Joint state and parameter estimation for Boolean dynamical systems

In a recent publication, a novel state-space signal model was proposed for discrete-time Boolean dynamical systems. The optimal recursive MMSE estimator for this model is called the Boolean Kalman filter (BKF), and an efficient algorithm was presented for its exact computation. In the present paper, we consider the system identification problem, i.e., the problem of parameter estimation for the case where only incomplete knowledge about the system is available. To solve this problem, we propose the application of the BKF in the context of the well-known paradigm of joint estimation of state and parameters. The approach is illustrated via a network inference example.