Automatic optimization of measurement accuracy through advanced machine learning techniques

摘要

The machine learning technique is used to predict the value of a fixed parameter when a reference value of a critical parameter is given. A neural network is a low-dimensional real associated with a spectrum, such as, for example, one or more critical parameters, and, for example, a spectroscopic ellipsometry spectrum or a specular reflectance spectrum. -valued vector). Another neural network can map a low-dimensional real-valued vector. When using two neural networks, one neural network can be trained to map the spectrum to a low-dimensional real-valued vector. Another neural network can be trained to predict fixed parameters based on low-dimensional real-valued vectors and critical parameters from this other neural network.