Figure 2
From: Machine learning analysis of rogue solitons in supercontinuum generation
Schematic of the feed-forward neural network used in this work. (a) The input of the network is a single-shot supercontinuum spectral intensity vector \({{\bf{X}}}_{n}=[{x}_{1},{x}_{2},\cdots ,{x}_{N}]\) yielding the output of the network y that corresponds to the peak power, duration or temporal delay of the rogue solitons. The network consists of two fully connected hidden layers and a single output neuron. (b) shows the operation of a single neuron. The output of a generic neuron \({n}_{i}^{(k)}\) (ith neuron in layer k) is calculated as a weighted sum between the outputs from the previous layer k − 1 and the weights of each connection \({w}_{ij}^{(k)}\) which is followed by adding a bias term \({b}_{i}^{(k)}\) and nonlinearity. See Methods for more details.
