Pure Exploration

We look at structured active sequential hypothesis testing problems. We are interested in quantifying how many samples are necessary, and in finding near-optimal sampling strategies. One driving example is (stochastic) game tree search.

Optimal best-arm identification methods for tail-risk measures
Shubhada Agrawal, Wouter M. Koolen, Sandeep Juneja
ArXiv, August 2020. Submitted to Advances in Neural Information Processing Systems (NeurIPS).
Structure adaptive algorithms for stochastic bandits
Rémy Degenne, Han Shao, Wouter M. Koolen
In Proceedings of the 37th International Conference on Machine Learning (ICML), July 2020.
Non-asymptotic pure exploration by solving games
Rémy Degenne, Wouter M. Koolen, Pierre Ménard
In Advances in Neural Information Processing Systems (NeurIPS) 32, pages 14465–14474. Curran Associates, Inc., December 2019.
Pure exploration with multiple correct answers
Rémy Degenne, Wouter M. Koolen
In Advances in Neural Information Processing Systems (NeurIPS) 32, pages 14564–14573. Curran Associates, Inc., December 2019.
Sequential test for the lowest mean: From Thompson to Murphy sampling
Emilie Kaufmann, Wouter M. Koolen, Aurélien Garivier
In Advances in Neural Information Processing Systems (NeurIPS) 31, pages 6333–6343. Curran Associates, Inc., December 2018.
Mixture martingales revisited with applications to sequential tests and confidence intervals
Emilie Kaufmann, Wouter M. Koolen
Preprint, October 2018.
Monte-Carlo tree search by best arm identification
Emilie Kaufmann, Wouter M. Koolen
In Advances in Neural Information Processing Systems (NeurIPS) 30, pages 4904–4913, December 2017.
Maximin action identification: A new bandit framework for games
Aurélien Garivier, Emilie Kaufmann, Wouter M. Koolen
In Proceedings of the 29th Annual Conference on Learning Theory (COLT), pages 1028 – 1050, June 2016.