Identification of Recurrent Neural Networks by Bayesian Interrogation Techniques

We introduce novel online Bayesian methods for the identification of a family of noisy recurrent neural networks (RNNs). We present Bayesian active learning techniques for stimulus selection given past experiences. In particular, we consider the unknown parameters as stochastic variables and use A-optimality and D-optimality principles to choose optimal stimuli. We derive myopic cost functions in order to maximize the information gain concerning network parameters at each time step. We also derive the A-optimal and D-optimal estimations of the additive noise that perturbs the dynamical system of the RNN. Here we investigate myopic as well as non-myopic estimations, and study the problem of simultaneous estimation of both the system parameters and the noise. Employing conjugate priors our derivations remain approximation-free and give rise to simple update rules for the online learning of the parameters. The efficiency of our method is demonstrated for a number of selected cases, including the task of controlled independent component analysis.