Stochastic inversion of magnetotelluric data using deep reinforcement learning

We propose a new tool to invert magnetotelluric data for 1D model based on deep Q-networks (DQN), which works as a stochastic optimization method. By transforming the inversion problem into a Marko...