RLgraph: Modular Computation Graphs for Deep Reinforcement Learning
MetadataShow full item record
Schaarschmidt, M., Mika, S., Fricke, K., & Yoneki, E. RLgraph: Modular Computation Graphs for Deep Reinforcement Learning. https://doi.org/10.17863/CAM.47584
Reinforcement learning (RL) tasks are challenging to implement, execute and test due to algorithmic instability, hyper-parameter sensitivity, and heterogeneous distributed communication patterns. We argue for the separation of logical component composition, backend graph definition, and distributed execution. To this end, we introduce RLgraph, a library for designing and executing reinforcement learning tasks in both static graph and define-by-run paradigms. The resulting implementations are robust, incrementally testable, and yield high performance across different deep learning frameworks and distributed backends.
cs.LG, cs.LG, cs.AI, stat.ML
This record's DOI: https://doi.org/10.17863/CAM.47584
This record's URL: https://www.repository.cam.ac.uk/handle/1810/300510
All rights reserved