Exploiting similarity in system identification tasks with recurrent neural networks.
Sigurd Spieckermann, Siegmund Düll, Steffen Udluft, Alexander Hentschel, Thomas A. Runkler · 2014
Abstract. A new dual-task learning approach based on recurrent neural networks with factored tensor components for system identification tasks is presented. The overall goal is to identify the underlying dynamics of a system given few observations which are augmented by auxiliary data from similar systems. The resulting system identification is motivated by various real-world industrial use cases, e.g. gas or wind turbine modeling for optimization and monitoring. The problem is formalized and the effec-tiveness of the proposed method is assessed on the cart-pole benchmark. 1