(The talk will be given in English)

Speaker:     Dr. Rami Katz  University of Trento

011 hall, Electrical Engineering-Kitot Building

Monday, June 25^th, 2023

15:00 - 16:00

Constrained system identification of reaction-diffusion equations: a bridge between control and inverse problems

Abstract

System identification uses mathematical methods to reconstruct models of systems from partial and noisy data. The identification of reaction-diffusion (RD) systems is a mathematically intricate and significant research domain within systems and control, and finds application across diverse fields, spanning from multi-agent systems to chemical reactors. Many identification algorithms necessitate an infinite number of measurements and often fail to yield explicit bounds on recovery errors, in the presence of noise. In this talk, I will consider the problem of recovering the first dominant modes and initial condition of an unknown one-dimensional RD system from a finite set of filtered (noisy) state measurements. We demonstrate that Prony’s method for spike deconvolution, commonly applied in super-resolution problems, is suitable for solving this identification task. Moreover, building on recent results on the error sensitivity of Prony’s method, we derive new explicit identification guarantees, by analyzing the first-order condition numbers with respect to the measurement filtering error and showing their (super) exponential decay to zero in several identification regimes. The developed tools establish a hitherto unexplored connection between systems and control theory and inverse problems.

Joint with D. Batenkov from the School of Mathematics at Tel Aviv University.

Short Bio

Rami Katz is a post-doctoral researcher at the University of Trento, Italy. Rami recieved his B.Sc. and M.Sc in applied mathematics and his Ph.D. in electrical engineering from Tel Aviv University. He is the recipient of the KLA and Weinstein fellowships, as well as several excellence awards for both studies and teaching. He is an ECC21 best student paper award finalist and the recipient of the June 2020 editor's choice in Automatica. His research interests include control of nonlinear and distributed parameter system, systems biology and inverse problems in signal processing and control. 

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בטופס הנוכחות שיועבר באולם במהלך הסמינר

(The talk will be given in English)

Speaker:     Prof. Alexander Fradkov 

Institute for Problems of Mechanical Engineering of RAS and Saint Petersburg University

011 hall, Electrical Engineering-Kitot Building

Wednesday, December 20^th, 2023

11:00 - 12:00

A Brief History of Adaptation and learning

Abstract

Machine learning and artificial intelligence have attracted a lot of attention during recent years. They are applied to various  new problems and it looks like they are based upon completely new ideas in the applied science. However, there exist strong links between  machine learning and lassical adaptation methods which are much lesser known and almost not exploited nowadays.

In this talk a brief overview of the historical evolution of the machine learning field will be presented and its relations to adaptation, optimization and adaptive control will be discussed.

A number of little-known facts published in hard-to-reach sources are presented.

Short Bio

Alexander L. Fradkov received the Diploma degree in mathematics from Leningrad (currently St.Petersburg) State University in 1971, the Candidate of Sciences (Ph.D.) degree in technical cybernetics in 1975 from Leningrad Mechanical Institute, and the Doctor of Sciences (Habilitation) degree in control engineering in 1986 from Leningrad Electrotechnical University. He is currently the Head of the “Control of Complex Systems” Lab of the Institute for Problems in Mechanical Engineering of Russian Academy of Sciences and Professor of the Department of Theoretical Cybernetics at Saint Petersburg University.

Prof. Fradkov is the coauthor of more than 800 journal and conference papers, 18 books and textbooks, and a holder of ten patents. In his book "Cybernetical Physics" (Nauka, 2003; Springer-Verlag, 2007) an emerging boundary field between Physics and Control areas is pioneered. His research interests include nonlinear and adaptive and learning control, control of oscillatory and chaotic systems, dynamics and control of complex physical systems and networks. Prof. Fradkov was recipient of the best paper award from the Elsevier journal Annual Reviews of Control for 2020-2022. He is IEEE and IFAC.

השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בטופס הנוכחות שיועבר באולם במהלך הסמינר