entered in natural language as "Countermeasure: Change the structure so that MM does not NN." Additionally, "Necessary Design Change" is entered as "Necessary," the "Issuance Number" is entered as "123456," and the "Comments" field is left blank.

[0077] The "Attached Files" group accepts files such as documents provided to the user as a result of an investigation by the manufacturer, for example, regarding a problem with the circuit board processing device 3. In this example, a file titled "Report Document File (2023 / 11 / 14)" is attached, and the information processing device 1 stores the attached file in the knowledge DB 5 along with the information on the permanent action report.

[0078] In this embodiment, the report is created by the user inputting information, but this is not the only way. The information processing device 1 may create the report using the language model 7 and display the created report on the display unit 14 in the manner shown in Figures 10 and 11. In this case, the information processing device 1 receives operations from the user, such as adding and modifying information on the displayed report, and stores the report reflecting the received operations in the knowledge database 5.

[0079] <Recipe Generation Process> Figure 12 is a flowchart showing an example of the procedure for generating a recipe or cost function performed by the information processing device 1 according to this embodiment. The recipe generation unit 11c of the processing unit 11 of the information processing device 1 according to this embodiment displays an appropriate input screen on the display unit 14 and receives instructions from the user in natural language regarding the operation of the substrate processing device 3 (step S31). At this time, for example, the user can input instructions to the information processing device 1 in natural language such as, "For base recipe XX, please apply the L8 experimental design to parameter list YY and execute it." The instructions input by the user may also include examples of queries and answers.

[0080] The recipe generation unit 11c decomposes the user's natural language operation instruction received in step S31 into multiple subtasks using the language model 7 (step S32). At this time, the recipe generation unit 11c generates a prompt based on the operation instruction received in step S31 and a pre-stored prompt template, such as "For base recipe XX, please apply the L8 experimental design to parameter list YY and execute it. To execute this instruction, please first decompose it into subtasks," and inputs it to the language model 7. The prompt may also include examples of queries and answers. In response to this prompt input, the language model 7 outputs the result of decomposing the operation instruction into multiple subtasks as, for example, the following text information.

[0081] "1. Obtain base recipe XX. 2. Load parameter list yy. 3. Prepare L8 experimental design template. 4. Apply parameters to the template. 5. Output recipes corresponding to each experiment. 6. Check if hardware constraints are met. If not, modify the parameter list and return to step 4."

[0082] Next, the recipe generation unit 11c generates program code corresponding to each subtask obtained in step S32 using the language model 7 (step S33). At this time, the recipe generation unit 11c generates a prompt for each of the above subtasks 1 to 6, for example, "Generate program code corresponding to subtask 1," and inputs it to the language model 7. The prompt may also include example queries and answers. In response to the input of this prompt, the language model 7 generates and outputs program code corresponding to the subtask. An example of the code for each subtask output by the language model 7 is shown below.

[0083] #1. Get the base recipe XX. xx = get_recipe("xx") #2. Read the parameter list yy. parameters = pd.read("parameter_list.csv") #3. Prepare the L8 experimental design template. DoE = get_DoE_template("L8") #4. Apply the parameters to the template. for i in range(len(patrameters)): if DoE[i]["p1"] == 0: DoE[i]["p1"] = parameters ["p1"][0] else: DoE[i]["p1"] = parameters ["p1"][1] #5. Output the recipe corresponding to each experiment. for i in range(len(DoE)): recipe = xx.copy() for par in parameters.columns: recipe[per] = DoE[i][par] recipe.to_csv("recipe"+str(i)+".csv") #6. Check if the hardware constraints are met. check_constraint(recipe)

[0084] Next, the recipe generation unit 11c uses the language model 7 to add constraints to the recipe or cost function to be generated (step S34). At this time, the countermeasure generation unit 11b generates a prompt based on a pre-stored prompt template, such as "Generate a function that shows the constraints of the operating range for device model A," and inputs it to the language model 7. Various constraints can be generated by the language model 7, such as power limits, gas type limits, gas flow rate limits, temperature setpoint limits, pressure limits, time setpoint limits, and legal restrictions. The language model 7 can generate constraints based, for example, on manual information for the substrate processing device 3 stored in the knowledge DB 5.

[0085] Next, the recipe generation unit 11c generates a recipe or cost function for operating the substrate processing device 3 based on the program code generated in step S33 and the constraints added in step S34 (step S35). At this time, the recipe generation unit 11c generates a recipe or cost function by, for example, executing the generated program code, and if the generated recipe or cost function contains parts that do not satisfy the constraints, it can generate a recipe or cost function for operating the substrate processing device 3 by correcting or deleting these parts.

[0086] Next, the simulation processing unit 11d of the processing unit 11 simulates substrate processing according to the recipe or cost function generated in step S35 using a simulator that reproduces the operation of a pre-prepared substrate processing apparatus 3 (step S36). Based on the simulation results in step S36, the simulation processing unit 11d determines whether or not there is a risk in the substrate processing based on the recipe or cost function generated in step S35 (step S37). At this time, the simulation processing unit 11d may use the constraint conditions added in step S34 as one of the conditions for determining whether or not there is a risk, by checking whether or not the simulation results satisfy the constraint conditions. Alternatively, for example, the simulation processing unit 11d may generate a cost function for calculating the cost related to substrate processing using the language model 7, calculate the cost related to the recipe using the generated cost function, and determine whether or not there is a risk by checking whether or not the calculated cost exceeds a predetermined threshold. The costs calculated using a cost function may include a variety of factors, such as prohibited actions, conditions to be avoided, feasibility, predicted output, processing time, electricity consumption, water consumption, gas emissions, environmental impact indicators, cost indicators, throughput differences between upstream and downstream processes, load on equipment, risks to employees and the factory, or material prices at that time.

[0087] If the simulation processing unit 11d determines that there is no risk (S37: NO), the display processing unit 11f of the processing unit 11 displays the recipe or cost function generated in step S35 on the display unit 14 (step S38). Figure 13 is a schematic diagram showing one example of a recipe generated by the information processing device 1. In the information processing system according to this embodiment, the recipe for operating the substrate processing device 3 defines the values ​​of multiple settable parameters for multiple steps of the operation related to substrate processing. The display processing unit 11f displays the recipe generated in step S35 in a table format, for example, where the steps of substrate processing and the types of parameters are mapped vertically and horizontally. The display processing unit 11f may also highlight important values ​​of the recipe based on information entered by the user, for example. As an example, in Figure 13, the display processing unit 11f highlights the value of parameter 3 in step S3 (highlighting is indicated by hatching in this figure).

[0088] Furthermore, if the display processing unit 11f generates a cost function in step S35, it may display the generated cost function formula on the display unit 14, for example, as follows: "def cost(): cost = term1() + term2() + … return cost"

[0089] The user confirms whether the recipe or cost function displayed on the display unit 14 of the information processing device 1 is appropriate for the operation instruction entered in step S31, and approves or rejects the recipe or cost function to the information processing device 1. The recipe generation unit 11c determines whether the recipe or cost function displayed in step S38 has been approved by the user (step S39). To accept user approval, the information processing device 1 may, for example, provide a button on the screen displaying the recipe or cost function that accepts the operation to approve / reject the recipe or cost function.

[0090] If the simulation processing unit 11d determines that there is a risk in step S37 (S37: YES), or if the user does not approve the recipe or cost function in step S39 (S39: NO), the recipe generation unit 11c receives instructions from the user in natural language to modify the recipe or cost function (step S40). The recipe generation unit 11c returns to step S32 and regenerates the recipe or cost function.

[0091] If the user approves the recipe or cost function (S39: YES), the control processing unit 11e of the processing unit 11 sends the generated recipe or cost function to the substrate processing device 3 and issues a command to perform substrate processing, thereby causing the substrate processing device 3 to perform substrate processing (step S41). The information collection unit 11a of the processing unit 11 acquires information regarding the results of the substrate processing performed in step S41, such as log information, from the substrate processing device 3 (step S42). The information collection unit 11a stores the recipe or cost function generated in step S35 and the results of the substrate processing acquired in step S42 in the knowledge DB 5 in association (step S43), and terminates the process.

[0092] <Summary> In the information processing system according to this embodiment with the above configuration, the information processing device 1 acquires trouble information related to troubles with the substrate processing device 3 and log information output by the substrate processing device 3 that performed the substrate processing. Based on the acquired trouble information and log information, the language model 7, which uses a knowledge DB 5 that stores information related to the substrate processing device 3, generates countermeasure information regarding countermeasures for troubles, and displays (outputs) the generated countermeasure information. As a result, the information processing system according to this embodiment is expected to help resolve troubles that occur in the substrate processing device 3 and to support users of the substrate processing device 3.

[0093] Furthermore, in the information processing system according to this embodiment, the information processing device 1 generates multiple pieces of information on countermeasures for a problem using the language model 7, and displays the generated multiple pieces of information on countermeasures in an ordered manner. In addition, in the information processing system according to this embodiment, the information processing device 1 accepts natural language input of information on the ordering criteria, and based on the information on the criteria received, displays the multiple pieces of information on countermeasures in an ordered manner using the language model 7. As a result, the information processing system according to this embodiment can display multiple pieces of information on countermeasures in an ordered manner according to the criteria requested by the user, and is expected to support the user using the substrate processing device 3.

[0094] Furthermore, in the information processing system according to this embodiment, the knowledge DB 5 used by the information processing device 1 includes a database that stores examples of troubles in the substrate processing device 3 and countermeasures for these troubles, as well as a database that stores the manual for the substrate processing device 3. As a result, the information processing system according to this embodiment can use the language model 7 to obtain the information necessary to deal with troubles from the knowledge DB 5 and is expected to present countermeasures to the user.

[0095] Furthermore, in the information processing system according to this embodiment, the information processing device 1 extracts information similar to the acquired trouble information and log information from the knowledge DB 5, and generates countermeasure information based on the extracted information. As a result, the information processing system according to this embodiment is expected to extract appropriate information from the knowledge DB 5 and generate countermeasure information.

[0096] Furthermore, in the information processing system according to this embodiment, the log information of the substrate processing device 3 includes measurement information from sensors. The information processing device 1 receives the measurement information as input and inputs the measurement information contained in the acquired log information into a learning model that has been pre-trained to predict events that occur in substrate processing. The information processing device 1 obtains the event prediction result output by the learning model and generates countermeasure information based on the acquired prediction result and trouble information. Thus, the information processing system according to this embodiment is expected to generate countermeasure information for troubles corresponding to the log information obtained from the substrate processing device 3.

[0097] Furthermore, in the information processing system according to this embodiment, the information processing device 1 receives input from the user regarding the results of countermeasures implemented based on the countermeasure information, and stores the trouble information, log information, countermeasure information, and countermeasure results in the knowledge DB 5 in association with each other. As a result, the information processing system according to this embodiment can accumulate information on how to deal with troubles in the knowledge DB 5, and is expected to provide the user with better countermeasures for troubles that occur in the future.

[0098] Furthermore, in the information processing system according to this embodiment, the information processing device 1 receives natural language input of instructions for operation to the substrate processing device 3, generates a recipe or cost function for operating the substrate processing device 3 using a language model 7 based on the received instructions, and causes the substrate processing device 3 to perform substrate processing based on the generated recipe or cost function. As a result, the information processing system according to this embodiment is expected to support users of the substrate processing device 3 by allowing users to input substrate processing conditions etc. in natural language and automatically generate a recipe or cost function.

[0099] Furthermore, in the information processing system according to this embodiment, the information processing device 1 uses a language model 7 to decompose the operation related to the received instruction into multiple subtasks and generates a recipe or cost function that causes the substrate processing device 3 to perform the operation of each subtask. As a result, the information processing system according to this embodiment is expected to be able to automatically generate recipes or cost functions with greater accuracy.

[0100] Furthermore, in the information processing system according to this embodiment, the information processing device 1 simulates the substrate processing of the substrate processing device 3 based on the generated recipe or cost function, and determines whether or not there is a risk in the operation based on the recipe based on the simulation results. As a result, the information processing system according to this embodiment can check for the presence or absence of risk before actually performing the substrate processing based on the generated recipe or cost function in the substrate processing device 3, and it is expected that this will prevent the execution of substrate processing based on a risky recipe or cost function.

[0101] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of this disclosure is indicated by the claims, not in the sense described above, and all modifications within the meaning and scope equivalent to the claims are intended.

[0102] The matters described in each embodiment can be combined with each other. Furthermore, the independent and dependent claims described in the claims can be combined with each other in any combination, regardless of the form of reference. In addition, the claims use a form in which claims referencing two or more other claims (multi-claim form), but are not limited to this. A form in which multi-claims referencing at least one multi-claim (multi-multi-claim) may also be used.

[0103] 1 Information Processing Unit (Computer) 3 Circuit Board Processing Unit 5 Knowledge DB 7 Language Model 11 Processing Unit 11a Information Gathering Unit 11b Countermeasure Generation Unit 11c Recipe Generation Unit 11d Simulation Processing Unit 11e Control Processing Unit 11d Display Processing Unit 12 Storage Unit 12a Program (Computer Program) 12b Learning Model Storage Unit 13 Communication Unit 14 Display Unit 15 Operation Unit

Claims

1. A computer program that causes a computer to execute a process to acquire trouble information related to a problem with a circuit board processing device and log information output by the circuit board processing device that performed the circuit board processing, generate countermeasure information related to countermeasures for the problem using a language model that uses a knowledge database storing information related to the circuit board processing device based on the acquired trouble information and log information, and output the generated countermeasure information.

2. The computer program according to claim 1, which generates multiple pieces of countermeasure information using the language model and outputs the generated multiple pieces of countermeasure information in an ordered manner.

3. A computer program according to claim 2, which receives input information regarding ordering criteria and outputs a plurality of countermeasure information in order based on the received information regarding the criteria.

4. The computer program according to claim 1, wherein the knowledge database includes a database storing examples of troubles in the substrate processing apparatus and countermeasures for such troubles, and a database storing the manual for the substrate processing apparatus.

5. A computer program according to claim 1, which extracts information similar to the acquired trouble information and log information from the knowledge database, and generates the countermeasure information based on the extracted information.

6. The computer program according to claim 1, wherein the log information includes measurement information from a sensor of the substrate processing apparatus, the measurement information included in the acquired log information is input to a learning model that has been trained to receive the measurement information as input and predict events that occur in substrate processing, the prediction results of events output by the learning model in response to the input of the measurement information is acquired, and the countermeasure information is generated based on the acquired prediction results and the trouble information.

7. The computer program according to claim 1, which receives input of the results of countermeasures implemented based on the countermeasure information, and stores the acquired trouble information and log information, the generated countermeasure information, and the received countermeasure results in the knowledge database in association with each other.

8. A computer program according to claim 1, which receives an input of an instruction to operate the substrate processing apparatus, generates setting information for operating the substrate processing apparatus using the language model based on the received instruction, and causes the substrate processing apparatus to perform substrate processing based on the generated setting information.

9. A computer program that receives input of instructions for operation to a substrate processing device, generates setting information for operating the substrate processing device using a language model that uses a knowledge database storing information about the substrate processing device based on the received instructions, and causes a computer to execute a process to make the substrate processing device perform substrate processing based on the generated setting information.

10. A computer program according to claim 8 or 9, which decomposes the operation related to the received instruction into a plurality of subtasks and generates a plurality of setting information for causing the substrate processing device to perform the operations of the subtasks.

11. A computer program according to claim 8 or 9, which simulates the operation of the substrate processing apparatus based on the generated setting information and determines whether or not there is a risk of operation based on the setting information.

12. A computer program according to claim 11, which obtains a cost function relating to the substrate processing apparatus and determines whether or not the risk exists based on the obtained cost function.

13. An information processing method comprising: an information processing device acquiring trouble information relating to a problem with a substrate processing device and log information output by the substrate processing device that performed the substrate processing; generating countermeasure information relating to countermeasures for the trouble using a language model that uses a knowledge database storing information relating to the substrate processing device, based on the acquired trouble information and log information; and outputting the generated countermeasure information.

14. An information processing method comprising: an information processing device receiving an input of an instruction to operate a substrate processing device; generating setting information for operating the substrate processing device using a language model that uses a knowledge database storing information about the substrate processing device based on the received instruction; and causing the substrate processing device to perform substrate processing based on the generated setting information.

15. An information processing device comprising a processing unit, wherein the processing unit acquires trouble information relating to a problem with the substrate processing device and log information output by the substrate processing device that performed the substrate processing, generates countermeasure information relating to countermeasures for the trouble using a language model that uses a knowledge database storing information relating to the substrate processing device, based on the acquired trouble information and log information, and outputs the generated countermeasure information.

16. An information processing device comprising a processing unit, wherein the processing unit receives an input of an instruction to operate a substrate processing device, generates setting information for operating the substrate processing device using a language model that uses a knowledge database storing information about the substrate processing device based on the received instruction, and causes the substrate processing device to perform substrate processing based on the generated setting information.

Images