Maxim Raginsky: The limits of control: An information-theoretic viewpoint
Presented February 22nd, 2013 at the University of Florida by the Laboratory for Cognition and Control in Complex Systems.
More info: http://ccc.centers.ufl.edu/?q=CCCW_Ma...
Visit us! http://ccc.centers.ufl.edu
Abstract:
Adaptive dynamical systems arise in a multitude of contexts, e.g., optimization, control, communications, signal processing, and machine learning. A precise characterization of their fundamental limitations is therefore of paramount importance. In this talk, I consider the general problem of adaptively controlling and/or identifying a stochastic dynamical system, where a priori knowledge allows us to place the system in a subset of a metric space (the uncertainty set). I will present an information-theoretic meta-theorem that captures the trade-off between the metric complexity (or richness) of the uncertainty set, the amount of information acquired online in the process of controlling and observing the system, and the residual uncertainty remaining after the observations have been collected. Following the approach of G. Zames, I quantify a priori information by the Kolmogorov (metric) entropy of the uncertainty set, while the information acquired online is expressed as a sum of information divergences. I will then use the meta-theorem to derive new minimax lower bounds on the metric identification error, as well as to give a simple derivation of the minimum time needed to stabilize an uncertain stochastic linear system.
Видео Maxim Raginsky: The limits of control: An information-theoretic viewpoint канала cccUF
More info: http://ccc.centers.ufl.edu/?q=CCCW_Ma...
Visit us! http://ccc.centers.ufl.edu
Abstract:
Adaptive dynamical systems arise in a multitude of contexts, e.g., optimization, control, communications, signal processing, and machine learning. A precise characterization of their fundamental limitations is therefore of paramount importance. In this talk, I consider the general problem of adaptively controlling and/or identifying a stochastic dynamical system, where a priori knowledge allows us to place the system in a subset of a metric space (the uncertainty set). I will present an information-theoretic meta-theorem that captures the trade-off between the metric complexity (or richness) of the uncertainty set, the amount of information acquired online in the process of controlling and observing the system, and the residual uncertainty remaining after the observations have been collected. Following the approach of G. Zames, I quantify a priori information by the Kolmogorov (metric) entropy of the uncertainty set, while the information acquired online is expressed as a sum of information divergences. I will then use the meta-theorem to derive new minimax lower bounds on the metric identification error, as well as to give a simple derivation of the minimum time needed to stabilize an uncertain stochastic linear system.
Видео Maxim Raginsky: The limits of control: An information-theoretic viewpoint канала cccUF
Показать
Комментарии отсутствуют
Информация о видео
Другие видео канала
Peter Caines: Recent Developments in Mean Field ControlGregory Puleo - Community Detection via Correlation Clustering Algorithm ImprovementsKarl McDermott: Regulatory Lag and How to Recover the Costs of New InfrastructureZap Q-learning with Nonlinear Function ApproximationAdithya M Devraj - Differential TD Learning to Approximate the Solution to Poisson's EquationYves Atchadé - Iterated Filtering AlgorithmsKarl McDermott: The Home as a Power PlantJohanna Mathieu: Harnessing Residential Loads for Demand ResponseMaxim Raginsky: Concentration of Measure Inequalities in Information Theory, Communications + CodingMuriel Medard - Going on Twenty - Network CodingProf. Joe Chow's lectureMichelle Effros: On Communication, Control, and the Cost of SeperationSean Meyn - Modeling Nature in Linearly Scalable MDPsMaryam Shanechi USC Next-Generation Brain-Machine Interface SystemsJean-Yves Le Boudec: Real Time Control of Distribution NetworksShaofeng Zou University of Buffalo Robust Reinforcement Learning Under Model UncertaintyVivek Borkar: Relative Value for Ergodic Control of DiffusionsEric Moulines: On stochastic proximal gradient algorithmsAnand Radhakrishnan - Variance reduction in MCMC using control variatesMurial Medard: Joint source, channel and network coding- a simple ad-hoc approach