Algorithms or Actions? a Study in Large-Scale Reinforcement Learning
Abstract
Large state and action spaces are very challenging to reinforcement learning. However, in many domains there is a set of algorithms available, which estimate the best action given a state. Hence, agents can either directly learn a performance-maximizing mapping from states to actions, or from states to algorithms. We investigate several aspects of this dilemma, showing sufficient conditions for learning over algorithms to outperform over actions for a finite number of training iterations. We present synthetic experiments to further study such systems. Finally, we propose a function approximation approach, demonstrating the effectiveness of learning over algorithms in real-time strategy games.
Cite
Text
Tavares et al. "Algorithms or Actions? a Study in Large-Scale Reinforcement Learning." International Joint Conference on Artificial Intelligence, 2018. doi:10.24963/IJCAI.2018/377Markdown
[Tavares et al. "Algorithms or Actions? a Study in Large-Scale Reinforcement Learning." International Joint Conference on Artificial Intelligence, 2018.](https://mlanthology.org/ijcai/2018/tavares2018ijcai-algorithms/) doi:10.24963/IJCAI.2018/377BibTeX
@inproceedings{tavares2018ijcai-algorithms,
title = {{Algorithms or Actions? a Study in Large-Scale Reinforcement Learning}},
author = {Tavares, Anderson Rocha and Anbalagan, Sivasubramanian and Marcolino, Leandro Soriano and Chaimowicz, Luiz},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2018},
pages = {2717-2723},
doi = {10.24963/IJCAI.2018/377},
url = {https://mlanthology.org/ijcai/2018/tavares2018ijcai-algorithms/}
}