/ EURO PhD School / Program

The goal of the school is to make PhD students in stochastic modeling familiar with Reinforcement Learning techniques by DIY in a hackathon-like workshop. The school will cover the following techniques (preliminary list): Stochastic Dynamic Programming, Monte-Carlo Simulation, Temporal Difference Learning, Value Function Approximation, Eligibility Traces, Policy-gradient methods, Bayesian Dynamic Programming, Multi-armed Bandits, Deep Reinforcement Learning. There will be lectures by specialists, both in-person and online, with time to work in small groups on various problems. Specialists will actively participate in Q&A sessions with the groups.

Technical program:

  • Sun 17 Jul: Arrival before dinner; Non-technical overview of RL and problems
  • Mon 18 Jul: Lectures on techniques; Discussion of techniques; Identification of problems to solve; Group formation
  • Tue 19 Jul: Implementation in groups; Interaction with specialists
  • Wed 20 Jul: Lectures on techniques; Discussion of techniques; Identification of problems to solve
  • Thu 21 Jul: Implementation in groups; Interaction with specialists
  • Fri 22 Jul: Lectures on techniques; Discussion of techniques; Identification of problems to solve
  • Sat 23 Jul: Implementation in groups; Interaction with specialists; Presentation of results
  • Sun 24 Jul: Evaluation of group work; Discussion of the future of RL for OR; Departure after lunch

Social program: There will be two organized social events complemented with several informal ones to allow for frequent interactions.

The detailed program can be found here.