Sample preparation and evaluation system, control method for the sample preparation and evaluation system, program, and recording medium.

Abstract

This technology provides advantages in quickly obtaining the optimal conditions for producing high-quality samples. [Solution] The sample preparation and evaluation system comprises a sample preparation device that prepares a sample based on preparation conditions, a measuring device that measures the sample, and an information processing device that evaluates the sample based on the measurement results of the measuring device and acquires the evaluation results of the sample. If the evaluation results do not meet predetermined conditions, the information processing device notifies that correction of the evaluation results is necessary.

Description

Technical Field

[0001] This disclosure relates to a technique for preparing a sample and evaluating the prepared sample.

Background Art

[0002] In the development of materials and the like, a sample is prepared by selecting raw materials, determining the blending amounts of the raw materials, and performing processes such as stirring or heating the materials. Also, in development, the prepared sample is measured and the measurement results are evaluated. The preparation conditions of the sample are updated so that this evaluation satisfies the target conditions. In such development, data-driven development that focuses on improving the performance of materials and development efficiency is known. Data-driven development derives the preparation conditions under which a desired characteristic value can be obtained from, for example, the preparation conditions of a sample and the data of the measured characteristic values.

[0003] Specifically, determine the preparation conditions of the sample, prepare the sample under those preparation conditions, measure the prepared sample to obtain characteristic values, perform data analysis using the preparation conditions and measurement results of the samples obtained so far, and determine the preparation conditions of the next sample. This series of operations is called a search cycle, and the search cycle is repeatedly executed to derive the preparation conditions under which a desired characteristic value can be obtained.

[0004] Data analysis techniques such as Bayesian optimization are often used to determine the preparation conditions of the sample, and the preparation and measurement of the sample were often performed manually.

[0005] Patent Document 1 discloses automating a series of search cycles by automatically performing the preparation and measurement of a sample using a robot or the like.

Prior Art Documents

Patent Documents

[0006]

Patent Document 1

Summary of the Invention

[0007] In such systems, it is desirable to obtain the conditions for producing high-quality samples in a short period of time. [Means for solving the problem]

[0008] This disclosure provides a technology that is advantageous for obtaining production conditions that produce high-quality samples in a short period of time.

[0009] A first aspect of this disclosure is a sample preparation and evaluation system comprising: a sample preparation apparatus for preparing a sample based on preparation conditions; a measuring apparatus for measuring the sample; and an information processing apparatus for evaluating the sample based on the measurement results of the measuring apparatus and obtaining an evaluation result of the sample, wherein the information processing apparatus notifies that correction of the evaluation result is necessary if the evaluation result does not meet predetermined conditions.

[0010] A second aspect of the present disclosure is a sample preparation and evaluation system comprising: a sample preparation apparatus for preparing a sample based on preparation conditions; a first measuring apparatus for measuring the sample; a second measuring apparatus for measuring the sample with higher accuracy than the first measuring apparatus; and an information processing apparatus for evaluating the sample based on the measurement results of the first measuring apparatus and acquiring the evaluation results of the sample, wherein if the evaluation results do not satisfy predetermined conditions, the information processing apparatus causes the second measuring apparatus to measure the sample, evaluates the sample based on the measurement results of the second measuring apparatus, and updates the evaluation results of the sample.

[0011] A third aspect of this disclosure is a sample preparation and evaluation system comprising: a sample preparation apparatus for preparing a sample based on preparation conditions; a measuring device for measuring the sample; and an information processing device for evaluating the sample based on the measurement results of the measuring device and obtaining an evaluation result of the sample, wherein the information processing device, if at least one first evaluation result obtained from at least one first measurement result of the sample prepared according to the preparation conditions satisfies predetermined conditions, causes the sample preparation apparatus to prepare the sample again, has the measuring device measure the again prepared sample and obtains at least one second measurement result from the measuring device, obtains at least one second evaluation result based on the at least one second measurement result, and updates the preparation conditions based on the at least one second evaluation result.

[0012] A fourth aspect of this disclosure is a control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, characterized in that the method comprises: a step of the sample preparation device preparing a sample based on preparation conditions; a step of the measuring device measuring the sample; a step of the information processing device evaluating the sample based on the measurement results of the measuring device and obtaining an evaluation result of the sample; and a step of the information processing device notifying that if the evaluation result does not meet predetermined conditions, the evaluation result needs to be corrected.

[0013] A fifth aspect of this disclosure is a control method for a sample preparation and evaluation system having a sample preparation device, a first measuring device, a second measuring device having higher measurement accuracy than the first measuring device, and an information processing device, the method comprising: a step of the sample preparation device preparing a sample based on preparation conditions; a step of the first measuring device measuring the sample; a step of the information processing device evaluating the sample based on the measurement results of the first measuring device and obtaining the evaluation results of the sample; a step of the second measuring device measuring the sample if the evaluation results do not satisfy predetermined conditions; and a step of the information processing device evaluating the sample based on the measurement results of the second measuring device and updating the evaluation results of the sample.

[0014] A sixth aspect of the present disclosure is a control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, characterized in that the method comprises: a step of the sample preparation device preparing a sample based on preparation conditions; a step of the measuring device measuring the sample and obtaining a first measurement result; a step of the information processing device obtaining a first evaluation result based on the first measurement result; a step of the sample preparation device preparing the sample again if the first evaluation result satisfies predetermined conditions; a step of the measuring device measuring the again prepared sample and obtaining a second measurement result; a step of the information processing device obtaining a second evaluation result based on the second measurement result; and a step of the information processing device updating the preparation conditions based on the second evaluation result. [Brief explanation of the drawing]

[0015] [Figure 1] (a) is a block diagram of the sample preparation and evaluation system according to the first embodiment. (b) is a block diagram of the information processing device according to the first embodiment. [Figure 2] (a) is an explanatory diagram of the information processing unit according to the first embodiment. (b) is an explanatory diagram showing an example of training data used in machine learning according to the first embodiment. [Figure 3] This is a flowchart of the sample preparation and evaluation method using the sample preparation and evaluation system according to the first embodiment. [Figure 4] (a) is an explanatory diagram showing an example of a user interface image according to the first embodiment. (b) is an explanatory diagram showing the evaluation results according to the first embodiment. [Figure 5] This is an explanatory diagram of the user interface image according to the first embodiment. [Figure 6] This is a diagram illustrating the estimation function according to the first embodiment. [Figure 7] This is a diagram illustrating the estimation function according to the first embodiment. [Figure 8](a) is an explanatory diagram showing an example of an image displayed on the display device according to the first embodiment. (b) is an explanatory diagram showing an example of an image displayed on the display device according to the first embodiment. [Figure 9] It is an explanatory diagram showing an example of an image displayed on the display device according to the first embodiment. [Figure 10] It is a block diagram of a sample preparation evaluation system according to the second embodiment. [Figure 11] (a) is a schematic perspective view of a sample according to the second embodiment. (b) is an explanatory diagram of a shape error according to the second embodiment. [Figure 12] (a) is a schematic side view of a sample and a measuring device for measuring the sample according to the third embodiment. (b) is an explanatory diagram showing an example of preparation conditions when preparing the sample according to the third embodiment. [Figure 13] (a) is a graph showing an example of a first measurement result of a sample according to the third embodiment. (b) is a graph showing the difference between the discharge amount of the sample and the target discharge amount according to the third embodiment. (c) is a graph showing an example of a second measurement result of the sample according to the third embodiment. (d) is a graph showing the difference between the discharge amount of the sample and the target discharge amount according to the third embodiment. [Figure 14] It is a flowchart of a sample preparation evaluation method by a sample preparation evaluation system according to the third embodiment. [Figure 15] (a) is a graph showing an example of measurement results of a plurality of samples according to the third embodiment. (b) is a graph showing the difference between the average discharge amount of each of the plurality of samples and the target discharge amount according to the third embodiment. [Figure 16] (a) is a graph showing an example of measurement results of a plurality of samples according to the third embodiment. (b) is a graph showing the difference between the average discharge amount of each of the plurality of samples and the target discharge amount according to the third embodiment.

Mode for Carrying Out the Invention

[0016] Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the drawings. The embodiments shown below are illustrative, and for example, the details of the configuration can be modified as appropriate by those skilled in the art without departing from the spirit of the present disclosure. In the drawings referenced in the following description of embodiments, elements indicated by the same reference numerals have the same function unless otherwise specified.

[0017] [First Embodiment] Figure 1(a) is a block diagram of the sample preparation and evaluation system 100 according to the first embodiment. The sample preparation and evaluation system 100 comprises a sample preparation device 101, a measuring device 102, an information processing device 103, and a display device 104.

[0018] The sample preparation and evaluation system 100 automatically prepares a sample based on preparation conditions during the development of materials, automatically evaluates the sample, and repeatedly performs a series of search cycles to determine the next preparation conditions from a dataset of preparation conditions and evaluation results. By executing the search cycle multiple times, the sample preparation and evaluation system 100 searches for preparation conditions that can produce the desired sample.

[0019] Here, the preparation conditions are the conditions necessary when preparing a sample, and include material conditions such as the type, properties, and quantity of the material, and processing conditions such as stirring, heating, cooling, and curing of the material. In other words, the preparation conditions may include information on the material used to prepare the sample, information on the proportions of the material used to prepare the sample, and information on the processing temperature (e.g., curing temperature) during the sample preparation process. Any other conditions necessary when preparing the sample may also be included in the preparation conditions. The preparation conditions are command values ​​(target values) instructed to the sample preparation device 101.

[0020] The evaluation items evaluated by the information processing unit 105 are, for example, sensory evaluation items such as the appearance of the sample. There may be one type of evaluation item or multiple types. If there are multiple types of evaluation items, they may be optimized simultaneously. The information processing unit 103 determines the next production conditions by analyzing a dataset that correlates production conditions with evaluation results. Bayesian optimization may be used to determine the next production conditions, but other optimization methods such as response surface analysis, regression analysis, or genetic algorithms may also be used.

[0021] The sample preparation device 101 automatically prepares a sample under given preparation conditions in each of multiple exploration cycles. The information processing device 103 evaluates the sample prepared by the sample preparation device 101. The information processing device 103 is configured to have one or more computers. The following explanation will use the case where the information processing device 103 has one computer, i.e., one processor, as an example.

[0022] Figure 1(b) is a block diagram of an information processing device 103 according to the first embodiment. The information processing device 103 includes a CPU 301, which is an example of a processor; a RAM 302, which is a temporary storage device; a ROM 303 and an SSD 304, which are non-temporary storage devices (recording media); a recording disk drive 305; and an interface 306 for I / O, etc. The non-temporary storage device is an example of a recording media that can be read by a computer, and the non-temporary storage device stores a program 307 that causes the CPU 301 to execute information processing methods such as control processing and arithmetic processing that control each part of the entire device. The recording disk drive 305 can read data recorded on a recording disk 308, which is an example of a recording media.

[0023] In the first embodiment, the non-temporary recording medium readable by the computer is, for example, an SSD 304, and the program 307 is recorded on the SSD 304, but it is not limited to this. The program 307 may be recorded on any recording medium that is a non-temporary recording medium readable by the computer. Examples of recording media that can be used to supply the program 307 to the computer include flexible disks, hard disks, optical disks, magneto-optical disks, magnetic tapes, non-volatile memory, etc. Furthermore, the program 307 may be obtained from a network (not shown).

[0024] Furthermore, the information processing device 103 having a processor may be configured in ways other than those described above, such as a PLD (Programmable Logic Device) including an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), a general-purpose or dedicated computer with a program built in, or a combination of all or part of these.

[0025] The CPU 301 of the information processing device 103 functions as the information processing unit 105 shown in Figure 1(a) by executing the program 307. The information processing unit 105 performs a learning process to perform machine learning, an evaluation process to perform inference based on the trained machine learning model and obtain an evaluation result for the sample as an inference result, and a condition determination process to determine the manufacturing conditions used for preparing the sample from the evaluation result. The information processing unit 105 then outputs control commands, which are the manufacturing conditions, to the sample preparation device 101.

[0026] The sample preparation device 101 prepares the sample based on the preparation conditions corresponding to the control commands. Sample preparation involves processes such as weighing and dispensing the material, stirring the material, degassing, heating and cooling, pressurizing and unloading, and hardening. In the sample preparation device 101, sample preparation is performed using, for example, a robot or an automated machine.

[0027] In the first embodiment, the measuring device 102 is a visual sensor, which is an example of a sensory sensor. The visual sensor is, for example, a digital camera, and generates an image of the sample prepared by the sample preparation device 101 as visual information, which is an example of sensory information.

[0028] The evaluation items evaluated by the information processing unit 105 include, for example, a rank related to the amount of damage contained in the image of the sample, and the information processing unit 105 determines the manufacturing conditions that reduce the amount of damage.

[0029] Figure 2(a) is an explanatory diagram of the information processing unit 105 according to the first embodiment. The information processing unit 105 includes a learning unit 106 and an evaluation unit 107. The learning unit 106 performs machine learning processing, and the evaluation unit 107 performs evaluation processing and condition determination processing.

[0030] The learning unit 106 performs supervised learning as a form of machine learning. The learning unit 106 performs supervised machine learning using training data T1 which contains image data and correct answer data, and generates a trained machine learning model M1. The machine learning model M1 is stored, for example, in SSD 304 in Figure 1(b). The evaluation unit 107 performs an evaluation process by using the trained machine learning model M1 to perform inference on the input data, image IM2, and outputs the evaluation result E2, which is the inference result.

[0031] Image IM2 is digital data such as an image acquired from the measuring device 102. The measuring device 102 generates image IM2 by imaging the sample prepared by the sample preparation device 101.

[0032] Figure 2(b) is an explanatory diagram showing an example of training data T1 used in machine learning according to the first embodiment. The training data T1 includes a plurality of images IM1 and a plurality of rank values ​​A1 corresponding to each of the plurality of images IM1. The user (operator) creates the training data T1 by classifying each of the plurality of images IM1 into one of a plurality of ranks from rank A to rank G, according to the amount of damage to the sample captured in each of the plurality of images IM1 that was prepared in advance. In other words, each image IM1 is assigned a rank value A1 corresponding to one of the ranks from rank A to rank G as the correct answer data. Thus, the training data to be trained by the learning unit 106 is created by the user through sensory inspection performed by the user. Image IM1 is digital data such as captured images.

[0033] Among the various ranks, rank A represents the highest evaluation in the sensory evaluation, and rank G represents the lowest evaluation. In other words, ranks A, B, C, D, E, F, and G represent increasing levels of sensory evaluation in this order. A sample that appears in rank A is the sample that received the best evaluation in the sensory evaluation.

[0034] Multiple ranks can be quantified. For example, rank A can be assigned a numerical value such as "0" and rank G a numerical value such as "1", allowing each of the multiple ranks to be represented numerically.

[0035] Figure 3 is a flowchart of the sample preparation and evaluation method, i.e., the control method of the sample preparation and evaluation system 100, according to the first embodiment. Steps S2 onward shown in Figure 3 constitute a search cycle, which is repeated until the termination condition is met in step S6. Data input by the user to the information processing device 103 is performed via a graphical user interface (user interface image). Figure 4(a) is an explanatory diagram showing an example of the user interface image UI1 according to the first embodiment.

[0036] First, in step S1, the evaluation unit 107 displays a user interface image UI1 on the display device 104 and accepts the user's setting of search conditions via the user interface image UI1. The user can input the search conditions into the user interface image UI1 using an input device such as a mouse (not shown) or a keyboard (not shown). The display device 104 may be a touch panel display equipped with an input device.

[0037] The search conditions will now be explained in detail. The search conditions only need to be defined in order to operate the sample preparation device 101, and are not limited to those described below. The search conditions include search settings related to preparation conditions, search settings related to termination conditions, and search settings related to the overall operation.

[0038] The search settings related to the fabrication conditions define the search range and other parameters corresponding to the search items. The search items are the items of the fabrication conditions to be searched. For example, these include the type of material used to prepare the sample, the ratio of materials, and the curing temperature.

[0039] The search range is the range within which the manufacturing conditions corresponding to the search item can be varied. For example, if the search item is something that can be expressed numerically, such as material ratio or curing temperature, then the search range is set as one or more numerical values, or a continuous range of numerical values. In the example in Figure 4(a), the curing temperature range is set to 40-120°C and the material ratio range is set to 0.8-2.0. In this case, the manufacturing conditions are determined as follows: the curing temperature is determined from within the range of 40-120°C, and the material ratio is determined from within the range of 0.8-2.0.

[0040] Furthermore, the search range may also be non-numerical information, such as materials E001, E002, and E003. In this case, the manufacturing conditions are determined from materials E001, E002, and E003. In this way, the desired sample is searched for by changing the manufacturing conditions within the search range.

[0041] The overall search settings define the termination conditions and options for the search cycle. The termination condition for the search cycle may be that it ends after a specified number of cycles have been repeated, or it may be that it ends when the evaluated rank reaches a specified rank.

[0042] The evaluation unit 107 determines the preparation conditions based on the search conditions and outputs a control command to the sample preparation device 101. The initial preparation conditions may be the lower or upper limit of the search range, or random values.

[0043] In step S2, the evaluation unit 107 outputs a control command, which is the production condition, to the sample preparation device 101, and the sample preparation device 101 prepares the sample based on the production conditions corresponding to the control command.

[0044] Next, in step S3, the evaluation unit 107 causes the measuring device 102 to measure the sample. In this embodiment, since the measuring device 102 is a visual sensor, the visual sensor is made to image the sample, and an image IM2, which is an example of the measurement result of the measuring device 102 (visual information which is sensory information), is obtained from the measuring device 102.

[0045] Next, in step S4, the evaluation unit 107 evaluates the sample based on the image IM2, which is the measurement result of the measuring device 102, and obtains the evaluation result E2 of the sample. In this embodiment, the evaluation unit 107 obtains the evaluation result E2 for the image IM2, which is the measurement result, using a trained machine learning model M1. In other words, in the evaluation process of the information processing unit 105 of the information processing device 103, the sensory evaluation of the sample is not performed by the user, but is automatically performed by the information processing unit 105 of the information processing device 103 using the machine learning model M1.

[0046] Figure 4(b) is an explanatory diagram of the evaluation result E2 according to the first embodiment. The evaluation result E2 includes multiple rank values ​​A20. The evaluation unit 107 selects the evaluation value E20 from among the multiple rank values ​​A20. That is, one of the multiple rank values ​​A20 is the evaluation value E20.

[0047] Furthermore, the evaluation result E2 includes multiple probabilities P20, each corresponding to a multiple rank value A20. The evaluation unit 107 sets the rank value corresponding to the highest first probability P1 among the multiple probabilities P20 as the evaluation value E20. For example, suppose the evaluation result E2 has probabilities of being rank A (1%), rank B (2%), rank C (8%), rank D (86%), rank E (6%), rank F (2%), and rank G (1%). In this case, rank D, which has the highest probability, becomes the evaluation value E20. That is, the first probability P1, which has the highest probability, corresponds to rank D.

[0048] Next, in step S5, the evaluation unit 107 determines whether the evaluation result E2 satisfies predetermined conditions. In this embodiment, the evaluation unit 107 compares the difference ΔP (=P1-P2) between the top two probabilities P1 and P2 with a preset threshold. That is, the predetermined condition is that the difference ΔP (=P1-P2) between the first probability P1 and the second probability P2, which is the next highest probability among the multiple probabilities P20, is greater than or equal to the threshold. The threshold is, for example, 10%. Note that the predetermined conditions are not limited to the example described above.

[0049] If the evaluation result E2 does not meet the predetermined conditions, that is, if the difference ΔP is not equal to or greater than the threshold (if the difference ΔP is less than the threshold) (step S5: NO), in step S7, the evaluation unit 107 notifies the user that the evaluation result E2 (specifically the evaluation value E20) needs to be corrected. At that time, the evaluation unit 107 assigns a label to the evaluation value E20 indicating that re-evaluation is necessary, and saves the evaluation value E20 in a predetermined area of ​​the SSD 304, linked to the label.

[0050] Figure 5 is an explanatory diagram of the user interface image UI2 according to the first embodiment. In this embodiment, in step S7, the evaluation unit 107 notifies the user by displaying an image on the display device 104 indicating that the evaluation result E2 needs to be corrected, such as the user interface image UI2. Although the method of notifying the user is described as an image, it is not limited to this, and the user may also be notified by email, voice, etc.

[0051] To illustrate with a specific example, if the second highest probability of achieving rank C is P2 = 46%, and the highest probability of achieving rank D is P1 = 52%, then the difference ΔP is 52% - 46% = 6%, which is less than 10%. Therefore, the evaluation unit 107 proceeds to step S7. In this case, the evaluation value E20 is output as the value for rank D, which has the highest probability.

[0052] In step S8, the evaluation unit 107 determines whether or not it has accepted a modification to the evaluation value E20 in the user interface image UI2. If step S8 is YES, that is, if a modification to the evaluation value E20 has been accepted, in step S11, the evaluation unit 107 updates the evaluation value E20 to the accepted modification and proceeds to the process in step S9. If step S8 is NO, that is, if a modification to the evaluation value E20 has not been accepted, the evaluation unit 107 proceeds directly to the process in step S9.

[0053] Here, we will explain the process of step S7 in detail. In step S7, the evaluation unit 107 displays a user interface image UI2, such as the one shown in Figure 5, on the display device 104 and notifies the user that re-evaluation is necessary, that is, that the evaluation result E2 (evaluation value E20) needs to be corrected, with text such as "There is data that needs to be re-evaluated."

[0054] User interface image UI2 includes buttons B1 for "Yes," B2 for "No," and B3 for "Later." Buttons B1-B3 are user-operable buttons.

[0055] When the "Yes" button B1 is pressed, the evaluation unit 107 accepts the modification of the labeled evaluation value E20 stored in a predetermined area of ​​the SSD 304. When the "No" button B2 is pressed, the evaluation unit 107 erases the label associated with the evaluation value E20 and proceeds to the next step S8. When the "Later" button B3 is pressed, the evaluation unit 107 proceeds directly to the next step S8.

[0056] Thus, in the first embodiment, the evaluation unit 107 displays a user interface image UI2 on the display device 104 that accepts user modifications to the evaluation value E20 of the evaluation result E2, making it possible to easily modify the evaluation value E20.

[0057] In step S9, the evaluation unit 107 calculates an estimation function using the evaluation value E20. The estimation function is a function in which, for example, the manufacturing condition parameters (e.g., curing temperature) are independent variables and the evaluation value is the dependent variable. If the evaluation unit 107 receives a correction to the evaluation value E20, it calculates an estimation function using the corrected evaluation value E20.

[0058] Bayesian optimization is preferably used to create or update the estimation function. Bayesian optimization is a method that sequentially and probabilistically searches for the minimum (or maximum) value of a function, updating the function sequentially as the amount of data increases and searching for the minimum value.

[0059] In step S10, the evaluation unit 107 updates the fabrication conditions using the estimation function and returns to the process in step S2. Then, in step S2 from the second round onward of the search cycle shown in the flowchart in Figure 3, the evaluation unit 107 outputs a control command, which is the updated fabrication condition, to the sample preparation device 101, and the sample preparation device 101 prepares the sample based on the updated fabrication condition. In this way, in step S2, the sample preparation device 101 prepares the sample based on the updated fabrication condition each time the fabrication condition is updated.

[0060] Furthermore, if the difference ΔP in step S5 is greater than or equal to the threshold, i.e., if step S5 is YES, the evaluation unit 107 determines whether the termination condition is met. If the termination condition is not met, i.e., if step S6 is NO, the evaluation unit 107 proceeds to the processing of step S8. If the termination condition is met, i.e., if step S6 is YES, the evaluation unit 107 terminates the process.

[0061] In this way, by repeating the search cycle shown in Figure 3 multiple times, multiple evaluation values ​​E20 are obtained, and an estimation function is obtained using these multiple evaluation values ​​E20. Of the multiple evaluation values ​​E20, those that have been assigned labels become subject to modification by the user.

[0062] Figure 6 is a diagram illustrating the estimation function according to the first embodiment. In the first round of the search cycle in the flowchart shown in Figure 3, as shown in the left diagram of Figure 6, there is only one set of data points (108) for the manufacturing condition parameters and evaluation values, and an estimation function 109 is obtained in which the evaluation value is constant. Then, from this estimation function 109, the next manufacturing condition parameters 110 are obtained within the range of the search conditions set in step S1.

[0063] In the second round of the search cycle shown in the flowchart in Figure 3, as shown in the right-hand figure of Figure 6, the set data of the manufacturing condition parameters and evaluation values ​​are two points 108 and 111, and an estimated function 112 that passes through these two points 108 and 111 is obtained. That is, the estimated function 109 is updated to the estimated function 112. Then, from this estimated function 112, the next manufacturing condition parameter 113 is obtained within the range of the search conditions set in step S1.

[0064] In the first embodiment, the search cycle, i.e., sample preparation, sample measurement, and updating of sample preparation conditions, continues even if the evaluation values ​​have not been corrected. This is because, in this embodiment, optimization is performed to successively find better preparation conditions as the dataset of preparation condition parameters and evaluation values ​​increases, and for this purpose, it is necessary to increase the number of datasets. If the search cycle is stopped until the evaluation values ​​are re-evaluated by the user (i.e., corrected), the dataset will not increase, and as a result, it will take time to determine the sample preparation conditions.

[0065] The re-evaluation of evaluation value E20 is performed independently of the sample preparation, sample measurement, and updating of preparation conditions, i.e., the exploration cycle. For example, if the exploration cycle is performed automatically overnight, there may be cases where a user corrects two or more evaluation values ​​labeled for re-evaluation all at once in the morning.

[0066] Figure 7 is a diagram illustrating the estimation function according to the first embodiment. In the left diagram of Figure 7, we will explain the case in which two evaluation values ​​114 and 115, out of two or more evaluation values ​​labeled for re-evaluation, are modified. From the estimation function 121, the manufacturing condition parameter 119 that results in the minimum evaluation value can be obtained.

[0067] The two evaluation values ​​114 and 115 are, for example, assigned to rank D and modified to evaluation values ​​116 and 117, which correspond to rank C, a rank with a very small probability difference according to the user. By modifying the labeled evaluation values ​​114 and 115 to evaluation values ​​116 and 117, the estimation function 121 shown in the left figure of Figure 7 is updated to the estimation function 118 shown in the right figure of Figure 7. This makes it possible to make the minimum evaluation value in the estimation function 118 shown in the right figure of Figure 7 smaller than the minimum evaluation value in the estimation function 121 shown in the left figure of Figure 7. Then, the production condition parameter 120 that yields the minimum evaluation value can be obtained from the estimation function 118 shown in the right figure of Figure 7. In this way, the production condition parameter 120 that improves the sensory evaluation results compared to the production condition parameter 119 can be obtained.

[0068] In this way, by replacing evaluation values ​​114 and 115 with evaluation values ​​116 and 117, the following production conditions change. Therefore, it becomes possible to obtain production conditions that match the termination conditions of the search conditions, i.e., production conditions that produce high-quality samples, in a short period of time.

[0069] The image IM2 corresponding to the corrected evaluation value E20 can be used as training data, and the learning unit 106 may be subjected to machine learning again using image IM2 and evaluation value E20 to update the machine learning model M1. This improves the accuracy of ranking. Furthermore, all past evaluation values ​​E20 may be re-evaluated by the evaluation unit 107.

[0070] Furthermore, if the evaluation value is modified by the user, i.e., if the evaluation value is re-evaluated by the user, the evaluation unit 107 may display the image IM3 shown in Figure 8(a) on the display device 104 to notify the user of the estimated function before and after re-evaluation. Alternatively, the evaluation unit 107 may display the image IM4 shown in Figure 8(b) on the display device 104 to notify the user of the manufacturing conditions before and after re-evaluation.

[0071] Furthermore, the evaluation unit 107 may repeat the above series of search cycles multiple times and display the image IM5 shown in Figure 9 on the display device 104 to notify the user of the changes in the evaluation value E20 (search result). The evaluation unit 107 may then display the manufacturing conditions that resulted in the best evaluation value E20 at that time on the image.

[0072] Furthermore, even if the termination conditions shown in Figure 4(a) are met, if there are remaining datasets that require re-evaluation, it may be possible to derive even better production conditions, so a function that does not satisfy the termination conditions may be added.

[0073] Furthermore, if the number of data points requiring re-evaluation increases too much, the likelihood of low-rated samples being continuously produced increases. Therefore, if the number of evaluation results E2 (evaluation value E20) requiring correction reaches a predetermined number, the sample preparation device 101 may, under the command of the evaluation unit 107, stop preparing samples, for example, by interrupting sample preparation. Also, sample preparation and measurement may be performed one at a time or multiple samples simultaneously.

[0074] Furthermore, when improving multiple types of measurements simultaneously, you can create individual graphs for each measurement as shown in Figure 9, or you can create a single indicator that combines the quality of multiple measurements, such as hypervolume.

[0075] As described above, according to the first embodiment, the evaluation value E20, which is the evaluation result E2, can be modified at the user's discretion, and the estimation function is updated based on the result, and the next production conditions are determined. As a result, the number of evaluation results (data) necessary for optimization is obtained, and after the evaluation results that need to be modified are corrected to high-precision evaluation results, the search for the next production conditions can be performed with high precision, and even if the measurement results are ambiguous sensory information such as image IM2, production conditions that match the search conditions can be determined in a short period of time. In other words, production conditions that produce high-quality samples can be obtained in a short period of time.

[0076] [Example 1] In the first embodiment described above, the sensory sensor was a visual sensor, and the amount of damage to the sample was automatically evaluated using the image IM2. However, the method is not limited to this. It can also be applied to cases where the sensory sensor measures aroma, taste, texture, the balance of colors related to the appearance of the dish, and design aspects such as shape.

[0077] Taste can be measured using a taste sensor as the measuring device 102 (sensory sensor). For the training data, the user should prepare the measurement items measured by sensory measurement in advance, and the numerical measurement results such as bitterness, sweetness, sourness, and saltiness in those cases, and then the learning unit 106 should be trained using this training data.

[0078] The measuring device 102 may be configured to automatically measure the sample, and the evaluation unit 107 may have the sample preparation device 101 prepare the sample, have the measuring device 102 measure the sample, and then use the trained machine learning model M1 to evaluate the rank of the sample measurement result.

[0079] Then, the evaluation unit 107 assigns a re-evaluation label to evaluation values ​​E20 for which the probability difference ΔP is determined to be less than the threshold. Subsequently, the user re-evaluates the labeled evaluation values ​​E20, updating the estimation function and determining the next production conditions.

[0080] This allows us to gather the necessary data for optimization, and after obtaining highly accurate measurement results, we can then search for the next manufacturing conditions with high accuracy, enabling us to determine the manufacturing conditions in a short period of time.

[0081] Although the explanation used the example of a taste sensor, the explanation is not limited to this, and the sensor may also be an olfactory sensor or a pressure sensor. Furthermore, the sensor is not limited to one type; for example, the sensor may be at least one of a visual sensor, an olfactory sensor, a taste sensor, and a pressure sensor. For example, an olfactory sensor can be used to detect aroma, and a pressure sensor can be used to detect texture.

[0082] [Second Embodiment] A second embodiment will now be described. Hereinafter, elements denoted by the same reference numerals as those in the first embodiment will have substantially the same configuration and function as those described in the first embodiment unless otherwise specified. The differences from the first embodiment will be the main focus of this description.

[0083] In the first embodiment, a case where the user corrects the evaluation results was described, but in the second embodiment, a case where another measuring device remeasures the sample will be described.

[0084] Figure 10 is a block diagram of the sample preparation and evaluation system 100A according to the second embodiment. The sample preparation and evaluation system 100A comprises a sample preparation device 101, a measuring device 102A, a measuring device 102B, an information processing device 103, and a display device 104. In the second embodiment, measuring devices 102A and 102B are provided instead of the measuring device 102 of the first embodiment. Each of the measuring devices 102A and 102B is a device for measuring a sample.

[0085] Although measuring device 102A has lower measurement accuracy than measuring device 102B, the time required for measurement is shorter than that of measuring device 102B. Although measuring device 102B has higher measurement accuracy than measuring device 102A, the time required for measurement is longer than that of measuring device 102A. In other words, measuring device 102B can measure the sample with higher accuracy than measuring device 102A. Measuring device 102A is an example of a first measuring device, and measuring device 102B is an example of a second measuring device.

[0086] In the second embodiment, the information processing unit 105 replaces the measurement result of the low-precision measuring device 102A with the measurement result obtained by re-measuring with the high-precision measuring device 102B, and determines the sample preparation conditions.

[0087] Figure 11(a) is a schematic perspective view of sample W1 according to the second embodiment. Sample W1 is, for example, an optical element. Each of the measuring devices 102A and 102B is configured to measure the surface shape of sample W1 prepared by the sample preparation device 101. That is, the information processing unit 105 is configured to acquire the measured surface shape of sample W1 from the measuring devices 102A and 102B.

[0088] Figure 11(b) is an explanatory diagram of the shape error 126 according to the second embodiment. The shape error 126 is the difference between the designed shape of the surface of sample W1 and the measured shape of the surface of sample W1. The shape error 126 is the evaluation result, and the PV (peak to valley) value 127 of the shape error 126 is the evaluation value. The PV value 127 is the maximum value of the shape error 126. The smaller the evaluation value PV value 127, the better the product.

[0089] In the second embodiment, the information processing unit 105 has the measuring device 102A measure the sample preferentially, evaluates the sample W1 based on the measurement results of the measuring device 102A, and obtains the evaluation result of the sample W1. If the shape error 126, which is the evaluation result, does not meet the predetermined conditions, the information processing unit 105 has the measuring device 102B measure the sample W1, evaluates the sample W1 based on the measurement results of the measuring device 102B, and updates the evaluation result of the sample W1.

[0090] In the second embodiment, the predetermined condition is that the PV value 127 is equal to or greater than a threshold. That is, if the PV value 127 derived from the measurement result of the measuring device 102A is less than the threshold, the information processing unit 105 has the measuring device 102B measure the sample W1, evaluates the sample W1 based on the measurement result of the measuring device 102B, and updates the evaluation result of the sample W1. Note that the predetermined condition is not limited to the above example.

[0091] Thus, if the measurement result is above the threshold, the information processing unit 105 does not perform a remeasurement because the possibility of the sample being a good product is low, and if the measurement result is below the threshold, it performs a remeasurement because the possibility of the sample being a good product is high.

[0092] Furthermore, the information processing unit 105 is configured to obtain an estimation function using the evaluation value PV value 127 and to update the manufacturing conditions using the estimation function. The estimation function has the same configuration as in the first embodiment.

[0093] When the evaluation value is updated, i.e., when a remeasurement is performed, the information processing unit 105 calculates the estimation function using the updated evaluation value. That is, the information processing unit 105 updates the estimation function using the PV value 127 which has been replaced with the remeasured value, and determines the next manufacturing conditions.

[0094] In the second embodiment, a large number of samples W1 are measured in a short time using the measuring device 102A. This reduces the number of samples W1 that need to be measured using the measuring device 102B, which requires more measurement time, and shortens the period for determining the preparation conditions for sample W1. Preparation conditions that produce high-quality samples can be obtained in a short period of time.

[0095] [Third Embodiment] A third embodiment will now be described. In the following description, elements denoted by the same reference numerals as those in the above embodiments will have substantially the same configuration and function as those described in the above embodiments unless otherwise specified, and the differences from the above embodiments will be described primarily.

[0096] The first embodiment describes a case where the user modifies the evaluation results, and the second embodiment describes a case where another measuring device remeasures the sample. In the third embodiment, a case is described in which the evaluation value that exceeds the threshold for the target value is left as is, and the sample is prepared again using only the preparation conditions corresponding to the evaluation value within the range defined by the threshold, and the evaluation value is updated based on the results of the remeasurement.

[0097] The sample preparation and evaluation system according to the third embodiment is the same as that of the first embodiment. In the third embodiment, the information processing unit 105 re-prepares the sample and re-measures the sample, and determines the sample preparation conditions that satisfy predetermined conditions based on the measurement results.

[0098] Figure 12(a) is a schematic side view of a sample L1 and a measuring device 102 for measuring sample L1 according to the third embodiment. Sample L1 is, for example, a droplet ejected from the ejection head of an inkjet printer. The measuring device 102 is used to measure the ejection volume and ejection speed of the prepared sample L1 and includes, for example, an imaging system having a light source 102a and a high-speed camera 102b. The measuring device 102 is configured to image the sample L1 as it is ejected from the ejection head of the inkjet printer and in flight. That is, the information processing unit 105 controls the light source 102a to irradiate the flying sample L1 with light and controls the camera 102b to cause the camera 102b to image the flying sample L1, thereby acquiring an image from the camera 102b. The information processing unit 105 is configured to process the image obtained from the camera 102b and calculate the ejection volume and ejection speed of sample L1. As another example of a method for measuring the discharge amount, the information processing unit 105 may use the light source 102a and the camera 102b to image the sample L1, which is a droplet that has landed on the evaluation substrate, detect the shape of the droplet from the image obtained from the camera 102b, and calculate the discharge amount from the shape of the droplet. In this embodiment, the control of the measuring device 102, image acquisition, and image processing are performed by the information processing unit 105, but this is not limited to this. For example, the measuring device 102 may be an independent measuring device equipped with an information processing unit for the measuring device, and the data of the discharge amount and discharge speed measured by the measuring device may be transferred to the information processing unit 105.

[0099] Figure 12(b) is an explanatory diagram showing an example of the preparation conditions when preparing sample L1 according to the third embodiment. The preparation conditions for sample L1 are the drive waveform for driving the pressure generation mechanism of the discharge head. In the graph shown in Figure 12(b), the horizontal axis is time, and the vertical axis is the voltage applied to the pressure generation mechanism. The drive of the pressure generation mechanism is controlled by the drive waveform shown in Figure 12(b).

[0100] Figure 13(a) is a graph showing an example of the measurement result 128 of sample L1 according to the third embodiment. In the graph shown in Figure 13(a), the horizontal axis represents the discharge speed of sample L1, and the vertical axis represents the discharge volume of sample L1. By measuring sample L1 using the measuring device 102, the discharge volume 128a and discharge speed 128b of sample L1 are obtained as the measurement result 128. The discharge volume 128a and discharge speed 128b are examples of measured values. The measurement result 128 is an example of the first measurement result.

[0101] Figure 13(b) is a graph showing the difference between the discharge rate 128a and the target discharge rate 129 of sample L1 according to the third embodiment. In the graph shown in Figure 13(b), the horizontal axis represents the manufacturing condition parameters, and the vertical axis represents the discharge rate. The target discharge rate 129 is an example of a set target value. A threshold 131 is set with ΔP0 as the range, based on the target discharge rate 129. The threshold 131 is a threshold value based on the target discharge rate 129. The value of │ΔPv│ 130, which is the difference between the discharge rate 128a and the target discharge rate 129, is used as the evaluation value. The evaluation value │ΔPv│ 130 is an example of the first evaluation result. A smaller │ΔPv│ value 130 is better.

[0102] Figure 13(c) is a graph showing an example of the measurement result 128N for sample L1 according to the third embodiment. In the graph shown in Figure 13(c), the horizontal axis represents the discharge speed of sample L1, and the vertical axis represents the discharge volume of sample L1. The measurement result 128N shown in Figure 13(c) is an updated measurement result obtained by re-preparing and re-evaluating the sample using the re-presented preparation conditions. By measuring the re-prepared sample L1 using the measuring device 102, a discharge volume of 128aN and a discharge speed of 128bN are obtained as the re-evaluated measurement result 128N. The discharge volume 128aN and discharge speed 128bN are examples of measured values. The measurement result 128N is an example of the second measurement result.

[0103] Figure 13(d) is a graph showing the difference between the discharge rate 128aN and the target discharge rate 129 for sample L1 according to the third embodiment. In the graph shown in Figure 13(d), the horizontal axis represents the manufacturing condition parameters, and the vertical axis represents the discharge rate. The value of │ΔPvN│ 1301, which is the difference between the discharge rate 128aN and the target discharge rate 129, is used as the evaluation value. The evaluation value of │ΔPvN│ 1301 is an example of the second evaluation result. A smaller value of │ΔPvN│ 1301 is better.

[0104] Figure 14 is a flowchart of the sample preparation and evaluation method, i.e., the control method of the sample preparation and evaluation system 100, according to the third embodiment.

[0105] In step S1, the information processing unit 105 presents the preparation conditions to the sample preparation device 101. In step S2, the sample preparation device 101 prepares the sample. In step S3, the information processing unit 105 causes the measuring device 102 to measure the sample L1. Then, in step S4, the information processing unit 105 evaluates the sample L1 using the measurement result 128 obtained from the measuring device 102 and obtains the evaluation value of sample L1, the value of │ΔPv│ 130.

[0106] In step S5, if the value of │ΔPv│ 130, which is the evaluation value of sample L1, does not satisfy the predetermined conditions, that is, if step S5 is YES, the information processing unit 105 proceeds to step S6.

[0107] In step S6, the information processing unit 105 determines whether the termination condition is met. If the evaluation value does not meet the predetermined condition, i.e., if step S6 is No, the information processing unit 105 proceeds to step S9.

[0108] In step S9, the information processing unit 105 obtains an estimated function from the acquired evaluation values ​​and determines the next manufacturing conditions based on the estimated function. Then, in step S10, the information processing unit 105 updates the manufacturing conditions.

[0109] If the predetermined conditions are met in step S5, that is, if step S5 is NO, then in step S7, the information processing unit 105 presents the preparation conditions for sample L1 to the user again. Then, in step S12, the information processing unit 105 performs the preparation and evaluation again and obtains the measurement result 128N.

[0110] Next, in step S11, the information processing unit 105 performs an averaging process to average the measurement result 128 and the measurement result 128N, and re-evaluates based on the result of the averaging process, i.e., the average value. Then, the information processing unit 105 updates the evaluation value of sample L1 to the value of |ΔPvN| 1301, updates the estimation function in step S9 based on the evaluation value updated in step S11, and determines the next manufacturing conditions in step S10.

[0111] In the third embodiment, the evaluation value was set to the value of │ΔPvN│ 1301, which is the difference between the measured discharge rate 128aN and the target discharge rate 129, but it is not limited to this. For example, if multiple measurement results are obtained, such as a discharge rate of 128aN and a discharge rate of 128bN, the evaluation value may be a value determined based on a calculation formula consisting of combinations of these results.

[0112] The predetermined condition in the third embodiment is that the value of |ΔPv| 130 falls within the range |ΔP0| defined by a preset threshold 131. The threshold 131 may be, for example, a value of ±10% of the target discharge amount 129. Note that the predetermined condition is not limited to the example described above. Furthermore, the predetermined condition does not have to be executed from the initial stage of the search, but may be applied after a user-specified number of times.

[0113] In the third embodiment, the information processing unit 105 can perform manufacturing and evaluation a specified number of times for the presented manufacturing conditions and obtain measurement values ​​indicating the measurement results for the specified number of times. The specified number of times can be one or any number, and the user may specify and change it.

[0114] The information processing unit 105 performs sample preparation and evaluation a specified number of times according to the presented preparation conditions, and performs averaging processing to average the multiple measurement results 128 using a predetermined method. Examples of predetermined methods include calculating the average value of all multiple measurement values ​​as described above, or excluding measurement values ​​that exceed a predetermined range as outliers, and calculating the average value of the remaining measurement values ​​after excluding the outliers. Outliers may be, for example, those outside the range of mean ± 3 × standard deviation. Alternatively, averaging processing may be performed using a combination of measurement values ​​among the multiple measurement values ​​that is smaller than a predetermined standard deviation value. The method of averaging processing can be specified and changed by the user.

[0115] Figure 15(a) is a graph showing an example of the measurement results for multiple samples L1 to L4 according to the third embodiment. In Figure 15(a), the horizontal axis represents the discharge rate and the vertical axis represents the discharge volume. An example in which the specified number of preparation and evaluation cycles were performed 6 times and the number of exploration cycles 4 times will be described.

[0116] Measurement results 128, 132, 133, and 134 are sets of measurement results for samples L1 to L4, respectively. There is variability in the six measurement results for each of samples L1 to L4. Factors causing this variability include fluctuations originating from the measurement device, fluctuations in the control system of the sample preparation device, changes in material properties due to time, and fluctuations in ambient environmental conditions. These factors causing variability are not included in the preparation condition parameters.

[0117] Figure 15(b) is a graph showing the difference between the average discharge rate of each of the multiple samples L1 to L4 according to the third embodiment and the target discharge rate 129. Figure 15(b) shows the relationship between the average value of the measurement results for each of the samples L1 to L4, the target value, the target discharge rate 129, the difference ΔPv between the average value of the measurement results for each of the samples L1 to L4 and the target discharge rate 129, and the threshold 131. The average discharge rates 1280a, 1320a, 1330a, and 1340a represent the average values ​​of the measured discharge rates for the samples L1 to L4, respectively.

[0118] As shown in Figure 15(b), for samples L2 and L4, the evaluation value │ΔPv│ exceeds the threshold of 131, while for samples L1 and L3, the evaluation value │ΔPv│ is within the range ΔP0 of the threshold of 131.

[0119] The information processing unit 105, at the end of the fourth search cycle, re-presents the preparation conditions for samples L1 and L3 whose |ΔPv| values ​​satisfy the predetermined conditions, and then performs sample preparation and evaluation a specified number of times. As a result, as shown in Figure 16(a), the second measurement results, measurement result 128N and measurement result 133N, are obtained.

[0120] Figure 16(a) is a graph showing an example of the measurement results for several samples L1 to L4 according to the third embodiment. In Figure 16(a), the horizontal axis represents the discharge rate and the vertical axis represents the discharge volume. Figure 16(a) shows an example of updated measurement results for samples L1 and L3 that were re-prepared and re-evaluated according to the re-presented preparation conditions.

[0121] Figure 16(b) is a graph showing the difference between the average discharge rate of each of the multiple samples L1 to L4 according to the third embodiment and the target discharge rate 129. Figure 16(b) shows the relationship between the average value of the measurement results for each of the samples L1 to L4, the target value, the target discharge rate 129, the difference ΔPv(ΔPvN) between the average value of the measurement results for each of the samples L1 to L4 and the target discharge rate 129, and the threshold 131. The average discharge rates 1280aN, 1320a, 1330aN, and 1340a represent the average values ​​of the measured discharge rates for the samples L1 to L4, respectively.

[0122] In Figure 16(b), the average discharge rate of 1280a has been updated to an average discharge rate of 1280aN, and the average discharge rate of 1330a has been updated to an average discharge rate of 1330aN. As a result, the evaluation values ​​of |ΔPv| and |ΔPvN| have been updated accordingly.

[0123] The average value of all the multiple measurement results 128 before and 128 after sample L1 is set to a discharge volume of 1280 aN. Similarly, the average value of all the multiple measurement results 133 before and 133 after sample L3 is set to a discharge volume of 1330 aN.

[0124] The information processing unit 105 calculates the evaluation value │ΔPvN│ for each of sample L1 and sample L3. The information processing unit 105 updates the estimation function from the evaluation values ​​│ΔPv│ for each of samples L1 and L4, and the evaluation values ​​│ΔPvN│ for each of samples L2 and L3, and determines the next manufacturing conditions.

[0125] In the third embodiment, a case in which the information processing unit 105 performs averaging on the measured values ​​was described, but the invention is not limited to this. For example, the information processing unit 105 may perform averaging on the evaluation values ​​(first evaluation result and second evaluation result) obtained from the measured values. The user can specify as appropriate whether to average the measured values ​​or the evaluation values.

[0126] In other words, the information processing unit 105 of the information processing device 103 performs an averaging process to average multiple first measurement results when one or more first measurement results are multiple first measurement results, an averaging process to average one or more first measurement results and one or more second measurement results, an averaging process to average multiple first evaluation results when one or more first evaluation results are multiple first evaluation results, or an averaging process to average one or more first evaluation results and one or more second evaluation results.

[0127] Furthermore, in the third embodiment, an example was shown in which the information processing unit 105 instructs the reprocessing of samples collectively when the predetermined conditions are met, but it is not limited to this. For example, the information processing unit 105 may also perform reprocessing each time the predetermined conditions are met, and the user can choose whichever is appropriate.

[0128] If variations occur during sample preparation and measurement, creating an estimation function using evaluation values ​​obtained from the measurement results of a single trial and then performing the update process may result in discrepancies in the preparation conditions presented below. As a result, the number of searches may increase, and the number of trials required to obtain the target discharge volume may become enormous. If the discharge volume clearly deviates from the target discharge volume and does not meet the predetermined conditions, even if there are variations in the measurement results, the contribution of the evaluation value to the estimation function is small because the discharge volume is far from the target discharge volume, and it is inefficient to repeat sample preparation and measurement in the region where the evaluation value is large. On the other hand, if the predetermined conditions are met, the discharge volume is close to the target discharge volume, so the variations cannot be ignored.

[0129] In contrast, in the third embodiment, by adding the measurement result 128N only when predetermined conditions are met, the accuracy of the measurement result can be improved, and it becomes possible to quickly present the production conditions for producing high-quality samples.

[0130] Furthermore, in the third embodiment, if there is variation in the measured values ​​due to fluctuations in various devices of the sample preparation device 101 or the measuring device 102, the information processing unit 105 repeats the preparation and evaluation of multiple samples and uses the averaged values ​​for evaluation. The averaging process improves the accuracy of the measured and evaluated values, reduces the number of searches, and allows for the determination of sample preparation conditions in a short time. The third embodiment can also be applied to tuning the manufacturing process according to the manufacturing environment of the manufacturing equipment.

[0131] In summary, according to this disclosure, the conditions for producing high-quality samples can be obtained in a short period of time.

[0132] This disclosure is not limited to the embodiments described above, and many modifications are possible within the technical concept of this disclosure. For example, at least two of the embodiments and modifications described above may be combined. Furthermore, the effects described in this embodiment are merely a list of the most preferred effects arising from the embodiments of this disclosure, and the effects of the embodiments of this disclosure are not limited to those described in this embodiment.

[0133] (Other examples) This disclosure can also be implemented by supplying a program that implements one or more of the functions of the above-described embodiments to a system or device via a network or storage medium, and by having one or more processors in the computer of that system or device read and execute the program. It can also be implemented by a circuit (e.g., an ASIC) that implements one or more functions.

[0134] The above disclosure of embodiments includes the following sections.

[0135] (Section 1) A sample preparation device that prepares samples based on preparation conditions, A measuring device for measuring the aforementioned sample, The system includes an information processing device that evaluates the sample based on the measurement results of the measuring device and obtains the evaluation results of the sample. The information processing device notifies that the evaluation result needs to be corrected if the evaluation result does not meet the predetermined conditions. A sample preparation and evaluation system characterized by the following features.

[0136] (Section 2) The sample preparation apparatus prepares a sample based on the updated preparation conditions. The sample preparation and evaluation system according to item 1, characterized in that

[0137] (Section 3) The sample preparation apparatus prepares a sample based on the updated preparation conditions whenever the preparation conditions are updated. A sample preparation and evaluation system according to item 1 or 2, characterized by the above.

[0138] (Section 4) If the number of evaluation results requiring correction reaches a predetermined number, the sample preparation device will stop preparing the samples. The sample preparation and evaluation system according to item 3, characterized by the features described herein.

[0139] (Section 5) The aforementioned evaluation results include evaluation values, The information processing device is configured to obtain an estimation function using the evaluation value and update the manufacturing conditions using the estimation function. When the information processing device receives a request to modify the evaluation value, it uses the modified evaluation value to determine the estimation function. A sample preparation and evaluation system according to any one of items 1 to 4, characterized by the above.

[0140] (Section 6) The evaluation result includes multiple rank values, and any of the multiple rank values ​​is the evaluation value. The sample preparation and evaluation system according to item 5, characterized in that

[0141] (Section 7) The evaluation result includes a plurality of probabilities corresponding to each of the plurality of rank values, The information processing device uses the rank value corresponding to the highest first probability among the plurality of probabilities as the evaluation value. The sample preparation and evaluation system according to item 6, characterized in that

[0142] (Section 8) The predetermined condition is that the difference between the first probability and the second probability, which is the next highest among the multiple probabilities, is greater than or equal to a threshold. The sample preparation and evaluation system according to item 7, characterized in that

[0143] (Section 9) The information processing device acquires the evaluation result for the measurement result using a trained machine learning model. A sample preparation and evaluation system according to any one of claims 1 to 8, characterized by the above.

[0144] (Section 10) The aforementioned measuring device is a sensory sensor, The information processing device acquires sensory information of the sample from the sensory sensor as a measurement result. A sample preparation and evaluation system according to any one of claims 1 to 9, characterized by the above.

[0145] (Section 11) The aforementioned sensory sensor is at least one of a visual sensor, an olfactory sensor, a gustatory sensor, and a pressure sensor. The sample preparation and evaluation system according to item 10, characterized in that

[0146] (Section 12) It also has a display device, If the evaluation result of the sample does not meet predetermined conditions, the information processing device notifies the user by displaying an image on the display device indicating that the evaluation result needs to be corrected. A sample preparation and evaluation system according to any one of items 1 to 11, characterized by the above.

[0147] (Section 13) The information processing device displays a user interface image on the display device that accepts user modifications to the evaluation results. The sample preparation and evaluation system according to item 12, characterized in that

[0148] (Section 14) A sample preparation device that prepares samples based on preparation conditions, A first measuring device for measuring the aforementioned sample, A second measuring device that measures the sample with higher precision than the first measuring device described above, The system includes an information processing device that evaluates the sample based on the measurement results of the first measuring device and obtains the evaluation results of the sample, If the evaluation result does not meet predetermined conditions, the information processing device will have the second measuring device measure the sample, evaluate the sample based on the measurement results of the second measuring device, and update the evaluation result of the sample. A sample preparation and evaluation system characterized by the following features.

[0149] (Section 15) The aforementioned evaluation results include evaluation values, The information processing device is configured to obtain an estimation function using the evaluation value and update the manufacturing conditions using the estimation function. When the evaluation value is updated, the information processing device uses the updated evaluation value to determine the estimation function. The sample preparation and evaluation system according to item 14, characterized in that

[0150] (Section 16) Each of the first measuring device and the second measuring device is a device for measuring the surface shape of the sample, The measurement result is the measured shape of the surface, The evaluation result is the difference between the measured shape and the design shape. The sample preparation and evaluation system according to item 15, characterized in that

[0151] (Section 17) The aforementioned predetermined condition is the condition in which the maximum value of the difference is equal to or greater than a threshold. The sample preparation and evaluation system according to item 16, characterized in that

[0152] (Section 18) The sample preparation apparatus prepares a sample based on the updated preparation conditions. A sample preparation and evaluation system according to any one of claims 15 to 17, characterized by the above.

[0153] (Section 19) A sample preparation device that prepares samples based on preparation conditions, A measuring device for measuring the aforementioned sample, The system includes an information processing device that evaluates the sample based on the measurement results of the measuring device and obtains the evaluation results of the sample. The aforementioned information processing device is If at least one first evaluation result obtained from at least one first measurement result of the sample prepared under the above preparation conditions satisfies the predetermined conditions, the sample preparation device is made to prepare the sample again. The sample prepared again is measured by the measuring device, and at least one second measurement result is obtained from the measuring device. Based on the aforementioned at least one second measurement result, at least one second evaluation result is obtained. The manufacturing conditions are updated based on at least one of the second evaluation results. A sample preparation and evaluation system characterized by the following features.

[0154] (Section 20) The information processing device performs an averaging process to average the plurality of first measurement results when the at least one first measurement result is a plurality of first measurement results, an averaging process to average the at least one first measurement result and the at least one second measurement result, an averaging process to average the plurality of first evaluation results when the at least one first evaluation result is a plurality of first evaluation results, or an averaging process to average the at least one first evaluation result and the at least one second evaluation result. A sample preparation and evaluation system according to item 19, characterized in that it is a sample preparation and evaluation system.

[0155] (Section 21) A control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The steps include: the measuring device measuring the sample, The information processing device performs the steps of evaluating the sample based on the measurement results of the measuring device and obtaining the evaluation results of the sample. The information processing device includes the step of notifying that if the evaluation result does not meet predetermined conditions, the evaluation result needs to be corrected. A control method for a sample preparation and evaluation system characterized by the following.

[0156] (Section 22) A control method for a sample preparation and evaluation system having a sample preparation device, a first measuring device, a second measuring device having higher measurement accuracy than the first measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The first measuring device measures the sample, The information processing device performs the steps of evaluating the sample based on the measurement results of the first measuring device and obtaining the evaluation results of the sample. If the evaluation result does not satisfy the predetermined conditions, the second measuring device measures the sample. The information processing device includes the step of evaluating the sample based on the measurement results of the second measuring device and updating the evaluation results of the sample. A control method for a sample preparation and evaluation system characterized by the following.

[0157] (Section 23) A control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The steps include: the measuring device measuring the sample and obtaining a first measurement result; The information processing device acquires a first evaluation result based on the first measurement result, If the first evaluation result satisfies the predetermined conditions, the sample preparation device prepares the sample again. The steps include: the measuring device measuring the newly prepared sample to obtain a second measurement result; The information processing device acquires a second evaluation result based on the second measurement result, The information processing device includes the step of updating the manufacturing conditions based on the second evaluation result, A control method for a sample preparation and evaluation system characterized by the following.

[0158] (Section 24) A program for causing a computer to perform any one of the control methods described in items 21 to 23.

[0159] (Section 25) A computer-readable recording medium on which the program described in item 24 is recorded. [Explanation of Symbols]

[0160] 100, 100A…Sample preparation and evaluation system, 101…Sample preparation device, 102…Measurement device, 102A…Measurement device (first measurement device), 102B…Measurement device (second measurement device), 103…Information processing device, 104…Display device

Claims

1. A sample preparation device that prepares samples based on preparation conditions, A measuring device for measuring the aforementioned sample, The system includes an information processing device that evaluates the sample based on the measurement results of the measuring device and obtains the evaluation results of the sample. The information processing device notifies that the evaluation result needs to be corrected if the evaluation result does not meet the predetermined conditions. A sample preparation and evaluation system characterized by the following features.

2. The sample preparation apparatus prepares a sample based on the updated preparation conditions. The sample preparation and evaluation system according to claim 1.

3. The sample preparation apparatus prepares a sample based on the updated preparation conditions whenever the preparation conditions are updated. The sample preparation and evaluation system according to claim 1.

4. If the number of evaluation results requiring correction reaches a predetermined number, the sample preparation device will stop preparing the samples. The sample preparation and evaluation system according to claim 3.

5. The aforementioned evaluation results include evaluation values, The information processing device is configured to obtain an estimation function using the evaluation value and update the manufacturing conditions using the estimation function. When the information processing device receives a request to modify the evaluation value, it uses the modified evaluation value to determine the estimation function. The sample preparation and evaluation system according to claim 1.

6. The evaluation result includes multiple rank values, and any of the multiple rank values ​​is the evaluation value. The sample preparation and evaluation system according to claim 5.

7. The evaluation result includes a plurality of probabilities corresponding to each of the plurality of rank values, The information processing device determines the rank value corresponding to the highest first probability among the plurality of probabilities. The evaluation value is as follows: The sample preparation and evaluation system according to claim 6.

8. The predetermined condition is that the difference between the first probability and the second probability, which is the next highest among the multiple probabilities, is greater than or equal to a threshold. The sample preparation and evaluation system according to claim 7.

9. The information processing device acquires the evaluation result for the measurement result using a trained machine learning model. The sample preparation and evaluation system according to claim 1.

10. The aforementioned measuring device is a sensory sensor, The information processing device acquires sensory information of the sample from the sensory sensor as a measurement result. A sample preparation and evaluation system according to any one of claims 1 to 9, characterized by the above.

11. The aforementioned sensory sensor is at least one of a visual sensor, an olfactory sensor, a gustatory sensor, and a pressure sensor. The sample preparation and evaluation system according to claim 10.

12. It also has a display device, If the evaluation result of the sample does not meet predetermined conditions, the information processing device notifies the user by displaying an image on the display device indicating that the evaluation result needs to be corrected. The sample preparation and evaluation system according to claim 1.

13. The information processing device displays a user interface image on the display device that accepts user modifications to the evaluation results. The sample preparation and evaluation system according to claim 12.

14. A sample preparation device that prepares samples based on preparation conditions, A first measuring device for measuring the aforementioned sample, A second measuring device that measures the sample with higher precision than the first measuring device described above, The system includes an information processing device that evaluates the sample based on the measurement results of the first measuring device and acquires the evaluation results of the sample, If the evaluation result does not meet predetermined conditions, the information processing device will have the second measuring device measure the sample, evaluate the sample based on the measurement result of the second measuring device, and update the evaluation result of the sample. A sample preparation and evaluation system characterized by the following features.

15. The aforementioned evaluation results include evaluation values, The information processing device is configured to obtain an estimation function using the evaluation value and update the manufacturing conditions using the estimation function. When the evaluation value is updated, the information processing device uses the updated evaluation value to determine the estimation function. The sample preparation and evaluation system according to claim 14.

16. Each of the first measuring device and the second measuring device is a device for measuring the surface shape of the sample, The measurement result is the measured shape of the surface, The evaluation result is the difference between the measured shape and the design shape. The sample preparation and evaluation system according to claim 15.

17. The aforementioned predetermined condition is the condition in which the maximum value of the difference is equal to or greater than a threshold. The sample preparation and evaluation system according to claim 16.

18. The sample preparation apparatus prepares a sample based on the updated preparation conditions. A sample preparation and evaluation system according to any one of claims 15 to 17, characterized by the above.

19. A sample preparation device that prepares samples based on preparation conditions, A measuring device for measuring the aforementioned sample, The system includes an information processing device that evaluates the sample based on the measurement results of the measuring device and obtains the evaluation results of the sample. The aforementioned information processing device is If at least one first evaluation result obtained from at least one first measurement result of the sample prepared under the above preparation conditions satisfies predetermined conditions, the sample preparation device is made to prepare the sample again. The sample prepared again is measured by the measuring device, and at least one second measurement result is obtained from the measuring device. Based on the aforementioned at least one second measurement result, at least one second evaluation result is obtained. The manufacturing conditions are updated based on at least one of the second evaluation results. A sample preparation and evaluation system characterized by the following features.

20. The information processing device performs an averaging process to average the plurality of first measurement results when the at least one first measurement result is a plurality of first measurement results, an averaging process to average the at least one first measurement result and the at least one second measurement result, an averaging process to average the plurality of first evaluation results when the at least one first evaluation result is a plurality of first evaluation results, or an averaging process to average the at least one first evaluation result and the at least one second evaluation result. The sample preparation and evaluation system according to claim 19.

21. A control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The steps include: the measuring device measuring the sample, The information processing device performs the steps of evaluating the sample based on the measurement results of the measuring device and obtaining the evaluation results of the sample. The information processing device includes the step of notifying that if the evaluation result does not meet predetermined conditions, the evaluation result needs to be corrected. A control method for a sample preparation and evaluation system characterized by the following.

22. A control method for a sample preparation and evaluation system having a sample preparation device, a first measuring device, a second measuring device having higher measurement accuracy than the first measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The first measuring device measures the sample, The information processing device performs the steps of evaluating the sample based on the measurement results of the first measuring device and obtaining the evaluation results of the sample. If the evaluation result does not satisfy the predetermined conditions, the second measuring device measures the sample. The information processing device includes the step of evaluating the sample based on the measurement results of the second measuring device and updating the evaluation results of the sample. A control method for a sample preparation and evaluation system characterized by the following.

23. A control method for a sample preparation and evaluation system having a sample preparation device, a measuring device, and an information processing device, The steps include: the sample preparation apparatus preparing a sample based on the preparation conditions; The steps include: the measuring device measuring the sample and obtaining a first measurement result; The information processing device acquires a first evaluation result based on the first measurement result, If the first evaluation result satisfies the predetermined conditions, the sample preparation device prepares the sample again. The steps include: the measuring device measuring the newly prepared sample to obtain a second measurement result; The information processing device acquires a second evaluation result based on the second measurement result, The information processing device includes the step of updating the manufacturing conditions based on the second evaluation result, A control method for a sample preparation and evaluation system characterized by the following.

24. A program for causing a computer to execute the control method described in any one of claims 21 to 23.

25. A computer-readable recording medium having the program described in claim 24 recorded on it.

Images