Method of deep learning-based examination of a semiconductor specimen and system thereof

A computerized method of training a Deep Neural Network (DNN), the computerized method performed by a processing and memory circuitry (PMC) and comprising:

• training the DNN in a first training cycle using a first training set comprising first training samples, wherein each first training sample comprises at least one first training image synthetically simulated based on design data;
• and
• upon receiving a user feedback with respect to the DNN trained using the first training set, adjusting a second training cycle based on the user feedback by obtaining a second training set comprising augmented training samples and re-training the DNN using the second training set, wherein the augmented training samples are obtained by augmenting at least part of the first training samples using defect-related synthetic data, the augmenting comprising planting one or more new defects in a first training image using defects of known types;
• wherein the trained DNN is usable for examination of a semiconductor specimen.

A non-transitory computer readable storage medium tangibly embodying a program of instructions that, when executed by a computer, cause the computer to perform a method of training a Deep Neural Network (DNN), the method comprising:

• training the DNN in a first training cycle using a first training set comprising first training samples, wherein each first training sample comprises at least one first training image synthetically simulated based on design data;
• and
• upon receiving a user feedback with respect to the DNN trained using the first training set, adjusting a second training cycle based on the user feedback by obtaining a second training set comprising augmented training samples and re-training the DNN using the second training set, wherein the augmented training samples are obtained by augmenting at least part of the first training samples using defect-related synthetic data, the augmenting comprising planting one or more new defects in a first training image using defects of known types;
• wherein the trained DNN is usable for examination of a semiconductor specimen.