A Bayesian approach to sparse dynamic network identification

Date:
2011/10/11 16:30 - 17:45
Place:
吉田キャンパス工学部総合校舎111号室

Lecturer:
Alessandro Chiuso
Membership:
Padova University
Professional Affiliation:
Asociate Professor, Padova University

His research interest are mainly in Estimation, Identification Theory and Applications (subspace methods, stochastic realization, non-linear estimation, hybrid systems, adaptive optics), Computer Vision (structure from motion, texture and gait analysis) and Networked Estimation and Control.


Lecture Concept:
In this talk I shall review some recent work on ``Bayesian'' (linear) system identification and discuss extensions of this methodology for performing simultaneous identification and variable selection; this latter task is framed in the context of Granger causality graphs.
The proposed algorithms show very good performance on simulated data and have also been succesfully applied to real data for thermodinamic monitoring of buildings.
In order to understand some of the key features of this Bayesian variable selection problem we consider a simpler (group) linear regression problem and discuss the relation between group-Lasso (GLASSO), Multiple Kernel Leraning (MKL) and our method, as well with other Bayesian techniques developed in the statistics literature.
Our approach is nonconvex but one of its versions requires optimization with respect to only one scalar variable. Theoretical arguments, independent of the correctness of the priors entering the sparse model, are included to clarify the advantages of our nonconvex technique in comparison with the other two (MKL and GLASSO) convex estimators.
/Authors:/
*Alessandro Chiuso* and *Gianluigi Pillonetto*