BEGIN:VCALENDAR
PRODID:-//Columba Systems Ltd//NONSGML CPNG/SpringViewer/ICal Output/3.3-
 M3//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20220105T160746Z
DTSTART:20220208T140000Z
DTEND:20220208T150000Z
SUMMARY:Turing Fellow 'Spotlight': 8 February
UID:{http://www.columbasystems.com/customers/uom/gpp/eventid/}huv-kwwamw0
 8-w1ii0o
DESCRIPTION:The Digital Futures team are hosting a series of online talks
  by The University of Manchester's new Turing Fellows for 2021-22.\n\nPr
 esenting at this event are:\n\nManuel López-Ibáñez\, Senior Lecturer in 
 Decision Sciences and Business Analytics.\n\nManuel's main expertise is 
 on the application of metaheuristics\, including local search\, evolutio
 nary algorithms and ant colony optimization\, to optimization problems\,
  including continuous\, combinatorial\, and multi-objective problems. Hi
 s current research is on the experimental analysis and automatic configu
 ration and tuning of stochastic optimization algorithms\, in particular\
 , when applied to multi-objective optimization problems.\n\nTitle and ab
 stract to follow.\n\nNeil Walton\, Reader\, Department of Mathematics.\n
 \nNeil has conducted research visits at Microsoft Research Cambridge\, t
 he Basque Centre for Mathematics and the Automatic Control Laboratory ET
 H Zurich. Neil is co-investigator on the Turing project 'Artificial Inte
 lligence for Transport Planners'. He is an associate editor at the journ
 als Operations Research and Operations Research Letters.\n\nNeil's talk 
 is titled 'Learning and Information in Stochastic Networks and Queues'\n
 \nWe review the role of information and learning in the stability and op
 timization of queueing systems. In recent years\, techniques from superv
 ised learning\, bandit learning and reinforcement learning have been app
 lied to queueing systems supported by increasing role of information in 
 decision making. We present observations and new results that help ratio
 nalize the application of these areas to queueing systems.\n\nWe prove t
 hat the MaxWeight and BackPressure policies are an application of Blackw
 ell's Approachability Theorem. This connects queueing theoretic results 
 with adversarial learning. We then discuss the requirements of statistic
 al learning for service parameter estimation. As an example\, we show ho
 w queue size regret can be bounded when applying a perceptron algorithm 
 to classify service. Next\, we discuss the role of state information in 
 improved decision making. Here we contrast the roles of epistemic inform
 ation (information on uncertain parameters) and aleatoric information (i
 nformation on an uncertain state). Finally we review recent advances in 
 the theory of reinforcement learning and queueing\, as well as\, provide
  discussion on current research challenges.
STATUS:TENTATIVE
TRANSP:TRANSPARENT
CLASS:PUBLIC
END:VEVENT
END:VCALENDAR