Lyceum: An efficient and scalable ecosystem for robot learning

Colin Summers, Kendall Lowrey, Aravind Rajeswaran, Siddhartha Srinivasa, Emanuel Todorov
Proceedings of the 2nd Conference on Learning for Dynamics and Control, PMLR 120:793-803, 2020.

Abstract

We introduce Lyceum, a high-performance computational ecosystem for robot learning. Lyceum is built on top of the Julia programming language and the MuJoCo physics simulator, combining the ease-of-use of a high-level programming language with the performance of native C. In addition,Lyceum has a straightforward API to support parallel computation across multiple cores and machines. Overall, depending on the complexity of the environment,Lyceum is 5-30X faster compared to other popular abstractions like OpenAI’s Gym and DeepMind’s dm-control. This substantially reduces training time for various reinforcement learning algorithms; and is also fast enough to support real-time model predictive control through MuJoCo. The code, tutorials, and demonstration videos can be found at: www.lyceum.ml.

Cite this Paper

BibTeX
@InProceedings{pmlr-v120-summers20a, title = {Lyceum: An efficient and scalable ecosystem for robot learning}, author = {Summers, Colin and Lowrey, Kendall and Rajeswaran, Aravind and Srinivasa, Siddhartha and Todorov, Emanuel}, booktitle = {Proceedings of the 2nd Conference on Learning for Dynamics and Control}, pages = {793--803}, year = {2020}, editor = {Bayen, Alexandre M. and Jadbabaie, Ali and Pappas, George and Parrilo, Pablo A. and Recht, Benjamin and Tomlin, Claire and Zeilinger, Melanie}, volume = {120}, series = {Proceedings of Machine Learning Research}, month = {10--11 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v120/summers20a/summers20a.pdf}, url = {https://proceedings.mlr.press/v120/summers20a.html}, abstract = {We introduce Lyceum, a high-performance computational ecosystem for robot learning. Lyceum is built on top of the Julia programming language and the MuJoCo physics simulator, combining the ease-of-use of a high-level programming language with the performance of native C. In addition,Lyceum has a straightforward API to support parallel computation across multiple cores and machines. Overall, depending on the complexity of the environment,Lyceum is 5-30X faster compared to other popular abstractions like OpenAI’s Gym and DeepMind’s dm-control. This substantially reduces training time for various reinforcement learning algorithms; and is also fast enough to support real-time model predictive control through MuJoCo. The code, tutorials, and demonstration videos can be found at: www.lyceum.ml.} }
Endnote
%0 Conference Paper %T Lyceum: An efficient and scalable ecosystem for robot learning %A Colin Summers %A Kendall Lowrey %A Aravind Rajeswaran %A Siddhartha Srinivasa %A Emanuel Todorov %B Proceedings of the 2nd Conference on Learning for Dynamics and Control %C Proceedings of Machine Learning Research %D 2020 %E Alexandre M. Bayen %E Ali Jadbabaie %E George Pappas %E Pablo A. Parrilo %E Benjamin Recht %E Claire Tomlin %E Melanie Zeilinger %F pmlr-v120-summers20a %I PMLR %P 793--803 %U https://proceedings.mlr.press/v120/summers20a.html %V 120 %X We introduce Lyceum, a high-performance computational ecosystem for robot learning. Lyceum is built on top of the Julia programming language and the MuJoCo physics simulator, combining the ease-of-use of a high-level programming language with the performance of native C. In addition,Lyceum has a straightforward API to support parallel computation across multiple cores and machines. Overall, depending on the complexity of the environment,Lyceum is 5-30X faster compared to other popular abstractions like OpenAI’s Gym and DeepMind’s dm-control. This substantially reduces training time for various reinforcement learning algorithms; and is also fast enough to support real-time model predictive control through MuJoCo. The code, tutorials, and demonstration videos can be found at: www.lyceum.ml.
APA
Summers, C., Lowrey, K., Rajeswaran, A., Srinivasa, S. & Todorov, E.. (2020). Lyceum: An efficient and scalable ecosystem for robot learning. Proceedings of the 2nd Conference on Learning for Dynamics and Control, in Proceedings of Machine Learning Research 120:793-803 Available from https://proceedings.mlr.press/v120/summers20a.html.

Related Material