Follow
GHEORGHE COMANICI
GHEORGHE COMANICI
Research Scientist, DeepMind
Verified email at deepmind.com
Title
Cited by
Cited by
Year
The option keyboard: Combining skills in reinforcement learning
A Barreto, D Borsa, S Hou, G Comanici, E Aygün, P Hamel, D Toyama, ...
Advances in Neural Information Processing Systems 32, 2019
1042019
What can i do here? a theory of affordances in reinforcement learning
K Khetarpal, Z Ahmed, G Comanici, D Abel, D Precup
International Conference on Machine Learning, 5243-5253, 2020
732020
Androidenv: A reinforcement learning platform for android
D Toyama, P Hamel, A Gergely, G Comanici, A Glaese, Z Ahmed, ...
arXiv preprint arXiv:2105.13231, 2021
552021
Optimal policy switching algorithms for reinforcement learning
G Comanici, D Precup
Proceedings of the 9th International Conference on Autonomous Agents and …, 2010
482010
On-the-fly algorithms for bisimulation metrics
G Comanici, P Panangaden, D Precup
2012 ninth international conference on quantitative evaluation of systems …, 2012
242012
Basis function discovery using spectral clustering and bisimulation metrics
G Comanici, D Precup
International Workshop on Adaptive and Learning Agents, 85-99, 2011
212011
Representation discovery for mdps using bisimulation metrics
S Ruan, G Comanici, P Panangaden, D Precup
Proceedings of the AAAI Conference on Artificial Intelligence 29 (1), 2015
172015
Temporally abstract partial models
K Khetarpal, Z Ahmed, G Comanici, D Precup
Advances in Neural Information Processing Systems 34, 1979-1991, 2021
122021
An empirical analysis of off-policy learning in discrete mdps
C Păduraru, D Precup, J Pineau, G Comănici
European Workshop on Reinforcement Learning, 89-102, 2013
122013
Knowledge representation for reinforcement learning using general value functions
G Comanici, D Precup, A Barreto, DK Toyama, E Aygün, P Hamel, ...
102018
Basis refinement strategies for linear value function approximation in MDPs
G Comanici, D Precup, P Panangaden
Advances in neural information processing systems 28, 2015
92015
Vision-Language Models as a Source of Rewards
K Baumli, S Baveja, F Behbahani, H Chan, G Comanici, S Flennerhag, ...
arXiv preprint arXiv:2312.09187, 2023
82023
What can I do here
K Khetarpal, Z Ahmed, G Comanici, D Abel, D Precup
A theory of affordances in reinforcement learning. arXiv [cs. LG], 2020
62020
AndroidEnv: A Reinforcement Learning Platform for Android. abs/2105.13231 (2021)
D Toyama, P Hamel, A Gergely, G Comanici, A Glaese, Z Ahmed, ...
arXiv preprint cs.LG/2105.13231, 2021
52021
Finding increasingly large extremal graphs with alphazero and tabu search
A Mehrabian, A Anand, H Kim, N Sonnerat, M Balog, G Comanici, ...
arXiv preprint arXiv:2311.03583, 2023
42023
A study of off-policy learning in computational sustainability
C Paduraru, D Precup, J Pineau, G Comanici
European Workshop on Reinforcement Learning (EWRL) 24, 89-102, 2012
42012
Learning how to Interact with a Complex Interface using Hierarchical Reinforcement Learning
G Comanici, A Glaese, A Gergely, D Toyama, Z Ahmed, T Jackson, ...
arXiv preprint arXiv:2204.10374, 2022
12022
Controlling computing devices using hierarchical agents
GT Comanici, AMC Glaese, A Gergely, Z Ahmed, T Jackson, D Precup
US Patent App. 18/105,180, 2024
2024
Representation discovery for Markov decision processes using behavioural similarity
G Comanici
McGill University (Canada), 2016
2016
Optimal Time Scales for Reinforcement Learning Behaviour Strategies
G Comanici, D Precup
McGill University, 2010
2010
The system can't perform the operation now. Try again later.
Articles 1–20