When individuals are learning how to behave in an unknown environment, a statistically sensible thing to do is form posterior distributions over unknown quantities of interest (such as features of the environment and individuals’ preferences) then select an action by integrating with respect to these posterior distributions. However reasoning with such distributions is very troublesome, even in a machine learning context with extensive computational resources; Savage himself indicated that Bayesian decision theory is only sensibly used in reasonably “small” situations.
Random beliefs is a framework in which individuals instead respond to a single sample from a posterior distribution. This is a strategy known as Thompson sampling, after its introduction in a medical trials context by Thompson (1933), and is used by many Web providers both to select which adverts to show you and to perform website optimisation. I will demonstrate that such behaviour ‘solves’ the exploration-exploitation dilemma in a contextual bandit setting, which is the framework used by most current applications. I will also discuss more recent research related to online optimisation of websites.
15:15: Registration at Clarendon Wing Lecture Theatre
Clarendon Wing is linked to the Worsley Building by two walkways – one on Level 7 and one on Level 5. The level 7 walkway enters Clarendon Wing on Level D. The lecture theatre is on your left as you enter Clarendon Wing.
15:30: Mr Thomas Higgins- An age-period-cohort analysis under rapidly changing circumstances
15:45: Dr Mark Webster- Modelling and calibration for human cell dynamics
16:00: Professor David Leslie- Thompson sampling for website optimisation
17:00: Refreshments and networking
The talks will be followed by a networking reception in the LIDA staff room with drinks and nibbles.