A method for managing nonconformances in laminates. The method comprises recording, by a sensor system, layup information about a layup of layers on a workpiece platform, wherein the layup of layers forms a workpiece and recording inspection information about the laminate on an inspection platform, wherein the laminate is formed from curing the workpiece. An analyzer in a computer system identifies a laminate nonconformance in the laminate using the inspection information and a user input verifies the laminate nonconformance in the laminate is present. An artificial intelligence system is trained by the computer system using the layup information, the inspection information, and the user input verifying the laminate nonconformance.

Data is received that is derived from each of a plurality of inspection camera modules forming part of a quality assurance inspection system. The data includes a feed of images of a plurality of objects passing in front of the respective inspection camera module. Thereafter, the received data is separately analyzed by each inspection camera module using at least one image analysis inspection tool. The results of the analyzing can be correlated for each inspection camera module on an object-by-object basis. The correlating can use timestamps for the images and/or detected unique identifiers within the images and can be performed by a cloud-based server and/or a local edge computer. Access to the correlated results can be provided to a consuming application or process.

A coating production line system for coating work pieces comprises a coating powder, a coating apparatus, an inspection unit to measure the thickness of the applied coating, a conveyor unit to move the work pieces through the system, and a control unit to use thickness requirements and coating parameters to control the coating apparatus based on said coating parameters with a machine learning instance. A database comprises coating powder characteristics parameter as input vector for the machine learning instance for generating an output vector to control the coating apparatus being a first additional part vector. The control unit determines the coating quality based on a comparison between the thickness data acquired from the inspection unit and the retrieved thickness requirement data as second additional part vector. The first and second additional part vectors are fed back as additional parts to the next input vector for the machine learning instance.

Systems and methods are described for training a model for detecting manufacturing anomalies. A test response parameter is identified at a computing device, and a first plurality of component waveforms associated with the test response parameter are received at the computing device. Each waveform of the plurality of waveforms comprises a plurality of datapoints. A model is generated at the computing device, and the model is trained at the computing device and on the first plurality of component waveforms, thereby generating one or more parameters associated with the model. A second plurality of component waveforms associated with the test response parameter is received, and the trained model is accessed. It is indicated using the trained model, whether any of the second plurality of component waveforms comprises an anomaly. For each indicated waveform, the indicated waveform is reviewed and, for each reviewed waveform not comprising an anomaly, the waveform is labelled.

Aspects of the disclosed technology encompass the use of a deep learning controller for monitoring and improving a manufacturing process. In some aspects, a method of the disclosed technology includes steps for: receiving a plurality of control values from two or more stations, at a deep learning controller, wherein the control values are generated at the two or more stations deployed in a manufacturing process, predicting an expected value for an intermediate or final output of an article of manufacture, based on the control values, and determining if the predicted expected value for the article of manufacture is in-specification. In some aspects, the process can further include steps for generating control inputs if the predicted expected value for the article of manufacture is not in-specification. Systems and computer-readable media are also provided.

An additive manufacturing system comprises an apparatus arranged to distribute layer of metallic powder across a build plane and a power source arranged to emit a beam of energy at the build plane and fuse the metallic powder into a portion of a part. The system includes a processor configured to steer the beam of energy across the build plane and receive data generated by one or more sensors that detect electromagnetic energy emitted from the build plane when the beam of energy fuses the metallic powder. The received data is converted into one or more parameters that indicate one or more conditions at the build plane while the beam of energy fuses the metallic powder. The one or more parameters are used as input into a machine learning algorithm to detect one or more defects in the fused metallic powder.

The subject matter herein provides for AI-based prediction of production defects in association with a production system, such as a semiconductor manufacturing machine. In one embodiment, a method begins by receiving production data from the production system. The production data typically comprises non-homogeneous machine parameters and maintenance data, quality test data, and product and process data. Using the production data, a neural network is trained to model an operation of a given machine in the production system. Preferably, the training involves multi-task learning, transfer learning (e.g., using knowledge obtained with respect to a machine of the same type as the given machine), and a combination of multi-task learning and transfer learning. Once the model is trained, it is associated with the given machine operating environment, wherein it is used to provide quality assurance predictions.

It is disclosed a method and a predicting system for automatic prediction of porosity appearance generated during Laser Powder Bed Fusion (L-PBF), performed by an additive manufacturing system from at least one material. The method comprises steps for training a neural network comprising: generating labels of pore in every pixel using a porosity simulator; pre-training, comprising a first sub-step and a second sub-step, the second sub-step comprises using the data set created from the first sub-step to generate a pre-trained ML model; and training, comprising a first sub-step and a second sub-step, the second sub-step comprises using the data set created from the first sub-step to train the pre-trained ML model to generate a trained ML model.

Techniques are presented for the application of neural networks to the fabrication of integrated circuits and electronic devices, where example are given for the fabrication of non-volatile memory circuits and the mounting of circuit components on the printed circuit board of a solid state drive (SSD). The techniques include the generation of high precision masks suitable for analyzing electron microscope images of feature of integrated circuits and of handling the training of the neural network when the available training data set is sparse through use of a generative adversary network (GAN).

A method for determining a root cause affecting yield in a process for manufacturing devices on a substrate, the method including: obtaining yield distribution data including a distribution of a yield parameter across the substrate or part thereof; obtaining sets of metrology data, each set including a spatial variation of a process parameter over the substrate or part thereof corresponding to a different layer of the substrate; comparing the yield distribution data and metrology data based on a similarity metric describing a spatial similarity between the yield distribution data and an individual set out of the sets of the metrology data; and determining a first similar set of metrology data out of the sets of metrology data, being the first set of metrology data in terms of processing order for the corresponding layers, which is determined to be similar to the yield distribution data.