89 Hits in 2.0 sec

RLlib: Abstractions for Distributed Reinforcement Learning [article]

Eric Liang, Richard Liaw, Philipp Moritz, Robert Nishihara, Roy Fox, Ken Goldberg, Joseph E. Gonzalez, Michael I. Jordan, Ion Stoica
2018 arXiv   pre-print
Reinforcement learning (RL) algorithms involve the deep nesting of highly irregular computation patterns, each of which typically exhibits opportunities for distributed computation.  ...  We demonstrate the benefits of this principle through RLlib: a library that provides scalable software primitives for RL.  ...  Abstractions for Reinforcement Learning To leverage RLlib for distributed execution, algorithms must declare their policy π, experience postprocessor ρ, and loss L.  ... 
arXiv:1712.09381v4 fatcat:ihhwdewi4bfndags5x5c65mfaa

RLlib Flow: Distributed Reinforcement Learning is a Dataflow Problem [article]

Eric Liang, Zhanghao Wu, Michael Luo, Sven Mika, Joseph E. Gonzalez, Ion Stoica
2021 arXiv   pre-print
We propose RLlib Flow, a hybrid actor-dataflow programming model for distributed RL, and validate its practicality by porting the full suite of algorithms in RLlib, a widely adopted distributed RL library  ...  Researchers and practitioners in the field of reinforcement learning (RL) frequently leverage parallel computation, which has led to a plethora of new algorithms and systems in the last few years.  ...  In contrast to supervised learning, it is more difficult to provide a fixed set of abstractions for scaling RL training.  ... 
arXiv:2011.12719v4 fatcat:o7euvwohgrgtrazko3niasln4e

MALib: A Parallel Framework for Population-based Multi-agent Reinforcement Learning [article]

Ming Zhou, Ziyu Wan, Hanjing Wang, Muning Wen, Runzhe Wu, Ying Wen, Yaodong Yang, Weinan Zhang, Jun Wang
2021 arXiv   pre-print
Population-based multi-agent reinforcement learning (PB-MARL) refers to the series of methods nested with reinforcement learning (RL) algorithms, which produces a self-generated sequence of tasks arising  ...  both training and sampling, and meets the evaluation requirement of auto-curriculum learning; (3) a higher-level abstraction of MARL training paradigms, which enables efficient code reuse and flexible  ...  Yi Qu for her help on the design work.  ... 
arXiv:2106.07551v1 fatcat:xrdy7ppkbbfd3ogjjci3ace4wi

RL: Generic reinforcement learning codebase in TensorFlow

Bryan Li, Alexander Cowen-Rivers, Piotr Kozakowski, David Tao, Siddhartha Kamalakara, Nitarshan Rajkumar, Hariharan Sezhiyan, Sicong Huang, Aidan Gomez
2019 Journal of Open Source Software  
Vast reinforcement learning (RL) research groups, such as DeepMind and OpenAI, have their internal (private) reinforcement learning codebases, which enable quick prototyping and comparing of ideas to many  ...  Currently, there does not exist any RL codebase, to the author's knowledge, which contains all the five properties, particularly with TensorBoard logging and abstracting away cloud hardware such as TPU's  ...  Acknowledgements We would like to thank all other members of, for useful discussions and feedback.  ... 
doi:10.21105/joss.01524 fatcat:5jnp7brex5aydi2aswx2lcpvyu

Motion Planning by Reinforcement Learning for an Unmanned Aerial Vehicle in Virtual Open Space with Static Obstacles [article]

Sanghyun Kim, Jongmin Park, Jae-Kwan Yun, Jiwon Seo
2020 arXiv   pre-print
In this study, we applied reinforcement learning based on the proximal policy optimization algorithm to perform motion planning for an unmanned aerial vehicle (UAV) in an open space with static obstacles  ...  As the reinforcement learning progressed, the mean reward and goal rate of the model were increased.  ...  We used RLlib [30] to perform reinforcement learning in the Gym environment.  ... 
arXiv:2009.11799v1 fatcat:mi7vrghiyfeztexefbuqcl7uhu

RLgraph: Modular Computation Graphs for Deep Reinforcement Learning [article]

Michael Schaarschmidt, Sven Mika, Kai Fricke, Eiko Yoneki
2019 arXiv   pre-print
To this end, we introduce RLgraph, a library for designing and executing reinforcement learning tasks in both static graph and define-by-run paradigms.  ...  Reinforcement learning (RL) tasks are challenging to implement, execute and test due to algorithmic instability, hyper-parameter sensitivity, and heterogeneous distributed communication patterns.  ...  Further, we thank Lasse Espeholt for providing support to replicate IMPALA results. We also want to thank the RLlib authors for helping to replicate Ape-X results.  ... 
arXiv:1810.09028v2 fatcat:7q4mc2z4njdsnmrnfrriqqkfw4

PowerGridworld: A Framework for Multi-Agent Reinforcement Learning in Power Systems [article]

David Biagioni, Xiangyu Zhang, Dylan Wald, Deepthi Vaidhynathan, Rohit Chintala, Jennifer King, Ahmed S. Zamzam
2021 arXiv   pre-print
with existing training frameworks for reinforcement learning (RL).  ...  To highlight PowerGridworld's key features, we present two case studies and demonstrate learning MARL policies using both OpenAI's multi-agent deep deterministic policy gradient (MADDPG) and RLLib's proximal  ...  RLLib can be deployed on both cloud and high performance computing (HPC) systems, and it provides a number of training "abstractions," enabling users to develop custom, distributed RL algorithms.  ... 
arXiv:2111.05969v1 fatcat:jisj34ccbnaghldi2oiobllfk4

Mava: a research framework for distributed multi-agent reinforcement learning [article]

Arnu Pretorius, Kale-ab Tessera, Andries P. Smit, Claude Formanek, St John Grimbly, Kevin Eloff, Siphelele Danisa, Lawrence Francis, Jonathan Shock, Herman Kamper, Willie Brink, Herman Engelbrecht (+2 others)
2021 arXiv   pre-print
Breakthrough advances in reinforcement learning (RL) research have led to a surge in the development and application of RL.  ...  Mava provides useful components, abstractions, utilities and tools for MARL and allows for simple scaling for multi-process system training and execution, while providing a high level of flexibility and  ...  In this work, we present Mava, a research framework for distributed multi-agent reinforcement learning.  ... 
arXiv:2107.01460v1 fatcat:nuymlvbeendtpe3thy77qqp4ba

Learning to Fly – a Gym Environment with PyBullet Physics for Reinforcement Learning of Multi-agent Quadcopter Control [article]

Jacopo Panerati
2021 arXiv   pre-print
Vice versa, many reinforcement learning environments trade-off realism for high sample throughputs in toy-like problems.  ...  We demonstrate its use through several examples, either for control (trajectory tracking with PID control, multi-robot flight with downwash, etc.) or reinforcement learning (single and multi-agent stabilization  ...  We also thank the Vector Institute for providing access to its computing resources.  ... 
arXiv:2103.02142v3 fatcat:vzgoqo2sxja7tgjv3oc4o4bqoy

Running Alchemist on Cray XC and CS Series Supercomputers: Dask and PySpark Interfaces, Deployment Options, and Data Transfer Times [article]

Kai Rothauge, Haripriya Ayyalasomayajula, Kristyn J. Maschhoff, Michael Ringenburg, Michael W. Mahoney
2019 arXiv   pre-print
We also briefly discuss the combination of Alchemist with RLlib, an increasingly popular library for reinforcement learning, and consider the benefits of leveraging HPC simulations in reinforcement learning  ...  Alchemist is a system that allows Apache Spark to achieve better performance by interfacing with HPC libraries for large-scale distributed computations.  ...  matrix using the [VC, STAR] layout and a 100MB buffer is used for the messages, which is Alchemist's default setting.  ... 
arXiv:1910.01354v2 fatcat:5ho4jp775fc2rftmyk5phpwivi

Bellman: A Toolbox for Model-Based Reinforcement Learning in TensorFlow [article]

John McLeod, Hrvoje Stojic, Vincent Adam, Dongho Kim, Jordi Grau-Moya, Peter Vrancx, Felix Leibfried
2021 arXiv   pre-print
In the past decade, model-free reinforcement learning (RL) has provided solutions to challenging domains such as robotics.  ...  This demonstrates the necessity for a toolbox to push the boundaries for model-based RL.  ...  RLlib: Abstractions for distributed reinforcement learning.  ... 
arXiv:2103.14407v2 fatcat:c7mzfbm7w5g4vapexxzj4p66li

SURREAL-System: Fully-Integrated Stack for Distributed Deep Reinforcement Learning [article]

Linxi Fan, Yuke Zhu, Jiren Zhu, Zihua Liu, Orien Zeng, Anchit Gupta, Joan Creus-Costa, Silvio Savarese, Li Fei-Fei
2019 arXiv   pre-print
We present an overview of SURREAL-System, a reproducible, flexible, and scalable framework for distributed reinforcement learning (RL).  ...  The learning performances of our distributed algorithms establish new state-of-the-art on OpenAI Gym and Robotics Suites tasks.  ...  We propose SURREAL-SYSTEM, a reproducible, flexible, and scalable framework for distributed reinforcement learning. 2.  ... 
arXiv:1909.12989v2 fatcat:l7w77vhs4nfu7fjhw3hdn5iksy

Verified Probabilistic Policies for Deep Reinforcement Learning [article]

Edoardo Bacci, David Parker
2022 arXiv   pre-print
Deep reinforcement learning is an increasingly popular technique for synthesising policies to control an agent's interaction with its environment.  ...  In this paper, we tackle the problem of verifying probabilistic policies for deep reinforcement learning, which are used to, for example, tackle adversarial environments, break symmetries and manage trade-offs  ...  Conclusion We presented an approach for verifying probabilistic policies for deep reinforcement learning agents.  ... 
arXiv:2201.03698v1 fatcat:6q6tle2d45aphn6gicqci7h5f4

Wield: Systematic Reinforcement Learning With Progressive Randomization [article]

Michael Schaarschmidt, Kai Fricke, Eiko Yoneki
2019 arXiv   pre-print
Reinforcement learning frameworks have introduced abstractions to implement and execute algorithms at scale.  ...  We present Wield, a first-of-its kind system to facilitate task design for practical reinforcement learning.  ...  We are also grateful for receiving research credits from Google Cloud.  ... 
arXiv:1909.06844v1 fatcat:hsaisvqntzex7ihfxzmhhmfalm

Catalyst.RL: A Distributed Framework for Reproducible RL Research [article]

Sergey Kolesnikov, Oleksii Hrinchuk
2019 arXiv   pre-print
Despite the recent progress in deep reinforcement learning field (RL), and, arguably because of it, a large body of work remains to be done in reproducing and carefully comparing different RL algorithms  ...  Main features of our library include large-scale asynchronous distributed training, easy-to-use configuration files with the complete list of hyperparameters for the particular experiments, efficient implementations  ...  ., 2018] presented Dopamine 5 , a research framework for fast prototyping of reinforcement learning algorithms in TensorFlow.  ... 
arXiv:1903.00027v1 fatcat:4sdbs2r4f5hsxoe4kok4kpvpdi
« Previous Showing results 1 — 15 out of 89 results