Neural network in the inverse problem of liquid argon structure factor: from gas-to-liquid radial distribution function

Within the framework of the inverse problem theory, together with the dynamical Hopfield neural, the radial distribution function of liquid argon was obtained from neutron scattering data. A modest initial condition, the Boltzmann factor for a pair potential or an ideal gas radial distribution function, was used to propagate the neural differential equations. In both cases, the inverted data were obtained with great accuracy. The present work shows that a combination of the inverse theory approach, together with the Hopfield neural network, is a powerful method to obtain liquid properties. Results are obtained almost instantaneously and can be applied for more complex systems.