The BeNeRL workshop will discuss novel research, work in progress, and more general 'open problems' in Reinforcement Learning (RL) in Belgium and The Netherlands. For example, what are the current bottlenecks, why does not everybody use RL, what are the main practical and theoretical obstacles?
Building on last year's successful Reinforcement Learning workshop in Brussels, we will have another edition this year on Friday November 27 at CWI in Amsterdam. The workshop is organised by Sander Bohte and Davide Zambrano (CWI), and is hosted by the CWI Life Sciences group.
For questions, please send an e-mail to D.Zambrano@cwi.nl. The workshop is freely accessible after registration.The program:
09:45: coffee
10:00: Diederik Roijers, UVA: “multi-objective POMDPs.”
10:30: Matthijs Spaan, TU Delft: “Planning under uncertainty in smart grids.”
11:00: Michael Kaisers: “Learning in coopetitive energy systems”
11:30-12:00 break
12:00: Bert Kappen, RU: “Learning Optimal Control”
12:30: Marco Wiering, RUG: “Comparing exploration strategies for solving Markov Decision Processes with Q-learning”.
13:00-14:15 Lunch
14:15: Tim de Bruin, TU Delft: “The importance of bad experiences in deep reinforcement learning for control”.
14:30: Mihail Mihaylov, VUB: “Maximising renewable energy consumption in smart grids.”
15:00: Yamisleydi Salgueiro Sicilia, VUB: “Optimizing Stainless Steel using Reinforcement Learning”.
15:30-16:00 break
16:00: Pieter Roelfsema, NIN/VU: “Plausible models of reinforcement learning”
16:30: Davide Zambrano, CWI: “Continuous deep-time neural reinforcement learning”
17:00-19:00: drinks & fingerfood at CWI Praethuis
