Information processing device, information processing method, and article manufacturing method

Abstract

To reduce time to analyze a cause of an abnormality in an apparatus.SOLUTION: An information processing apparatus accumulates logs in a substrate processing apparatus that processes a substrate, and has: an accumulation unit that accumulates the logs; a control unit that controls to select specific logs from the logs accumulated in the accumulation unit; and a transmission unit that transmits the logs selected by the control unit to the outside. When an error occurs in the substrate processing apparatus, the control unit selects, from the logs accumulated in the accumulation unit, logs in which a degree of correlation exceeds a threshold on the basis of a learned model.SELECTED DRAWING: Figure 4

Description

[Technical field]

[0001] The present invention relates to an information processing device, an information processing method, and an article manufacturing method. [Background technology]

[0002] Exposure apparatuses have been used for some time when manufacturing display elements such as liquid crystal panels and semiconductor elements using photolithography technology. Exposure apparatuses can transfer a pattern drawn on an original (e.g., a mask or a reticle) by projecting the pattern onto a substrate (e.g., a glass substrate or a wafer).

[0003] An exposure apparatus is composed of a wide variety of drive units and measurement units, and a huge amount of data related to the exposure process is accumulated as a log. When an abnormality such as an apparatus failure or a decrease in performance related to the exposure process occurs, the accumulated logs are used to analyze the cause of the abnormality. Patent Document 1 discloses a method for determining an apparatus failure by generating a diagnostic model, which is a trained model for diagnosing the failure from the collected logs, and diagnosing the failure of the apparatus. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2019-213192 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, when a deterioration in performance related to the exposure process occurs, the operator who operates the exposure apparatus needs to select a log related to the abnormal state and send it to the analyst. As the number of types of logs and the amount of data accumulated increase with the sophistication of the exposure apparatus, selecting a log related to the abnormality of the apparatus requires specialized knowledge and experience. As a result, there is a risk that the operator will send a log that is not highly related to the abnormality to the analyst, making it difficult for the analyst to analyze the abnormality.

[0006] In the above cases, the analyst requests the operator to send the necessary additional logs, and the operator must resend the logs in response to the request, which results in a significant amount of time spent analyzing the cause of the anomaly. One possible method would be to send all accumulated logs (or more logs than are deemed necessary) to the analyst, but there may be circumstances in which it is not possible to send many logs to the analyst due to restrictions such as network bandwidth or customer security. Alternatively, there is a risk that a significant amount of time will be spent sending a huge amount of data.

[0007] SUMMARY OF THE PRESENT EMBODIMENTS Accordingly, an object of the present invention is to provide an information processing device which is advantageous in shortening the time required for analyzing the cause of an abnormality in the device. [Means for solving the problem]

[0008] In order to achieve the above object, an information processing apparatus according to one aspect of the present invention is an information processing apparatus that accumulates logs in a substrate processing apparatus that processes a substrate, the information processing apparatus including: a learning unit that creates a learned model by learning a correlation degree indicating a correlation with an error of the substrate processing apparatus from information of the log stored in the storage unit; and a control unit that controls the selection of a specific log from the logs stored in the storage unit, and a transmission unit that transmits the log selected by the control unit to an external device, the learning unit updates a learning model so that a degree of correlation of a log selected by an operator among the logs stored in the storage unit is increased, and updates the learning model so that a degree of correlation of a log not selected by the operator is decreased; When an error occurs in the substrate processing apparatus, the control unit The above Based on the trained model, among the logs stored in the storage unit, ,phase The method is characterized in that logs whose relevance exceeds a threshold are selected. Effect of the Invention

[0009] According to the present invention, it is possible to provide an information processing device that is advantageous in shortening the time required for analyzing the cause of an abnormality in the device. [Brief description of the drawings]

[0010] [Figure 1] 1 is a schematic diagram showing an information processing system according to a first embodiment. [Diagram 2] FIG. 2 is a diagram showing a learning model in the first embodiment. [Diagram 3] FIG. 4 is a diagram showing a learning screen in the first embodiment. [Figure 4] FIG. 13 is a flowchart showing log collection. [Diagram 5] FIG. 11 is a diagram showing a learning model in the second embodiment. [Figure 6] FIG. 11 is a diagram showing a learning screen in the second embodiment. [Figure 7] FIG. 13 is a diagram illustrating a network configuration in a third embodiment. [Figure 8] FIG. 13 is a schematic diagram showing an information processing system according to a fourth embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to refer to the same components, and duplicated explanations will be omitted.

[0012] First Embodiment FIG. 1 is a diagram showing an information processing system 10 in this embodiment. The information processing device 100 in this embodiment is applied to a substrate processing device (semiconductor manufacturing device) such as a pattern forming device that forms a pattern on a substrate. In the following description, an example of processing a log (processing data) of an exposure device that exposes a substrate to light and forms a latent image pattern will be described as the substrate processing device, but the present invention is not limited to this and may be applied to other substrate processing devices. The information processing device 100 in this embodiment can be applied to, for example, a coating device that coats a photosensitive medium or the like on the surface of a substrate, a developing device that develops a substrate to which a latent image pattern has been transferred, an inspection device that inspects the formation accuracy of a pattern, as well as an etching device and a film forming device.

[0013] The information processing system 10 in this embodiment includes an information processing device 100 including an operation unit 101, a data processing unit 103, a storage unit 104, a transmission unit 105, and a control unit 106. The control unit 106 controls each unit of the information processing device 100. The information processing system 10 also includes a unit control unit 110 including a sequence controller 111, a drive controller 112, a drive unit 113, a measurement controller 114, and a measurement unit 115.

[0014] The unit control unit 110 controls the drive controller 112 and the measurement controller 114 by the sequence controller 111 performing sequence control. The drive controller 112 controls the drive of the drive unit 113, and the measurement controller 114 controls the measurement in the measurement unit 115, thereby executing the exposure process of the exposure apparatus and various measurements required for the exposure process. Drive information of the drive unit 113 is collected by the drive controller 112, and measurement information of the measurement unit 115 is collected by the measurement controller 114. These pieces of information are notified to the information processing device 100 via the sequence controller 111.

[0015] In the information processing device 100, information notified from the unit control unit 110 is processed by the data processing unit 103 and stored in the accumulation unit 104 as a log for each unit such as the drive unit 113 and the measurement unit 115. Similarly, sequence execution information of the sequence controller 111 is stored in the accumulation unit 104 as a log.

[0016] An operator in charge of operating the exposure apparatus can select a specific log from the logs accumulated in the accumulation unit 104 by performing a log selection operation on the operation unit 101. The log selected by the operator can be transmitted by the transmission unit 105 to an external device 120 outside the apparatus itself. The selected log is used, for example, when an error occurs in the exposure apparatus, and is useful for analyzing the cause of the abnormality in the exposure apparatus. By transmitting the log to the external device 120, it becomes possible for an analyst located away from the factory where the exposure apparatus is installed to analyze the cause of the abnormality in the exposure apparatus.

[0017] However, as the exposure apparatus becomes more sophisticated, the types of logs and the amount of data accumulated are increasing, and selecting logs related to an error in the apparatus requires specialized knowledge and experience. In such a case, the analyst needs to request the operator to send additional necessary logs, and the operator needs to send the logs again in response to the request, which takes a lot of time to analyze the cause of the abnormality. In addition, a method of sending all the accumulated logs (or more logs than are considered necessary) to the analyst can be considered, but there may be circumstances in which it is not possible to send many logs to the analyst due to restrictions such as network bandwidth and customer security. Or, there is a risk that a lot of time will be taken due to the transmission of a huge amount of data.

[0018] Therefore, in the information processing device 100 of this embodiment, an optimal log is selected from the logs accumulated in the accumulation unit 104 without any operation by the operator, and is transmitted to the external device 120 by the transmission unit 105. This makes it possible to transmit logs necessary for analyzing the cause of an abnormality in the exposure device, regardless of the knowledge or experience of the operator. Furthermore, the information processing device 100 of this embodiment further includes a learning unit 102 for performing learning in order to select the optimal log. The contents of the learning performed by the learning unit 102 will be described below.

[0019] The learning performed by the learning unit 102 may be a method using so-called machine learning or a neural network, or a method using simple weighting. An example using weighting will be described below, but the learning method is not limited to this example. A learning table 200 corresponding to the type of error that occurs in the exposure apparatus is stored in the storage unit 104, and the learning unit 102 can update (learning, which will be described below) the learning table 200 stored in the storage unit 104. Note that, although it is assumed in this embodiment that the learning table 200 is stored in the storage unit 104, it may also be stored in the learning unit 102.

[0020] FIG. 2 is a diagram showing a learning table 200, which is information stored in the accumulation unit 104. The learning table 200 is composed of a plurality of learning models 201a to 201d corresponding to types of errors. The learning models 201a to 201d are composed of an ID 202 indicating the type of error, correlations 203a and 203b of logs to errors, and a reliability 204 of the learning model. The correlations 203a and 203b of logs and the reliability 204 are expressed as values ​​between 0 and 1. The correlations 203a and 203b of logs indicate the degree of association of logs corresponding to abnormal events, with higher values ​​indicating higher correlation. The reliability 204 indicates whether automatic collection of logs can be performed accurately, with higher values ​​indicating more accurate automatic collection of logs.

[0021] When an error occurs due to a unit failure or the like, the error ID is used as ID 202, and when a performance degradation or the like occurs without an error occurring, an arbitrarily determined unique ID corresponding to that event is used. Also, a learning model 201 having a reliability equal to or higher than a reference value is called a learned model that can be used for automatic collection of logs.

[0022] The learning method in this embodiment will be described with reference to FIG. 3. FIG. 3 is a diagram showing a learning screen 300 of the operation unit 101 for an operator to input information necessary for learning in the learning unit 102. The learning screen 300 displays an event name 301, an event occurrence time 302, a log selection screen 303, and a decision button 304. In the event name 301, the name of the type of error that has occurred and the corresponding error ID are automatically input according to the error that has occurred. If the error that has occurred is not an apparatus failure or the like but a deterioration in performance of the exposure process or the like, the error cannot be detected automatically, so the operator must input the event rather than automatically input it. Similarly to the above, if the error that has occurred is an apparatus failure or the like, the event occurrence time 302 is automatically input according to the error that has occurred, but if the error is a deterioration in performance of the exposure process or the like, the operator must input it. In addition, the event name 301 and the event occurrence time 302 can be freely edited by the operator.

[0023] The log selection screen 303 displays a list of logs stored in the storage unit 104, and each log has a check box. The operator selects a specific related log according to the type of error entered in the event name 301, or an ID indicating the type of error. Multiple logs may be selected. After the selection on the log selection screen 303 is completed, the operator presses the decision button 304 to execute learning. As learning is executed, the learning model 200 in FIG. 2 is updated. By repeating learning, logs can be automatically collected with higher accuracy.

[0024] Here, the log correlation and reliability of the learning model 200 in Fig. 2 will be described. The initial values ​​of the log correlation and reliability are 0, and the values ​​of the log correlation and reliability are updated by the above learning. The log correlation is calculated by the following formula (1). Log correlation = log selection (-1: not selected / 1: selected) × learning rate + log correlation before update (1)

[0025] If the result of the calculation using formula (1) is 0 or less, it is updated to 0, and if it is 1 or more, it is updated to 1. The selection of a log is determined by whether it is selected on log selection screen 303 in FIG. 3. The learning rate is a value between 0 and 1, is preset, and can be changed by operation from operation unit 101. The closer the learning rate is to 1, the greater the change in value from the log correlation degree before the previous update in one learning. The closer the learning rate is to 0, the less the change in value from the log correlation degree before the update.

[0026] The reliability is calculated using the following formula (2). Reliability = learning rate × reliability learning coefficient + reliability before update (2)

[0027] If the result of the calculation using formula (2) is 0 or less, it is updated to 0, and if it is 1 or more, it is updated to 1. The reliability learning coefficient is a value between 0 and 1, is preset, and can be changed by operation from the operation unit 101. If the reliability exceeds the reference value, the learning model can be used for automatic collection of logs as a trained model. For example, when the learning rate is 0.1, the learning coefficient is 1, and the reference value is 0.5, it is determined that the model is a trained model by performing learning five times from an unlearned state.

[0028] Next, a method of learning by the learning unit 102 in this embodiment and collection of logs using a learned model will be described with reference to the flowchart in Fig. 4. Fig. 4(a) describes a case where an error related to an apparatus failure occurs, and Fig. 4(b) describes a case where an error related to a deterioration in performance of the exposure process occurs.

[0029] When an error related to a device failure is detected, in step S401, an error notification is displayed on the operation unit 101. In step S402, it is determined whether learning related to the occurred error has been completed (whether there is a trained model related to the occurred error). Specifically, it is determined whether the trained model related to the occurred error is a trained model whose reliability exceeds a reference value. If there is a trained model related to the occurred error, the process proceeds to step S403, and if there is no trained model related to the occurred error, the process proceeds to step S404.

[0030] Here, the selection of the log will be specifically described with reference to FIG. 2. It is assumed that the reference value of the reliability is 0.5. When an error corresponding to the abnormal event ID 1 or 2 occurs, the learning models 201a and 201b are first referred to. The reliability of the learning model 201a is 0.4, and the reliability of the learning model 201b is 0.3, both of which are below the reference value. This indicates that the learning has not been performed a sufficient number of times, and the model has not become a trained model. In this case, the process proceeds to step S404. On the other hand, when an error corresponding to the abnormal event ID 3 or 4 occurs, the learning models 201c and 201d are first referred to. The reliability of the learning model 201c is 0.7, and the reliability of the learning model 201b is 0.9, both of which are above the reference value. This indicates that the learning has been performed a sufficient number of times, and the model has become a trained model. In this case, the process proceeds to step S403.

[0031] In step S403, a specific log is automatically selected using a trained model for the error that has occurred. The selection of a log is determined based on whether or not the value of log correlation 403 exceeds a threshold. If the log correlation exceeds the threshold, a log is selected, and if the log correlation is below the threshold, a log is not selected. Since the amount of logs selected varies depending on the threshold, the threshold may be set in advance so that a required amount of logs can be selected. After the selection of a specific log is performed, the result of the log selection is displayed on the operation unit 101.

[0032] Here, the selection of logs will be specifically described with reference to FIG. 2. It is assumed that the threshold value of the log correlation is 0.5. When an error corresponding to abnormal event ID 3 or 4 occurs, the learning models 201c and 201d are first referred to. In the learning model 201c, the correlation degree of log A is 0.1 and the correlation degree of log B is 0.6, so that only the correlation degree of log B exceeds the threshold. This indicates that log A is not related to the error corresponding to abnormal event ID 3, whereas log B is related to the error corresponding to abnormal event ID 3. In this case, log B is automatically selected, and log A is not selected. On the other hand, in the learning model 201d, the correlation degree of log A is 0.7 and the correlation degree of log B is 0.7, so that the correlation degrees of both log A and log B exceed the threshold. This indicates that both log A and log B are related to the error corresponding to abnormal event ID 4. In this case, both log A and log B are automatically selected.

[0033] In step S404, the learning screen 200 shown in FIG. 3 is displayed on the operation unit 201 in order to execute log learning in the learning unit 102. In step S405, the operator selects and collects logs. The learning method is the same as that described above. In step S406, the learning unit 102 updates the learning model 401.

[0034] In step S407, the selected specific log is transmitted to the outside by the transmission unit 105. The analyst who analyzes the abnormality of the device can give the operator necessary instructions to resolve the abnormality by analyzing the transmitted log.

[0035] If the error occurring in the exposure apparatus is not due to a malfunction or the like, but is due to a deterioration in the performance of the exposure process, processing is performed according to the flowchart in Fig. 4(b). Here, the performance of the exposure process is an important performance index of the exposure apparatus, such as overlay accuracy, focus accuracy, throughput, etc. These deteriorations in the performance of the exposure process are difficult to detect by the exposure apparatus alone, and therefore are not automatically detected. In other words, an error notification is not automatically displayed as in step S401 in Fig. 4(a).

[0036] In step S411, a learning screen as shown in Fig. 3 is displayed on the operation unit 101. The operator inputs the type of event name related to the deterioration of performance in the exposure process (for example, overlay accuracy, focus accuracy, throughput, etc.) and an ID for associating with the event. Unlike when an apparatus breaks down, this requires the intervention of the operator. The operator also inputs the time when the event occurred. Steps S412 to S417 are similar to steps S402 to S407 described above, and therefore description thereof will be omitted.

[0037] In the flowcharts of FIG. 4(a) and FIG. 4(b), the flow of automatically selecting logs using a trained model has been described, but if necessary, the trained model may be subjected to additional training. Additional training is required, for example, when the accuracy of automatic log selection is low. If the amount of logs sent is less than the required amount due to low accuracy of automatic log selection, it is difficult for the analyst to analyze the cause of the anomaly. Conversely, if the amount of logs sent is more than the required amount, the amount of data to be sent will be large and the burden on the analyst will be heavy, so it is desirable to be able to select appropriate logs by additional training. In the case of additional training, a learning rate different from that of normal training may be used. By performing this additional training, more accurate training can be performed.

[0038] As described above, in the information processing device 100 of this embodiment, by using the learned model, an optimal log is selected from the logs accumulated in the accumulation unit 104 without any operation by the operator, and is transmitted to the external device 120 by the transmission unit 105. As a result, regardless of the knowledge or experience of the operator, the log required to analyze the cause of an abnormality in the exposure device can be transmitted with high accuracy, thereby shortening the time required to analyze the cause of the abnormality.

[0039] <Second embodiment> In the first embodiment, the log acquisition period was set in advance, but in this embodiment, in addition to the learning in the first embodiment, the log acquisition period is also learned. Note that the information processing device 100 and the system configuration are the same as in the first embodiment, so the description will be omitted. Also, matters not mentioned in this embodiment follow the first embodiment.

[0040] The period of logs required for analysts to analyze the logs varies depending on the type of log and the type of error that has occurred. For example, log A requires 30 seconds of data before the abnormal event occurs, but log B may require only 10 seconds. The longer the period of logs to be acquired, the larger the data volume will be, so it is desirable to acquire logs for an appropriate period.

[0041] The learning model used in this embodiment will be described with reference to Fig. 5. Fig. 5 shows an example in which an item of an acquisition period 501 is added to the learning model 200 described in Fig. 2. The acquisition period 501 is set in seconds.

[0042] The flow of learning is the same as in the first embodiment, but in addition to learning the reliability and log correlation degree performed in the first embodiment, learning of the acquisition period is also performed. Fig. 6 adds a window for inputting the log acquisition period corresponding to each log on the log selection screen 303 of the learning screen 300 in Fig. 3. When the operator inputs the log acquisition period, the learning unit 202 also performs learning on the log acquisition period. The log acquisition period is updated for each log by the calculation of the following formula (3). Acquisition period = (input value - acquisition period before update) × acquisition period learning rate + acquisition period before update (3)

[0043] The acquisition period learning rate is preset to a value between 0 and 1, similar to the learning rate of the log correlation. However, it is managed separately from the learning rate of the log correlation, and the value can be set by the operation unit 101. For logs whose log correlation exceeds the threshold, only the logs for the acquisition period learned by the learning unit 201 are transmitted by the transmission unit 105 to the outside of the device.

[0044] A specific example in which the acquisition period is updated by the calculation of formula (3) will be described. Assume that the acquisition period learning rate is 0.5, the input value by the operator is 110 seconds, and the acquisition period before update is 100 seconds. From formula (3), the acquisition period is updated to 105. The closer the acquisition period learning rate is to 1, the greater the change in value from the acquisition period before update after one learning. The closer the acquisition period learning rate is to 0, the less the change in value from the acquisition period before update.

[0045] In this embodiment, additional learning may be performed on a trained model whose reliability exceeds a reference value, as in the first embodiment. For example, if an analyst determines that the automatically acquired log period is too short to perform analysis, additional learning may be performed to extend the acquisition period.

[0046] As described above, in the information processing device 100 of this embodiment, by using a trained model, logs are acquired from the logs accumulated in the accumulation unit 104 at an optimal period without operation by the operator, and are transmitted to the external device 120 by the transmission unit 105. As a result, regardless of the knowledge or experience of the operator, logs necessary for analyzing the cause of an abnormality in the exposure device can be transmitted with high accuracy, thereby shortening the time required for analyzing the cause of the abnormality.

[0047] <Third embodiment> In the first embodiment, the contents of applying a trained model to the apparatus itself to automatically acquire a log were described. That is, when there are multiple exposure apparatuses, it was assumed that each exposure apparatus would independently perform learning and log collection. In this embodiment, the contents of applying a trained model trained in one exposure apparatus to another exposure apparatus to automatically acquire a log will be described. Note that the configuration of the information processing apparatus 100 and the system is the same as in the first embodiment, and therefore description thereof will be omitted. Also, matters not mentioned in this embodiment will follow the first embodiment.

[0048] 7 is a diagram showing a network configuration in this embodiment. In this embodiment, exposure apparatuses 701a to 701c are connected to a shared server 700. The exposure apparatuses 701a to 701c may be connected to the shared server 700 by wired communication or wireless communication. In this embodiment, a trained model trained by the exposure apparatus 701a is uploaded to the shared server 700. Alternatively, a log of the exposure apparatus 701a may be uploaded to the shared server 700 and trained by the shared server 700.

[0049] When collecting logs, exposure apparatus 701b and exposure apparatus 701c first check whether there is a trained model in their own apparatus, and if there is not, make an inquiry to learning shared server 700. If there is a trained model corresponding to the corresponding abnormal event in shared server 700, the trained model in learning shared server 700 is acquired, and exposure apparatus 701b and exposure apparatus 701c automatically collect logs. If there is no trained model corresponding to the corresponding abnormal event in shared server 700, learning is performed within the own apparatus as in the first embodiment. At that time, if a trained model has been created, the trained model is uploaded to shared server 700.

[0050] Also, when additional learning is performed in exposure apparatus 701b or exposure apparatus 701c, selected information of the log for the error that occurred is uploaded to shared server 700. Shared server 700 updates the trained model based on information from exposure apparatus 701b or exposure apparatus 701c. It then transmits the trained model to connected exposure apparatuses 701a to 701c. Exposure apparatuses 701a to 701c replace the trained model in their own apparatus with the latest trained model based on the trained model transmitted from shared server 700.

[0051] As described above, in the information processing device 100 of this embodiment, by using a trained model, logs are acquired from the logs accumulated in the accumulation unit 104 for an optimal period without operation by an operator, and are transmitted to the external device 120 by the transmission unit 105. This makes it possible to transmit logs necessary for analyzing the cause of an abnormality in the exposure device, regardless of the knowledge or experience of the operator.

[0052] As described above, in the information processing device 100 of this embodiment, by using the trained model, the optimal log is acquired from the logs accumulated in the accumulation unit 104 without the intervention of an operator's operation. This makes it possible to transmit the log necessary to analyze the cause of an abnormality in the exposure device with high accuracy, regardless of the operator's knowledge or experience, thereby shortening the time required for analyzing the cause of the abnormality. In this embodiment, even in an exposure device that does not have a trained model, it is possible to automatically acquire logs by using a trained model trained in another device, so that even when a new exposure device is installed, the time required for analyzing the cause of the abnormality can be shortened.

[0053] <Fourth embodiment> In this embodiment, an example in which an external device 120 outside the exposure apparatus controls the settings and learning of the information processing device 100 will be described with reference to Fig. 8. Fig. 8 shows the contents of the system shown in Fig. 1 to which a receiving unit 800 for receiving information from an external device is added inside the information processing device 100. Also, an operation unit 801 and a learning unit 802 are added inside the external device 120.

[0054] The external device 120 in this embodiment may be installed outside the factory where the exposure apparatus is operating, and can be operated via a network even outside the country. An operation unit 801 in the external device 120 can perform log acquisition, log learning, and the like in the same way as the operation unit 101 in the information processing apparatus. Furthermore, a learning unit 802 can perform learning within the external device 120. The results of learning can also be transmitted to other devices via a network, as in the second embodiment. Therefore, there is no need for an operator to directly operate the exposure apparatus, and an analyst outside the factory can directly perform operations such as log acquisition and learning.

[0055] <Embodiments of processing items> The method for manufacturing an article according to an embodiment of the present invention is suitable for manufacturing, for example, a flat panel display (FPD). The method for manufacturing an article according to this embodiment includes a step of forming a latent image pattern on a photosensitive agent applied on a substrate using the above-mentioned exposure apparatus (a step of exposing the substrate), and a step of developing the substrate on which the latent image pattern has been formed in this step. Furthermore, this manufacturing method includes other well-known steps (oxidation, film formation, deposition, doping, planarization, etching, resist stripping, dicing, bonding, packaging, etc.). The method for manufacturing an article according to this embodiment is advantageous in at least one of the performance, quality, productivity, and production cost of the article compared to conventional methods.

[0056] <Other embodiments> The present invention can also be realized by a process in which a program for implementing one or more of the functions of the above-described embodiments is supplied to a system or device via a network or a storage medium, and one or more processors in a computer of the system or device read and execute the program. The present invention can also be realized by a circuit (e.g., ASIC) for implementing one or more of the functions.

[0057] Although the preferred embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various modifications and changes are possible within the scope of the gist of the present invention. [Explanation of symbols]

[0058] 100 Information processing device 104 Storage Unit 105 Transmitter 106 Control section

Claims

1. An information processing device that accumulates logs in a substrate processing device that processes substrates, a storage unit for storing logs; a learning unit that creates a learned model by learning a correlation degree indicating a correlation with an error of the substrate processing apparatus from information of the log stored in the storage unit; and a control unit that controls to select a specific log from the logs stored in the storage unit; a transmission unit that transmits the log selected by the control unit to an outside, the learning unit updates a learning model so that a degree of correlation of a log selected by an operator among the logs stored in the storage unit is increased, and updates the learning model so that a degree of correlation of a log not selected by the operator is decreased; The information processing device is characterized in that, when an error occurs in the substrate processing device, the control unit selects, based on the learned model, logs stored in the storage unit that have a correlation degree exceeding a threshold.

2. The learning unit causes a learning model to learn a period of a log required for analysis based on an anomaly that occurs; The information processing apparatus according to claim 1 , wherein the transmission unit transmits a log for a period during which the learning model is learned.

3. The information processing apparatus according to claim 1 or 2, wherein the transmission unit transmits the trained model to an external device so that the trained model can be used in a substrate processing apparatus other than the substrate processing apparatus.

4. 2. The information processing apparatus according to claim 1, wherein the learned model is a model created by having a learning model learn a correlation between log information of a substrate processing apparatus other than the substrate processing apparatus and an abnormality in the other substrate processing apparatus from the log information of the other substrate processing apparatus.

5. 5. The information processing apparatus according to claim 1, further comprising a receiving section for receiving information from an external device other than the substrate processing apparatus.

6. 1. An information processing method for accumulating a log in a substrate processing apparatus that processes a substrate, comprising: a log accumulating step; a learning process for generating a trained model by learning a correlation degree indicating a correlation with an error of the substrate processing apparatus from the log information accumulated in the accumulation process; a control step of controlling to select a specific log from the logs accumulated in the accumulation step; a transmission step of transmitting the log selected in the control step to an outside, The learning step updates a learning model so that a correlation degree of a log selected by an operator among the logs accumulated in the accumulation step is increased, and updates the learning model so that a correlation degree of a log not selected by the operator is decreased, The information processing method is characterized in that, when an error occurs in the substrate processing apparatus, the control process selects, based on the learned model, logs accumulated in the accumulation process that have a correlation degree exceeding a threshold value.

7. A program for causing a computer to execute each step of the information processing method according to claim 6.

8. An information processing device according to any one of claims 1 to 5, and a pattern forming means for transferring a pattern of an original onto a substrate to form a pattern on the substrate.

9. an exposure step of exposing a substrate using the pattern forming apparatus according to claim 8; A developing step of developing the substrate exposed in the exposure step, A method for manufacturing an article, comprising the steps of: manufacturing an article from the substrate developed in the developing step;

Images