BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Learning stable Recurrent Neural Networks for model predictive con
 trol
DTSTART:20240531T110000
DTEND:20240531T120000
DTSTAMP:20260928T024855Z
UID:e0d4125785ffc9d81c70b1ec46a11c14efde789fe584d422f8e5a942
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr Fabio Bonassi\,  Uppsala University\, Sweden.\nAbstract \
 nThis talk aims to explore how stable Recurrent Neural Networks (RNN) can 
 be used for indirect data-driven control. \nIn particular\, we unravel th
 e idea of learning RNN models of nonlinear dynamical systems with Incremen
 tal Input-to-State Stability (ISS) certificates\, based on which nonlinear
  Model Predictive Control (MPC) laws with nominal closed-loop properties c
 an be designed. This allows one to take advantage of RNNs’ modeling powe
 r while preserving\, at the same time\, the theoretical properties of MPC 
 schemes.\nTo this end\, it is essential to train provenly-ISS RNNs: we wil
 l therefore discuss how to enforce the stability of the GRU and LSTM archi
 tectures by an appropriate regularization\, and provide an overview of a r
 ecent\, structurally stable\, architecture known as Structured State-Space
  Model.\n\nBio\nFabio Bonassi is a postdoctoral researcher on “Machine L
 earning for Control” at the Uppsala University\, Sweden. He received th
 e M.Sc. degree and Ph.D. degree from Politecnico di Milano\, Italy\, in 20
 18 and 2023\, respectively. He is a recipient of the Dimitris N. Chorafas
  Ph.D. Award and of the “Claudio Maffezzoni” master thesis award. At t
 he 19th Symposium on System Identification\, he received the IFAC Best You
 ng Author Award.\nHis research interests are neural network architectures 
 for the identification and control of dynamic systems\, with a focus on th
 eir stability and robustness properties. 
LOCATION:ME C2 405 https://plan.epfl.ch/?room==ME%20C2%20405
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR