PAC-Bayesian bounds for learning LTI-ss systems with input from empirical loss
This provides theoretical guarantees for system identification algorithms, with potential applications to recurrent neural networks as a sub-class.
The paper tackles the problem of deriving finite-sample error bounds for learning linear time-invariant stochastic dynamical systems with inputs, resulting in a PAC-Bayesian bound that relates future prediction errors to empirical errors on training data.
In this paper we derive a Probably Approxilmately Correct(PAC)-Bayesian error bound for linear time-invariant (LTI) stochastic dynamical systems with inputs. Such bounds are widespread in machine learning, and they are useful for characterizing the predictive power of models learned from finitely many data points. In particular, with the bound derived in this paper relates future average prediction errors with the prediction error generated by the model on the data used for learning. In turn, this allows us to provide finite-sample error bounds for a wide class of learning/system identification algorithms. Furthermore, as LTI systems are a sub-class of recurrent neural networks (RNNs), these error bounds could be a first step towards PAC-Bayesian bounds for RNNs.