BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Human and Robot Decision Making in Multi-Armed Bandits - naomi@pri
 nceton.edu
DTSTART:20190502T130000Z
DTEND:20190502T140000Z
UID:TALK124456@talks.cam.ac.uk
CONTACT:Alberto Padoan
DESCRIPTION:Decision-making in explore–exploit tasks\, from resource all
 ocation to search in an uncertain environment\, can be modeled using multi
 -armed bandit (MAB) problems\, where the decision-maker must choose sequen
 tially in time among multiple options with uncertain rewards. Rigorous exa
 mination of the heuristics that humans use in these tasks can help in desi
 gning and evaluating strategies for performance in a wide range of decisio
 n-making scenarios that involve humans\, robots\, or both. I will discuss 
 results from multi-armed bandit experiments with human participants and fe
 atures of human decision-making captured by a model that relies on Bayesia
 n inference\, confidence bounds\, and Boltzmann action selection. I will p
 resent extensions to satisfying objectives and to distributed cooperative 
 decision making in multi-player multi-armed bandit problems in which agent
 s communicate according to a network graph.  I will show demonstrations of
  robots implementing the algorithms to search for peaks over an uncertain 
 distributed resource field.\n
LOCATION:Cambridge University Engineering Department\, Lecture Theatre 6
END:VEVENT
END:VCALENDAR