A method and device for predicting a defect. The method includes determining a sequence between a plurality of sub-models by modeling a production process into the plurality of sub-models, mapping production process data into each of the plurality of sub-models, determining, by a corresponding sub-model, output data comprising defect information on a potential defect occurring in a corresponding step, for each of the plurality of sub-models, predicting information associated with a defect in the production process based on the output data corresponding to each of the plurality of sub-models, and inputting the output data of each of the sub-models to a subsequent sub-model of the corresponding sub-model, based on the sequence.

Methods and systems for reducing substrate particle scratching using machine learning are provided. A machine learning model is trained to predict process recipe settings for a substrate temperature control process to be performed for a current substrate at a manufacturing system. First training data and second training data are generated for the machine learning model. The first training data includes historical data associated with prior process recipe settings for a prior substrate temperature control process performed for a prior substrate at a prior process chamber. The second training data is associated with a historical scratch profile of one or more surfaces of the prior substrate after performance of the prior substrate temperature control process according to the prior process recipe settings. The first training data and the second training data are provided to train the machine learning model to predict which process recipe settings for the substrate temperature control process to be performed for the current substrate correspond to a target scratch profile for one or more surfaces of the current substrate.

A non-transitory computer-readable medium comprising computer-executable instructions that, when executed, are configured to cause a processor to perform operations that include receiving image data after an operation is performed by an industrial automation device on a product; analyzing the image data based an object-based image analysis (OBIA) model to classify the product as one of a plurality of conditions related to manufacturing quality and the OBIA model includes property layers associated with features related to a manufacturing of the product; determining whether the one of the conditions indicates an anomaly being present in the product; sending a notification indicative of the one of the plurality of conditions is presently associated with the product; identifying a property layer associated with classifying the one of the plurality of conditions; and updating the OBIA model based on the property layer and the input indicative of the anomaly being incorrectly associated with the product.

A method includes receiving one or more fingerprint dimensions to be used to generate a fingerprint. The method further includes receiving trace data associated with a manufacturing process. The method further includes applying the one or more fingerprint dimensions to the trace data to generate at least one feature. The method further includes generating the fingerprint based on the at least one feature. The method further includes causing, based on the fingerprint, performance of a corrective action associated with one or more manufacturing processes.

A method including: obtaining one or more models configured for predicting a process metric of a manufacturing process based on inputting process data; and using a reinforcement learning framework to evaluate the one or more models and/or model configurations of the one more models based on inputting new process data to the one or more models and determining a performance indication of the one or more models and/or model configurations in predicting the process metric based on inputting the new process data.

Process recipe data associated a process to be performed for a substrate at a process chamber is provided as input to a trained machine learning model. A set of process recipe settings for the process that minimizes scratching on one or more surfaces of the substrate is determined based on one or more outputs of the machine learning model. The process is performed for the substrate at the process chamber in accordance with the determined set of process recipe settings.

Example embodiments are disclosed of systems and methods for predicting failure probabilities of future product tests of a testing sequence based on outcomes of prior tests. Predictions are made by a machine-learning-based model (MLM) trained with a set of test-result sequence records (TRSRs) including test values and pass/fail indicators (PRIs) of completed tests. Within training epochs over the set, iterations are carried out over each TRSR. Each iteration involves sub-iterations carried out successively over test results of the TRSR. Each sub-iteration involves (i) inputting to the MLM values of a given test and those of tests earlier in the sequence while masking those later in the sequence, (ii) computing probabilities of test failures for the masked tests found later in the sequence than the given test, and (iii) applying the PFIs of test results later in the sequence than the given test as ground-truths to update parameters of the MLM.

A process control system according to one embodiment of the present invention comprises: a first system for generating thickness information about an internal defect layer included in a carbon steel product; and a second system which receives the thickness information about the internal defect layer from the first system through a network, and which controls an etching process for removing at least a part of the internal defect layer from the carbon steel product by using the thickness information about the internal defect layer, wherein the first system provides the second system with a calculation module necessary for the second system to control the etching process, and the second system provides the first system with the information necessary for the first system to update the calculation module.

A computing system may include a virtual cross metrology engine configured to construct a given virtual metrology model. The given virtual metrology model may take, as inputs, process parameters applied for the given step of a semiconductor fabrication process. The virtual cross metrology engine may also be configured to construct a subsequent virtual metrology model, and the subsequent step is performed after the given step in the semiconductor fabrication process. Doing so may include determining inputs for the subsequent virtual metrology model from a combination of the process parameters applied for the given step of the semiconductor fabrication process, process parameters applied for the subsequent step of the semiconductor fabrication process, and a wafer value for the given step of the semiconductor fabrication process that the given virtual metrology model is configured to predict.

Provided is a method and system for detection of anomalous work pieces that includes computing at least one deviation data signal for a target data signal of a target work piece with respect to reference data signals recorded for a corresponding production process step of a set of reference work pieces, performing a stepwise anomaly detection by data processing of the at least one computed deviation data signal and a process type indicator indicating a type of the production process step using a trained anomaly detection data model to calculate for each time step or path length step of the production process step an anomaly probability that the respective time step or path length step is anomalous, and classifying the target work piece and/or the production process step as being anomalous or not anomalous on the basis of the calculated anomaly probabilities.