Analysis of Spectra from Laser Produced Plasmas Using a Neural Network

A backpropagation artificial neural network algorithm is applied to the analysis of $K$-shell x-ray line spectra from a well characterized laser produced plasma. After training on synthetic spectra produced by appropriate collisional radiative plasma emission models, the network correctly determines the electron temperature as a function of distance into the plasma. The results demonstrate the potential utility of neural networks for interpreting spectral data from plasma devices and sources.