System Identification with Time-Aware Neural Sequence Models
This addresses a specific challenge in system identification for industrial applications, but it is incremental as it extends existing models like GRUs.
The paper tackled the problem of dynamical system identification with unevenly sampled continuous variables by adapting neural sequence models to handle variable step sizes, demonstrating validity on two industrial input/output processes.
Established recurrent neural networks are well-suited to solve a wide variety of prediction tasks involving discrete sequences. However, they do not perform as well in the task of dynamical system identification, when dealing with observations from continuous variables that are unevenly sampled in time, for example due to missing observations. We show how such neural sequence models can be adapted to deal with variable step sizes in a natural way. In particular, we introduce a time-aware and stationary extension of existing models (including the Gated Recurrent Unit) that allows them to deal with unevenly sampled system observations by adapting to the observation times, while facilitating higher-order temporal behavior. We discuss the properties and demonstrate the validity of the proposed approach, based on samples from two industrial input/output processes.