Training a neural network is the process of finding a set of weight and bias values so that for a given set of inputs, the outputs produced by the neural network are very close to some target values.

Research on two-dimensional designer optical structures, or metasurfaces, has mainly focused on controlling the wavefronts of light propagating in free space. Here, we show that gradient metasurface ...

Back-propagation compares neural network actual outputs (for a given set of inputs, and weights and bias values) with target values, determines the magnitude and direction of the difference between ...