First workshop on multi-armed bandit co-organized by Lancaster University, Imperial College London and Google Deepmind

On 25^th and 26^th September, Imperial College hosted the first Multi Armed Bandit Workshop co-organized with the university of Lancaster and Google Deepmind.

Dr Grünewälder, one of the co-organizers from the university of Lancaster explains: “The multi-armed bandit problem has become classic because it is well suited for modelling many real world problems while being amenable to theoretical analyses due to its simplicity.

Researchers from different fields, especially statistics, operational research, applied probability and computer science, have made theoretical contributions that in the recent years started gaining wider popularity in practice. This workshop provides a forum for stimulating discussions, especially across the boundaries between computer science, statistics and operations research.”

The workshop, which was free of charge thanks to support from the London Mathematical Society, Google Deepmind, the Data Science Institute Lancaster, Amazon Research and our own Data Science Institute at Imperial College, included a series of talks by researchers from academia and companies, a poster session and a social event.

“We are very pleased with the turnout” – says Dr Andras Gyorgy, formerly at the department of Electrical and Electronic Engineering, now at Deepmind – “We had about 70 attendees and very lively discussions throughout the days”.