A method used in making a product, wherein a characteristic of the product is at least in part determined by values of parameters used in making the product, the method including the steps of: (a) applying a machine-based transfer learning process to prior result data, the application of the transfer learning process resulting in the generation of predictive data; (b) selecting one or more parameter values to be used in making the product based on the generated predictive data; (c) making the whole or a part of the product using the selected one or more parameter values.

The invention may be embodied in a system and method for monitoring and controlling feedback control in a manufacturing process, such as an integrated circuit fabrication process. The process control parameters may include translation, rotation, magnification, dose and focus applied by a photolithographic scanner or stepper operating on silicon wafers. Overlay errors are used to compute measured parameters used in the feedback control process. Statistical parameters are computed, normalized and graphed on a common set of axes for at-a-glance comparison of measured parameters and process control parameters to facilitate the detection of problematic parameters. Parameter trends and context relaxation scenarios are also compared graphically. Feedback control parameters, such as EWMA lambdas, may be determined and used as feedback parameters for refining the APC model that computes adjustments to the process control parameters based on the measured parameters.

Production plant (1) for producing at least one end product (3) from at least one primary starting material (2), comprising at least one processing station (41-43) which processes at least one starting material (21-23) to form at least one product (31, 32, 33), and a process controller (51-53) which can control at least one variable (71-73), which is a measure of a quality feature of the product (31-33) and/or is correlated with a quality feature of the product (31-33), by influencing at least one manipulated variable (61-63) acting on the processing station (41-43), wherein the process controller (51-53) is additionally designed to control the production rate (31a-33a) of the processing station (41-43) for the product (31-33) and/or the consumption rate (21a-23a) of the processing station (41-43) with regard to starting material (21-23) by acting on the manipulated variable (61-63).

An intelligent processing modulation system comprises: a detector detecting the condition of the processing equipment before the process for processing a raw material into a product begins; a data storage recording the information of the product; a basic database storing the information of the detector and the data storage corresponding to the product and the condition of the product for each time the process is completed; a mode database receiving the information of the detector, the data storage and the basic database and setting up the processing model; a processing equipment operating according to the parameters of the processing model received by the processing information manager; a processing information manager modifying the processing model promptly according to the result of the product in order to improve the yield rate of the product.

In a control method for the movement of a tool with a machine tool, the machine tool involves a numerically controlled machine tool, in order to produce an arbitrary required surface of a workpiece by machining. A numeric path program is created which describes the machining of the workpiece with the tool at machining points and which controls the control device. The numeric path program produces a path with respect to the geometric nature of the surface of the workpiece to be machined, with the path including a plurality of sample points and individual paths, with each individual path connecting a pair of the sample points to each other. The numeric path program is evaluated and selected on the basis of a geometric quality criterion, with the geometric quality criterion having continuity as at least one criterion.

To provide a trusted, secure, and immutable record of transactions within a process plant, techniques are described for utilizing a distributed ledger in process control systems. The distributed ledger may be maintained by nodes which receive transactions broadcasted from field devices, controllers, operator workstations, or other devices operating within the process plant. The transactions may include process plant data, such as process parameter data, product parameter data, configuration data, user interaction data, maintenance data, commissioning data, plant network data, and product tracking data. The distributed ledgers may also be utilized to execute smart contracts to allow machines such as field devices to transact by themselves without human intervention. In this manner, recorded process parameter values and product parameter values may be retrieved to verify the quality of products. Moreover, regulatory data may be recorded in response to triggering events so that regulatory agencies can review the data.

A system for rapidly adapting production of a product based on classification of production data using a classifier trained on prior production data is provided. A production control system includes a learning system and an adaptive system. The learning system trains a production classifier to label or classify previously collected production data. The adaptive system receives production data in real time and classifies the production data in real time using the production classifier. If the classification indicates a problem with the manufacturing of the product, the adaptive system controls the manufacturing to rectify the problem by taking some corrective action. The production classifier is trained using bootstrap data and corresponding example data extracted from prior production data. Once the bootstrap data is labeled, the corresponding example data is automatically labeled for use as training data.

This disclosure relates to a device, a method, and a program capable of removing erroneous data from learning data used for machine learning used in natural language processing, for example. The method includes storing a forward direction learned model of a discrete series converter. The model is trained based on a plurality of pairs of discrete series of texts. Each pair comprises a first discrete series indicates an input of discrete series. A second discrete series indicates an output of discrete series. The first discrete series and the second discrete series are correctly associated. The method further includes converting the first discrete series to the second discrete series, and generating a quality score using the forward direction learned model, using a second learning pair of discrete series texts including an error in relationship.