Learning dynamical models from the data is a fundamental process in science and engineering. System identification is the problem of building dynamical models from measured data of temporal signals in order to determine predictive relationships between them. The identified models can be used for forecasting the future outputs of the system, predicting responses to new input signals, or controlling the system. Satisfactorily solving this challenging problem requires flexible modeling tools that allow to incorporate structural knowledge that is specific of dynamical systems, as well as efficient computational techniques that are able to handle complex datasets. The machine learning community has been producing a large amount of modeling techniques for structured and large scale data that may potentially have a tremendous impact in the field of dynamical system identification.
This workshop aims at bringing together researchers from machine learning and system identification communities to foster discussion and elicit open problems on modeling and identification of dynamical systems. The workshop will consist of invited talks, contributed presentations, and posters. We plan to include an opening tutorial on system identification and an overview of the state-of-the-art techniques. Invited talks will be given by leading experts from both machine learning and system identification communities.