The AI Economist: Optimal Economic Policy Design via Two-level Deep Reinforcement Learning

@article{Zheng2021TheAE,
  title={The AI Economist: Optimal Economic Policy Design via Two-level Deep Reinforcement Learning},
  author={Stephan Zheng and Alexander Trott and Sunil Srinivasa and David C. Parkes and Richard Socher},
  journal={ERN: Efficiency; Optimal Taxation (Topic)},
  year={2021}
}
  • Stephan Zheng, Alexander Trott, +2 authors R. Socher
  • Published 5 August 2021
  • Computer Science, Economics
  • ERN: Efficiency; Optimal Taxation (Topic)
AI and reinforcement learning (RL) have improved many areas, but are not yet widely adopted in economic policy design, mechanism design, or economics at large. At the same time, current economic methodology is limited by a lack of counterfactual data, simplistic behavioral models, and limited opportunities to experiment with policies and evaluate behavioral responses. Here we show that machine-learning-based economic simulation is a powerful policy and mechanism design framework to overcome… Expand
2 Citations
Building a Foundation for Data-Driven, Interpretable, and Robust Policy Design using the AI Economist
TLDR
The AI Economist framework enables effective, flexible, and interpretable policy design using two-level reinforcement learning (RL) and data-driven simulations, and finds that log-linear policies trained using RL significantly improve social welfare, based on both public health and economic outcomes, compared to past outcomes. Expand
WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU
TLDR
WarpDrive is a fast and extensible multi-agent RL platform to significantly accelerate research and development and enables orders-of-magnitude faster RL compared to common implementations that blend CPU simulations and GPU models. Expand

References

SHOWING 1-10 OF 12 REFERENCES
Determination and reduction of translocator protein (TSPO) ligand rs6971 discrimination† †The authors declare no competing interests.
The 18 kDa translocator protein (TSPO) is a target for development of diagnostic imaging agents for glioblastoma and neuroinflammation.
Annual Review of Economics, (https://www.dropbox.com/s/ xca67zq04v03zqr/Stantcheva_Dynamic_Taxation_Final.pdf?dl=0) (2020)
  • 2020
Reinforcement Learning: An Introduction, en, Google-Books-ID: uWV0DwAAQBAJ
  • (MIT Press, Oct. 2018),
  • 2018
http://arxiv.org/abs/1806.04067) (June 2018)
  • [Cs], arXiv: 1806.04067,
  • 2018
https://budgetmodel.wharton.upenn.edu/issues/2018/ 2/6/w2018-1) (2018)
  • 2018
Kawachi
  • Inequality Matters : Report of the World Social Situation
  • 2013
The New Dynamic Public Finance (Princeton University Press
  • STU Student edition,
  • 2010
We thank Kathy Baxter for the ethical review
    developed the economic simulator, implemented the reinforcement learning platform, and performed experiments
      ...
      1
      2
      ...