Date : Nov. 28, 2019, 1 p.m. - Room :Salle du conseil

Rewards, Results and Challenges from Real-world Reinforcement Learning

OLIVER Bent - Oxford University

The talk will attempt to give an introduction and go deeper on tackling particular obstacles in applied reinforcement learning. While Oliver has recently been working on applying these approaches to the problem of malaria control strategy, the content will focus on this application but is extensible to other examples of reinforcement learning applied to computationally expensive simulators of real-world dynamics.  

Oliver is currently finishing writing his PhD thesis with Oxford’s Machine Learning Research Group. His research looks at applications of Machine Learning (ML) using heterogeneous sources of data, simulation and algorithms to inform decisions in the real-world. With a current focus on how ML may be assistive to Policy makers in achieving their goals towards the elimination of malaria. Having recently presented, demonstrated and published this work at AAAI 2018, NeurIPS 2018 (2019) and KDD 2019.  In other activities this year Oliver has run an introductory Machine Learning Lecture series in Nairobi Kenya, served as part of the judging panel for the IBM Watson AI Xprize $5 million award and become a first-time father.