Machine learning model generation and updating for manufacturing equipment

A system, comprising memory and a processing device coupled to the memory, wherein the processing device is configured to:

• determine that performance metrics of a processing chamber have changed since a trained machine learning model associated with the processing chamber was trained, wherein the trained machine learning model is configured to receive as input sensor data associated with a process of the processing chamber and produce as output a first indication of performance of the processing chamber;
• determine whether a change in the performance metrics of the processing chamber is a gradual change or a sudden change;
• if the change in the performance metrics of the processing chamber is a gradual change, perform a first training process to generate a new trained machine learning model configured to receive input sensor data of the processing chamber and produce as output a second indication of performance of the processing chamber after the gradual change;
• and
• if the change in the performance metrics of the processing chamber is a sudden change, perform a second training process to generate the new trained machine learning model configured to receive as input sensor data of the processing chamber and produce as output a third indication of performance of the processing chamber after the sudden change, wherein the second training process is different from the first training process, and wherein the second training process comprises generating synthetic training data for performing the second training process that is based on adjusting historical process data collected before the sudden change based on one or more differences in process data caused by the sudden change.

A non-transitory machine-readable storage medium storing instructions which, when executed, cause a processing device to perform operations comprising:

• determining that performance metrics of a processing chamber have changed since a trained machine learning model associated with the processing chamber was trained, wherein the trained machine learning model is configured to receive as input sensor data associated with a process of the processing chamber and produce as output a first indication of performance of the processing chamber;
• determining whether a change in the performance metrics of the processing chamber is a gradual change or a sudden change;
• if the change in the performance metrics of the processing chamber is a gradual change, performing a first training process to generate a new trained machine learning model configured to receive input sensor data of the processing chamber and produce as output a second indication of performance of the processing chamber after the gradual change;
• and
• if the change in the performance metrics of the processing chamber is a sudden change, performing a second training process to generate the new trained machine learning model configured to receive as input sensor data of the processing chamber and produce as output a third indication of performance of the processing chamber after the sudden change, wherein the second training process is different from the first training process, and wherein the second training process comprises generating synthetic training data for performing the second training process that is based on adjusting historical process data collected before the sudden change based on one or more differences in process data caused by the sudden change.