Home * People * Sylvain Gelly

Sylvain Gelly ^[1]

Sylvain Gelly,
a French computer scientist, deep learning researcher at Google Brain Zurich ^[2], and former member of the Learning and Optimisation Group (A&O) ^[3] in the Laboratoire de recherche en informatique (LRI) under the direction of Michèle Sebag and Nicolas Bredèche at Paris-Sud 11 University. He defended his Ph.D. thesis A Contribution to Reinforcement Learning; Application to Computer Go in 2007 ^[4]. His research interests covers machine learning and he is one of the authors of OpenDP ^[5], a general and featured framework of reinforcement learning, and in particular of dynamic programming, and co-authored the top level Go playing program Mogo, using Monte-Carlo Tree Search which uses patterns in the simulations and improvements in UCT ^[6]^[7].

Selected Publications

2005 ...

Sylvain Gelly, Nicolas Bredèche, Michèle Sebag (2005). From Factorial and Hierarchical HMM to Bayesian Network : A Representation Change Algorithm. Proceedings of the Symposium on Abstraction, Reformulation and Approximation 2005, p107-120 (SARA 2005). Reprinted in Lecture Notes in Computer Science, Vol. 3607

2006

Sylvain Gelly, Olivier Teytaud, Nicolas Bredèche, Marc Schoenauer (2006). Universal Consistency and Bloat in GP. Some theoretical considerations about Genetic Programming from a Statistical Learning Theory viewpoint. pdf (draft)
Sylvain Gelly, Olivier Teytaud (2006). Bayesian networks : a better than frequentist approach for parametrization, and a more accurate structural complexity measure than the number of parameters. pdf (draft)
Sylvain Gelly, Jérémie Mary, Olivier Teytaud (2006). Learning for stochastic dynamic programming. pdf
Sylvain Gelly, Jérémie Mary, Olivier Teytaud (2006). On the ultimate convergence rates for isotropic algorithms and the best choices among various forms of isotropy. PPSN, 2006, pdf
Olivier Teytaud, Sylvain Gelly (2006). General lower bounds for evolutionary algorithms. pdf
Sylvain Gelly, Yizao Wang (2006). Exploration exploitation in Go: UCT for Monte-Carlo Go. pdf
Sylvain Gelly, Yizao Wang, Rémi Munos, Olivier Teytaud (2006). Modiﬁcation of UCT with Patterns in Monte-Carlo Go. INRIA

2007

Yizao Wang, Sylvain Gelly (2007). Modifications of UCT and Sequence-Like Simulations for Monte-Carlo Go. IEEE Symposium on Computational Intelligence and Games, Honolulu, USA, 2007, pdf
Sylvain Gelly, Yizao Wang (2007). MoGo wins 19x19 Go tournament. ICGA Journal, Vol. 30, No. 2 » 12th Computer Olympiad
Sylvain Gelly (2007). A Contribution to Reinforcement Learning; Application to Computer Go. Ph.D. thesis, pdf
Sylvain Gelly, David Silver (2007). Combining Online and Offline Knowledge in UCT. pdf
Sylvain Gelly, Olivier Teytaud, Jérémie Mary (2007). Active learning in regression, with application to stochastic dynamic programming. ICINCO and CAP, 2007, pdf

2008

Sylvain Gelly, David Silver (2008). Achieving Master Level Play in 9 x 9 Computer Go. pdf
Guillaume Chaslot, Louis Chatriot, Christophe Fiter, Sylvain Gelly, Jean-Baptiste Hoock, Julien Pérez, Arpad Rimmel, Olivier Teytaud (2008). Combining expert, offline, transient and online knowledge in Monte-Carlo exploration. pdf
Sylvain Gelly, Jean-Baptiste Hoock, Arpad Rimmel, Olivier Teytaud, Yann Kalemkarian (2008). The Parallelization of Monte-Carlo Planning - Parallelization of MC-Planning. ICINCO-ICSO 2008: 244-249, pdf, slides as pdf

2009