Laser management server and laser management method

Abstract

SOLUTION: A laser management server includes: a query input processor for generating a first query item requiring external information and a second query item without requiring external information; an agent action processor for acquiring necessary external information from non-structured data on a laser device and structured data on the laser device and generating an agent response to the first query item; a response output processor for receiving the agent response and generating a query item prompt related to the second query item; and a large language model processor for receiving the query item prompt from the response output processor and generating a query item response. The response output processor receives the query item response and configures a query response before transmitting it to a transmission / reception processor.EFFECT: A laser management system comprises a large language model processor, which prevents a form of a query and a method for displaying a query response from being limited.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present disclosure relates to a laser management server and a laser management method. [Background technology]

[0002] In recent years, semiconductor exposure devices have been required to improve their resolution in response to the miniaturization and high integration of semiconductor integrated circuits. To this end, the wavelength of light emitted from exposure light sources has been shortened. For example, KrF excimer laser devices that output laser light with a wavelength of approximately 248 nm and ArF excimer laser devices that output laser light with a wavelength of approximately 193 nm are used as gas laser devices for exposure.

[0003] The spectral linewidth of the spontaneously oscillating light from KrF excimer laser devices and ArF excimer laser devices is as wide as 350 to 400 pm. Therefore, if a projection lens is constructed using a material that transmits ultraviolet light, such as KrF and ArF laser light, chromatic aberration may occur. As a result, resolution may decrease. Therefore, it is necessary to narrow the spectral linewidth of the laser light output from the gas laser device to a level where chromatic aberration is negligible. Therefore, a line narrowing module (LNM) containing a line narrowing element (e.g., an etalon or grating) may be installed inside the laser resonator of the gas laser device to narrow the spectral linewidth. Hereinafter, a gas laser device with a narrowed spectral linewidth is referred to as a line narrowing gas laser device. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-19441 [Patent Document 2] Japanese Patent Application Publication No. 1-216299 [Patent Document 3] Overview of International Publication No. 2020 / 161865

[0005] a query input processor that receives the query from the transmission / reception processor, decomposes the query, and generates a first query item that requires external information and a second query item that does not require external information; an agent action processor that receives the first query item from the query input processor, acquires necessary external information from unstructured data including a laser device manual and maintenance report and structured data including operation data of the laser device, and generates an agent response that is a response to the first query item; a response output processor that receives the second query item from the query input processor, receives the agent response from the agent action processor, and generates a query item prompt related to the second query item; and a large-scale language model processor that receives the query item prompt from the response output processor and generates a query item response that is a response to the query item prompt, wherein the response output processor receives the query item response from the large-scale language model processor, constructs a query response based on the agent response and the query item response, and sends the query response to the transmission / reception processor.

[0006] A laser management method for a laser device according to another aspect of the present disclosure includes a first step in which a transceiver processor receives a query from outside; a second step in which a query input processor breaks down the query received from the transceiver processor and generates a first query item that requires external information and a second query item that does not require external information; a third step in which an agent action processor receives the first query item from the query input processor, acquires the required external information from unstructured data including a manual and a maintenance report for the laser device and structured data including operation data for the laser device, and generates an agent response that is a response to the first query item; and a third step in which a response output processor outputs the second query item to the query input processor. a fourth step in which the large scale language model processor receives the query item prompt from the response output processor, receives an agent response from the agent action processor, and generates a query item prompt for the second query item; a fifth step in which the large scale language model processor receives the query item prompt from the response output processor and generates a query item response that is a response to the query item prompt; a sixth step in which the response output processor receives the query item response from the large scale language model processor and constructs a query response based on the agent response and the query item response; and a seventh step in which the transmission / reception processor receives the query response from the response output processor and transmits the query response to an external device. [Brief explanation of the drawings]

[0007] Some embodiments of the present disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which: [Figure 1] FIG. 1 is a diagram showing the configuration of an exemplary laser device. [Figure 2] FIG. 2 is a diagram showing the configuration of a laser management system according to a comparative example. [Figure 3] FIG. 3 is a diagram illustrating a laser management system according to the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating an operation flow of the laser management server according to the first embodiment. [Figure 5] FIG. 5 is a diagram showing examples of query items generated from a user's query according to the first embodiment. [Figure 6] FIG. 6 is a diagram showing an example of a reply to a user's query according to the first embodiment. [Figure 7] FIG. 7 is a diagram showing a laser management system according to a modified example of the first embodiment. [Figure 8] FIG. 8 is a diagram showing an operation flow of the laser management server according to the modified example of the first embodiment. [Figure 9] FIG. 9 is a diagram illustrating a laser management system according to the second embodiment. [Figure 10] FIG. 10 is a diagram showing an example of a reply to a user's query according to the second embodiment. Embodiment

[0008] -table of contents- 1. Explanation of terms 2. Comparative Example 2.1 Laser device 2.1.1 Configuration 2.1.2 Operation 2.2 Laser Management System 2.2.1 Configuration 2.2.2 Operation 2.3 Challenges 3. Embodiment 1 3.1 Configuration 3.2 Operation 3.3 Actions and Effects 4. Modification of the First Embodiment 4.1 Configuration 4.2 Operation 4.3 Actions and Effects 5. Embodiment 2 5.1 Configuration 5.2 Operation 5.3 Actions and Effects 6.Other Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. The embodiments described below show some examples of the present disclosure and do not limit the content of the present disclosure. Furthermore, not all of the configurations and operations described in each embodiment are necessarily essential as the configurations and operations of the present disclosure. Note that the same components are given the same reference symbols, and redundant explanations will be omitted.

[0009] 1. Explanation of terms The terms used in this specification are defined as follows: "Standard software" is application software that performs at least one task of managing, monitoring, and analyzing a specific device based on predefined functionality and display specifications.

[0010] A "web app" is application software that can be accessed over the internet using a web browser. Web apps can be accessed on a variety of devices, including PCs, smartphones, and tablets, as long as they are connected to the internet, without the need to install application software.

[0011] A "standard operation screen" is a pre-designed standard user interface (UI) in application software.

[0012] An "interactive operation screen" is an operation screen in which a user asks questions or gives commands in natural language, and the application software responds to the questions or commands from the user.

[0013] "Unstructured data" is data that lacks regularity or continuity and is difficult to manage in a typical database or table. Examples of unstructured data include text documents, images, and audio files.

[0014] "Operation data" is operational information generated by the laser device.

[0015] A "specialized information processor" is a chat agent with advanced expertise in a specific field. Unlike general chat agents, a specialized information processor can respond appropriately to conversations related to terms and concepts in a specialized field.

[0016] A "query" is an inquiry or instruction received by the entire system from a user. A query is an instruction string written in natural language that specifies the information to be obtained and the format in which that information is to be provided, or that specifies the operation to be performed on the laser device. An example of a query is "Report the operational diagnosis of laser 65400011 in standard format."

[0017] A "query response" is a result written in natural language that a system returns in response to a specific query. Depending on the content, the query response may contain not only text but also graphs, images, and other formats.

[0018] The Large Language Model Processor (LLM) performs natural language processing using an AI (Artificial Intelligence) model trained on large amounts of text data. The LLM understands and executes complex language tasks.

[0019] A "prompt" is a natural language text instruction or question that you send to an LLM so that the LLM knows what to respond to or what task to perform.

[0020] The Query Input Processor (QIP) analyzes the user's request (query) and performs the process of replacing the query with an appropriate input prompt for the LLM.

[0021] The "Response Output Processor (ROP)" performs processing to configure the content returned as the LLM's reply response in a format and content that conforms to the user's specifications.

[0022] The Agent Action Processor (AAP) analyzes the output of the LLM, obtains the information necessary to form an appropriate response from external sources, and performs the process of assembling an appropriate response that matches the user's query.

[0023] 2. Comparative Example 2.1 Laser device 2.1.1 Configuration 1 is a diagram showing the configuration of an exemplary laser apparatus 10. The laser apparatus 10 is a discharge-pumped gas laser apparatus and includes an oscillator (OSC) 20, an amplifier (AMP) 50, a monitor module 70, and a laser processor 80. The processor of the present disclosure is a processing device that includes a storage device in which a control program is stored and a CPU that executes the control program. The processor is specially configured or programmed to perform various processes.

[0024] The OSC 20 includes a line narrowing module (LNM) 22 , a chamber 24 , an output coupler (OC) 26 , a pulsed power module (PPM) 28 , and a charger 32 .

[0025] The LNM 22 includes prisms 36 and 38, a grating 42, and a rotation stage 44 that rotates the prism 38. The LNM 22 controls the central wavelength of the pulsed laser light by rotating the prism 38 to change the angle of incidence on the grating 42.

[0026] The chamber 24 includes a pair of discharge electrodes 46, 47 and two windows 48, 49 through which the laser light passes. An excimer laser gas is introduced into the chamber 24. The excimer laser gas includes, for example, a rare gas (Ar gas or Kr gas), a halogen gas (F2 gas), and a buffer gas (Ne gas).

[0027] The OC26 is a partial reflector that reflects part of the pulsed laser light and transmits the other part.

[0028] The LNM 22 and the OC 26 together constitute an optical resonator, and the chamber 24 is disposed on the optical path of the optical resonator.

[0029] The AMP 50 includes a rear mirror (RM) 52 , a chamber 54 , an output coupler (OC) 56 , a pulsed power module (PPM) 58 , and a charger 62 .

[0030] The RM 52 is a partial reflector that reflects part of the pulsed laser beam and transmits the other part. The reflectance of the RM 52 may be 80% to 90%.

[0031] The chamber includes a pair of discharge electrodes 64, 65 and two windows 66, 67 through which the laser light passes. An excimer laser gas is introduced into the chamber .

[0032] The OC 56 is a partial reflector that reflects and transmits a part of the pulsed laser light. The reflectance of the OC 56 may be 10% to 30%.

[0033] The RM 52 and the OC 56 together constitute an optical resonator, and the chamber 54 is disposed on the optical path of the optical resonator. The optical resonator may be a Fabry-Perot type optical resonator.

[0034] The monitor module 70 includes beam splitters 72 and 74, a spectral detector 76 that measures the wavelength and spectral linewidth of the pulsed laser light, and an optical sensor 78 that detects the pulse energy of the pulsed laser light. The spectral detector 76 may be an etalon spectrometer. The optical sensor 78 may be a photodiode.

[0035] 2.1.2 Operation The laser processor 80 receives a target center wavelength λt and a target pulse energy Et from an external device such as an exposure device (not shown). Then, the laser processor 80 sets the charging voltage V1 of the charger 32 and the charging voltage V2 of the charger 62 so as to obtain a pulsed laser beam having the target pulse energy Et.

[0036] A first charging capacitor (not shown) in the PPM 28 is charged with a charging voltage V1, and a second charging capacitor (not shown) in the PPM 58 is charged with a charging voltage V2.

[0037] When the laser processor 80 receives a light emission trigger Trt from an external device such as an exposure device, it sends a light emission trigger Tr1 to the switch 33 in the PPM 28. When the switch 33 operates, the charge stored in the first charging capacitor is converted into a high-voltage pulse corresponding to the charging voltage V1 in the PPM 28 and applied between the discharge electrodes 46 and 47 in the chamber 24.

[0038] As a result, a discharge occurs between the discharge electrodes 46 and 47 in the chamber 24, exciting the laser gas. Then, the wavelength narrowed by the optical resonator composed of the OC 26 and the LNM 22 is output from the OSC 20 as seed light with an ultraviolet wavelength of 150 nm to 380 nm. The wavelength of the seed light may be the oscillation wavelength of an ArF excimer laser or a KrF excimer laser.

[0039] In addition, when the laser processor 80 receives the light emission trigger Trt, it sends a light emission trigger Tr2 to the switch 59 of the PPM 58 so that a discharge occurs between the discharge electrodes 64, 65 when the seed light output from the OSC 20 enters the discharge space of the chamber 54 of the AMP 50.

[0040] When the switch 59 is operated, the charge stored in the second charging capacitor is converted into a high voltage pulse corresponding to the charging voltage V2 in the PPM 58 and applied between the discharge electrodes 64 and 65 in the chamber 54.

[0041] As a result, a discharge occurs between the discharge electrodes 64, 65 in the chamber 54, and the laser gas is excited. At this timing, the seed light output from the OSC 20 passes through the RM 52 and enters the discharge space in the chamber 54. The incident seed light is amplified by the optical resonator composed of the RM 52 and the OC 56, and is output from the AMP 50.

[0042] The pulsed laser light output from the AMP 50 enters the monitor module 70. A portion of the pulsed laser light that enters the monitor module 70 is reflected by a beam splitter 72, and a further portion is reflected by a beam splitter 74 and enters a spectrum detector 76. The pulsed laser light that passes through the beam splitter 74 then enters an optical sensor 78.

[0043] The spectrum detector 76 measures the central wavelength of the pulsed laser light, and the optical sensor 78 measures the pulse energy of the pulsed laser light.

[0044] The laser processor 80 may control the rotation stage 44 in the LNM 22 so that the center wavelength measured by the spectrum detector 76 becomes the target center wavelength λt. The rotation stage 44 may be a rotation stage including a piezoelectric element.

[0045] The laser processor 80 may control the charging voltage V2 output from the charger 62 so that the pulse energy measured by the optical sensor 78 becomes the target pulse energy Et.

[0046] 2.2 Laser Management System 2.2.1 Configuration 2 is a diagram showing the configuration of a laser management system 100 according to a comparative example. The comparative example of the present disclosure is a configuration that the applicant recognizes as being known only by the applicant, and is not a publicly known example that the applicant acknowledges. The laser management system 100 performs at least one of management, monitoring, and analysis of the laser device 10.

[0047] The laser management system 100 includes a laser device 10, an operation data server 104, a laser management server 110, and a terminal 150.

[0048] The operation data server 104 is a data server in which operation data of the laser device 10 is stored.

[0049] The terminal 150 is, for example, a personal computer, a smartphone, a tablet, or the like.

[0050] The laser management server 110 includes a standard software processor (SSP) 128 on which the standard software 122 is stored.

[0051] The SSP 128 acquires operational data of the laser device 10 from the operational data server 104 via the network.

[0052] The standard software 122 is provided in the form of a web application that is operated via a network.

[0053] The standard operation screen 124 of the standard software 122 is displayed on the terminal 150 and is designed so that the user can operate it.

[0054] 2.2.2 Operation The laser device 10 transmits the operation data to the operation data server 104, and the operation data server 104 stores the operation data.

[0055] The laser management server 110 receives a request from the user via the standard operation screen 124 to view predetermined information, etc.

[0056] The standard software 122 of the laser management server 110 uses the accumulated operational data to perform analysis of a specific laser device 10. The standard software 122 provides the user with at least one of the functions of managing, monitoring, and analyzing a predetermined laser device 10.

[0057] The routine software 122 of the laser management server 110 displays the requested information and the like on a routine operation screen 124 .

[0058] The user views the information on the routine operation screen 124 .

[0059] Different standard software 122 or standard operation screen 124 are provided depending on the purpose of each of different users, such as a field service engineer (FSE), an equipment owner, and a research and development engineer.

[0060] 2.3 Challenges At semiconductor manufacturing sites, standard software 122 is used to manage, monitor, and maintain the operation of laser device 10. Standard software 122 can only display or operate based on predetermined operating specifications.

[0061] In recent semiconductor manufacturing sites, as manufacturing processes have become more complex and diverse, the need for data analysis, monitoring, and retrieval using application software has become more sophisticated. However, it is difficult to meet the needs of various users using only the operations or screens predetermined by the standard software 122.

[0062] Furthermore, the data that application software can currently handle is limited to structured data with clear values, such as equipment parameters and sensor data. Furthermore, even handwritten reports, manuals, and images contain information that is useful for increasingly sophisticated data analysis. However, because this information is unstructured data with no regularity or continuity, it is difficult to handle it as analytical data.

[0063] 3. Embodiment 1 3.1 Configuration Fig. 3 is a diagram showing a laser management system 100A according to embodiment 1. Differences between the configuration shown in Fig. 3 and Fig. 2 will be described.

[0064] Compared to the laser management system 100, the laser management system 100A differs in the configuration within the laser management server 110A, in the fact that a document server 106 is connected to the laser management server 110A, and in the fact that the screen that the user actually uses via the terminal 150 has been changed from the standard operation screen 124 to the interactive operation screen 126.

[0065] The laser management server 110A includes a sending and receiving processor (SRP) 130, a query input processor (QIP) 132, a large-scale language model processor (LLM) 134, an agent action processor (AAP) 136, and a response output processor (ROP) 138.

[0066] The SRP 130, the QIP 132, the LLM 134, the AAP 136, and the ROP 138 may be application software or hardware such as a CPU. Each piece of application software may be implemented as a single processor.

[0067] The user's input and output to the laser management system 100A is performed on the interactive operation screen 126 of the terminal 150 via the SRP 130.

[0068] The document server 106 includes a laser device technical document database (DB) that stores unstructured data such as technical documents, such as manuals and maintenance reports for the laser device 10 .

[0069] At least one of the operational data server 104 and the document server 106 may be implemented within the laser management server 110A.

[0070] 3.2 Operation FIG. 4 is a diagram showing an operation flow of the laser management server 110A according to the first embodiment.

[0071] [Step 1] The SRP 130 receives the query W0 from the terminal 150. Step 1 is an example of the "first step" in the present disclosure. The method by which the user inputs the query W0 to the terminal 150 is not limited to key input from a physical or application software keyboard, but may also be voice input. Furthermore, for example, the query W0 may also be input via communication using at least one of email, short message, and voicemail.

[0072] [Step 2] The QIP 132 decomposes the query W0 received from the SRP 130 to generate a first query term W1 that requires external information and a second query term W2 that does not require external information. Step 2 is an example of the “second step” in this disclosure.

[0073] Fig. 5 is a diagram showing an example of a first query item W1 and a second query item W2 generated from a query W0. Fig. 5 shows an example of query item generation when the query W0 received by the QIP 132 is "Report the operational diagnosis of laser 65400011 in standard format." Note that "laser 65400011" is a name that identifies the model of the laser.

[0074] In this case, the QIP 132 breaks down the received query W0 into four query items. The query item with item number 1 is "Information about the model of laser 65400011," the query item with item number 2 is "Information about the standard format of the diagnostic report," the query item with item number 3 is "Analysis of the operation data of laser 65400011," and the query item with item number 4 is "Generate a diagnostic report based on the information."

[0075] Of these four query items, query items with item numbers 1 and 3 are difficult to answer using LLM134 alone and require external information to obtain an appropriate answer, so they are classified as the first query item W1. Query items with item numbers 2 and 4 do not require external information to be answered, so they are classified as the second query item W2.

[0076] In this case, external information refers to information present in the operational data server 104 or the document server 106. The criterion for determining whether a query item requires external information is, for example, whether a highly accurate response can be generated by the LLM 134 alone. For query items classified as the first query item W1, it is difficult to generate a highly accurate response by the LLM 134 alone. For query items classified as the second query item W2, it is possible to generate a highly accurate response by the LLM 134 alone. A highly accurate response is a response that is based on facts and has few misidentifications or errors.

[0077] Depending on the content of the query W0 received from the SRP 130, one of the first query item W1 and the second query item W2 may be generated.

[0078] [Step 3] The QIP 132 sends the second query term W2 to the ROP 138.

[0079] [Step 4] The QIP 132 sends the first query term W1 to the AAP 136. The first query term W1 may also be sent to the ROP 138 for use in verifying information.

[0080] [Step 5] The AAP 136 assigns a corresponding external information processing program to the first query item W1. The external information processing program acquires necessary information from the outside (the historical data server 104 or the document server 106), analyzes it, and compiles the results.

[0081] For example, for the first query item W1 of item number 3 in Figure 5, AAP136 obtains from the operation data server 104 the applied voltage (HV) between the discharge electrodes 46 and 47 in the chamber 24 of OSC20 of laser 65400011, the gas pressure in the chamber 24 of OSC20, and the total number of shots in the device, analyzes this data, and summarizes the results.

[0082] The information acquired by the external information processing program may include, for example, an operation diagnosis report, a parameter graph, a part replacement procedure, a light source technical explanation, and a record of past maintenance.

[0083] [Step 6] The AAP 136 executes all of the assigned external information processing programs. The AAP 136 then performs external information construction processing to compile all information related to the execution results and generate an agent response W3, which is a response to the first query item W1. Steps 5 to 6 are an example of the "third step" in this disclosure. Note that an example of the agent response W3 is the AAP generation results 152 and 154 in FIG. 6, which will be described later, and is not limited to text information but may include graphs, tables, etc.

[0084] [Step 7] The AAP 136 sends the agent response W3 to the ROP 138.

[0085] [Step 8] The ROP 138 receives the second query term W2 from the QIP 132.

[0086] [Step 9] The ROP 138 may receive the agent response W3 from the AAP 136 and check whether the required information is included.

[0087] [Step 10] The ROP 138 generates a query item prompt W4 related to the response to the second query item W2. Steps 8 to 10 are an example of the "fourth step" in this disclosure. For example, the query item prompt W4 for the second query item W2 with item number 2 shown in FIG. 5 is "What is the standard format for a diagnostic report?". Also, the query item prompt W4 for the second query item W2 with item number 4 shown in FIG. 5 includes the agent response W3 required for the response and the prompt "Generate a report for this agent response."

[0088] [Step 11] The LLM 134 receives the query item prompt W4 from the ROP 138 and generates a query item response W5 that is a response to the query item prompt W4. Step 11 is an example of the “fifth step” in this disclosure.

[0089] For example, the query item response W5 to the query item prompt W4 of the second query item W2 with item number 2 shown in Fig. 5 is "Report on the overall diagnostic results and individual performance based on the device's operation data." Also, the query item response W5 to the query item prompt W4 of the second query item W2 with item number 4 shown in Fig. 5 is, for example, as follows:

[0090] 1. Comprehensive diagnosis Laser 65400011 is currently in good working order. The trends in OSC chamber gas pressure and HV over the past four weeks indicate a slight upward trend in gas pressure. It is still too early to replace the part, but if the current trend continues, it may need to be replaced in up to three months.”

[0091] [Step 12] The ROP 138 receives the query item response W5 from the LLM 134. The ROP 138 then constructs a query response W6 based on the agent response W3 and the query item response W5. Step 12 is an example of the "sixth step" of this disclosure.

[0092] [Step 13] The ROP 138 sends the constructed query response W6 to the SRP 130.

[0093] [Step 14] The SRP 130 transmits the query response W6 to the terminal 150. Step 14 is an example of the "seventh step" in this disclosure.

[0094] [Step 15] The query response W6 is displayed on the interactive operation screen 126 of the terminal 150.

[0095] Fig. 6 is a diagram showing an example of a query response W6 displayed on terminal 150. Fig. 6 is a display example of the query response W6 when query W0 is "Please report the operational diagnosis of laser 65400011 in standard format." Upon receiving query W0, laser management server 110A returns a query response W6 with the content shown in Fig. 6 to the user.

[0096] 6, the query response W6 displayed on the interactive operation screen 126 includes LLM generation results 151 and 153 and AAP generation results 152 and 154. The LLM generation results 151 and 153 are results generated by the LLM 134. The AAP generation results 152 and 154 are results generated by the AAP 136. The AAP generation results 152 and 154 may include graphs, tables, etc. Note that the creation of graphs of operational data may be performed by the AAP 136 or the LLM 134.

[0097] On the interactive operation screen 126, the AAP generation result 152 may be displayed immediately below the LLM generation result 151. In the case of Fig. 6, the AAP generation result 152 is information related to the LLM generation result 151, and the AAP generation result 154 is information related to the LLM generation result 153. Therefore, in order to display this related information on the screen in an organized manner that is easy to view, it is preferable that the query response W6 is configured with a layout in which the LLM generation result 151, the AAP generation result 152, the LLM generation result 153, and the AAP generation result 154 are arranged alternately, for example.

[0098] The ROP 138 combines the LLM generation results 151 and 153 generated by the LLM 134 with the AAP generation results 152 and 154 received from the AAP 136 to generate a query response W6 that realizes the screen display shown in FIG.

[0099] When the SRP 130 receives a query W0 again in response to the query response W6 displayed on the interactive operation screen 126, the laser management server 110A again performs the processes of steps 2 to 15. Steps 1 to 15 are an example of the "laser management method" of the present disclosure.

[0100] 3.3 Actions and Effects The laser management system 100A according to the first embodiment includes the LLM 134 capable of language processing in the laser management server 110A, and therefore there are no limitations on the format of the query W0 or the display method of the query response W6.

[0101] A first query item W1 requiring external information is acquired by the AAP 136. The acquired external information is transmitted to the LLM 134 via the ROP 138. This allows the LLM 134 to utilize information specific to the laser device 10, such as unstructured data such as technical documentation of the laser device 10 that has not been learned, and structured data such as operational data of the laser device 10.

[0102] Since the AAP 136 acquires external information each time, the LLM 134 can generate a query item response W5 that includes the latest external information.

[0103] The laser management server 110A according to the first embodiment can generate a query response W6 in response to a query W0 from various users such as an FSE, a device owner, a research and development engineer, etc. This makes it possible to provide a laser management server 110A that meets the needs of various users.

[0104] The first query item W1 is used by the AAP 136 to acquire and analyze external information and create an agent response W3. The first query item W1, which requires external information, is then transmitted to the LLM 134 in the form of the agent response W3. Therefore, in the laser management server 110A according to the first embodiment, the amount of information received by the LLM 134 is reduced compared to when all external information is transmitted to the LLM 134, and the LLM 134 is prevented from exceeding its prompt reception capacity.

[0105] Furthermore, the laser management server 110A of the first embodiment does not require a variety of standard operation screens 124, and therefore can reduce the time and cost required for the UI development process compared to the laser management server 110 according to the comparative example.

[0106] 4. Modification of the First Embodiment 4.1 Configuration Fig. 7 is a diagram showing a laser management system 100B according to a modified example of the first embodiment. Differences between the configuration shown in Fig. 7 and Fig. 3 will be described below. The laser management server 110B shown in Fig. 7 differs from the laser management server 110A shown in Fig. 3 in that the QIP 132 and the LLM 134, and the AAP 136 and the LLM 134 communicate with each other. The other configurations may be the same as those shown in Fig. 3.

[0107] 4.2 Operation 8 is a diagram showing the operation flow of laser management server 110B according to a modification of embodiment 1. Differences between the operation flow of laser management server 110B shown in FIG. 8 and FIG. 4 will be described.

[0108] The QIP 132 of the laser management server 110B may generate a selection prompt W7 when performing language processing to generate the first query item W1 and the second query item W2. The selection prompt W7 is, for example, "break down the query into items."

[0109] The LLM 134 may receive the screening prompt W7 from the QIP 132 and generate a screening response W8 that is a response to the screening prompt W7. The screening response W8 may be, for example, an itemized query item as shown in FIG.

[0110] The QIP 132 may receive a screening response W8 from the LLM 134.

[0111] When AAP 136 obtains unstructured data and creates agent response W3, it may generate unstructured data prompt W9, which may include, for example, the laser device 10 manual and a prompt such as, "Where can I find instructions on how to replace the chamber?"

[0112] The LLM 134 may receive the unstructured data prompt W9 from the AAP 136 and generate an unstructured data response W10 that is a response to the unstructured data prompt W9. For example, the unstructured data response W10 may be, "The method for replacing the chamber is described from line BB on page AA to line DD on page CC."

[0113] The AAP 136 may receive the unstructured data response W10 from the LLM 134. The AAP 136 analyzes the unstructured data response W10 received from the LLM 134 and generates the agent response W3. For example, when the AAP 136 receives the unstructured data response W10, "The method for replacing the chamber is described from line BB on page AA to line DD on page CC," the AAP 136 allocates an appropriate external information processing program and generates the agent response W3 from the results. Other operations may be similar to those shown in FIG. 4.

[0114] 4.3 Actions and Effects Laser management server 110B provides the same effects as laser management server 110A. Laser management server 110B also provides improved language processing capabilities in QIP 132 and AAP 136.

[0115] 5. Embodiment 2 5.1 Configuration Fig. 9 is a diagram showing a laser management system 100C according to embodiment 2. Differences between the configuration shown in Fig. 9 and Fig. 7 will be described.

[0116] The laser management system 100C differs from the laser management system 100B in that an AI prediction processing device 108 is connected to a laser management server 110C.

[0117] The AI ​​prediction processing device 108 includes a life prediction model for predicting the life of consumables of the laser device 10 and a laser performance prediction model for predicting future laser performance of the laser device 10. The life prediction model and the laser performance prediction model are learned AI models trained by machine learning to perform target tasks, and are implemented within the AI ​​prediction processing device 108. Other configurations may be similar to those in FIG. 7.

[0118] 5.2 Operation The lifespan prediction model may be, for example, a trained model created by the machine learning method described in Patent Document 3. The machine learning method described in Patent Document 3 is a machine learning method for creating a learning model for predicting the lifespan of a consumable of laser device 10, and includes the steps of acquiring first lifespan-related information including data on lifespan-related parameters of the consumable recorded corresponding to different numbers of oscillation pulses during a period from when the consumable is first used to when it is replaced, dividing the first lifespan-related information into multiple levels representing the degree of deterioration of the consumable according to the numbers of oscillation pulses, and creating training data in which the first lifespan-related information is associated with the levels representing the degree of deterioration, performing machine learning using the training data to create a learning model that predicts the degree of deterioration of the consumable from the data on the lifespan-related parameters, and saving the created learning model.

[0119] The learning model is, for example, a neural network model, and is essentially a program that causes a computer to execute a process for predicting the degree of deterioration of consumables of the laser device 10.

[0120] The AI ​​prediction processing device 108 can predict the lifespan of each consumable scheduled for replacement in the laser device 10 by using a corresponding lifespan prediction model based on the lifespan-related information of the consumable. The prediction result based on the lifespan prediction model is sent from the AAP 136 to the ROP 138 as an agent response W3.

[0121] The laser performance prediction model is a trained model that can predict future changes in the laser performance of the laser device 10 according to the number of pulses or date and time in any part replacement scenario. The prediction results by the laser performance prediction model are transmitted from the AAP 136 to the ROP 138 as an agent response W3.

[0122] In the ROP 138, a query response W6 is constructed based on the agent response W3 and the query item response W5, and the query response W6 is displayed on the interactive operation screen 126 of the terminal 150.

[0123] Fig. 10 is a diagram showing an example of a query response W6 displayed on the terminal 150 in the second embodiment. Fig. 10 is a display example of the user's reply as the query response W6 when the query W0 is "Please report the operation diagnosis of laser 65400011 in the standard format." Differences between Fig. 10 and Fig. 6 will be described.

[0124] 10, compared to FIG. 6, prediction results for the spectral bandwidth (spectral linewidth) of the pulsed laser beam and the gas pressure in the chamber 24 of the OSC 20 are added. That is, in FIG. 10, the interactive operation screen 126 includes an LLM generation result 159 and an AAP generation result 160. The LLM generation result 159 is a result generated by the LLM 134. The AAP generation result 160 is a result obtained when the AAP 136 uses the laser performance prediction model of the AI ​​prediction processing device 108. The AAP generation result 160 may include a graph, a table, or the like. Note that the spectral bandwidth of the pulsed laser beam or the gas pressure in the chamber 24 of the OSC 20 is an example of the "laser device performance" in this disclosure.

[0125] 5.3 Actions and Effects The laser management system 100C according to the second embodiment can achieve the same effects as those of the first embodiment. Furthermore, the laser management system 100C can generate the agent response W3 using information on the predicted results of the lifespan of each consumable item to be replaced and the predicted results of the laser performance of the laser device 10, which are obtained from the AI ​​prediction processing device 108. This improves the quality of the query response W6 to the prediction-related query W0.

[0126] 6.Other The above description is intended to be illustrative rather than limiting. Thus, it will be apparent to one skilled in the art that modifications can be made to the disclosed embodiments without departing from the scope of the claims. It will also be apparent to one skilled in the art that the disclosed embodiments can be used in combination.

[0127] Terms used throughout this specification and claims should be construed as "open ended" unless expressly stated otherwise. For example, words such as "comprise," "have," "comprise," and "equip" should be construed as meaning "without excluding the presence of elements other than those listed." In addition, the modifier "a" should be construed as meaning "at least one" or "one or more." In addition, the term "at least one of A, B, and C" should be construed as "A," "B," "C," "A+B," "A+C," "B+C," or "A+B+C." Furthermore, it should be construed to include combinations of these with elements other than "A," "B," and "C."

Claims

1. A laser management server for a laser device, a transceiver processor that receives a query from an external device and transmits a query response to the query to the external device; a query input processor that receives the query from the transceiver processor, decomposes the query, and generates a first query term that requires external information and a second query term that does not require external information; an agent action processor that receives the first query item from the query input processor, acquires necessary external information from unstructured data including a manual and a maintenance report of the laser device and structured data including operation data of the laser device, and generates an agent response that is a response to the first query item; a response output processor that receives the second query term from the query input processor and the agent response from the agent action processor, and generates a query term prompt related to the second query term; a large scale language model processor that receives the query item prompt from the response output processor and generates a query item response that is a response to the query item prompt; The response output processor receives the query item response from the large scale language model processor, constructs the query response based on the agent response and the query item response, and transmits the query response to the transceiver processor.

2. 2. The laser management server of claim 1, the query item prompt includes the agent response; Laser management server.

3. 2. The laser management server of claim 1, the agent action processor acquires the structured data including a prediction result of a lifespan of a consumable of the laser device according to a lifespan prediction model; Laser management server.

4. 2. The laser management server of claim 1, the agent action processor acquires the structured data including a predicted result of performance of the laser device according to a laser performance prediction model; Laser management server.

5. 2. The laser management server of claim 1, the query input processor generates screening prompts that decompose the query when processing the first query term and the second query term; the large scale language model processor receives the screening prompt from the query input processor and generates a screening response that is a response to the screening prompt; the query input processor receives the filtered response from the large scale language model processor; Laser management server.

6. 2. The laser management server of claim 1, the agent action processor generates an unstructured data prompt when retrieving the unstructured data and creating the agent response; the large scale language model processor receives the unstructured data prompt from the agent action processor and generates an unstructured data response in response to the unstructured data prompt; The agent action processor receives the unstructured data response from the large scale language model processor.

7. 7. The laser management server of claim 6, The unstructured data prompt includes the unstructured data.

8. 2. The laser management server of claim 1, The response output processor combining the information generated by the large scale language model processor and the agent response received from the agent action processor to form the query response; Laser management server.

9. 2. The laser management server of claim 1, the transceiver processor receives the query from a terminal connected via a network, and transmits the query response to the terminal; Laser management server.

10. 10. The laser management server of claim 9, receiving the query from an interactive operation screen of the terminal; the query response is displayed on an interactive operation screen of the terminal. Laser management server.

11. A laser management method for a laser device, comprising: A first step in which a transceiver processor receives a query from an external device; a second step in which a query input processor decomposes the query received from the transceiver processor to generate a first query term requiring external information and a second query term not requiring external information; a third step in which an agent action processor receives the first query item from the query input processor, acquires necessary external information from unstructured data including a manual and a maintenance report of the laser device and structured data including operation data of the laser device, and generates an agent response that is a response to the first query item; a fourth step in which a response output processor receives the second query term from the query input processor, receives the agent response from the agent action processor, and generates a query term prompt related to the second query term; a fifth step in which a large scale language model processor receives the query item prompt from the response output processor and generates a query item response that is a response to the query item prompt; a sixth step in which the response output processor receives the query item response from the large scale language model processor and constructs a query response based on the agent response and the query item response; a seventh step in which the transceiver processor receives the query response from the response output processor and transmits the query response to an external device; A laser management method comprising:

12. 12. The laser management method of claim 11, the third step includes the agent action processor acquiring the structured data including a prediction result of a lifespan of a consumable of the laser device according to a lifespan prediction model; Laser management methods.

13. 12. The laser management method of claim 11, the third step includes the agent action processor acquiring the structured data including a prediction result of laser performance of the laser device according to a laser performance prediction model; Laser management methods.

14. 12. The laser management method of claim 11, the third step includes, when performing language processing to generate the first query term and the second query term, generating a screening prompt that decomposes the query into the first query term and the second query term, the large-scale language model processor receiving the screening prompt from the query input processor, and generating a screening response that is a response to the screening prompt, the query input processor receiving the screening response from the large-scale language model processor; Laser management methods.

15. 12. The laser management method of claim 11, generating an unstructured data prompt when the agent action processor retrieves the unstructured data and creates the agent response; the large scale language model processor receives the unstructured data prompt from the agent action processor and generates an unstructured data response that is a response to the unstructured data prompt; the agent action processor receiving the unstructured data response from the large scale language model processor; Laser management methods.

16. 16. The laser management method of claim 15, the unstructured data prompt includes the unstructured data; Laser management methods.

17. 12. The laser management method of claim 11, the query item prompt includes the agent response; Laser management methods.

18. 12. The laser management method of claim 11, the sixth step includes the response output processor combining information generated by the large scale language model processor and the agent response received from the agent action processor to construct the query response; Laser management methods.

19. 12. The laser management method of claim 11, the first step includes the transceiver processor receiving the query from a terminal connected via a network; the seventh step includes transmitting the query response to the terminal; Laser management methods.

20. 20. The laser management method of claim 19, comprising: receiving the query from an interactive operation screen of the terminal; the query response is displayed on an interactive operation screen of the terminal. Laser management methods.

Images