Building a Foundation for Data-Driven, Interpretable, and Robust Policy Design using the AI Economist

  title={Building a Foundation for Data-Driven, Interpretable, and Robust Policy Design using the AI Economist},
  author={Alexander Trott and Sunil Srinivasa and Douwe van der Wal and Sebastien Haneuse and Stephan Zheng},
  journal={Machine Learning eJournal},
Optimizing economic and public policy is critical to address socioeconomic issues and trade-offs, e.g., improving equality, productivity, or wellness, and poses a complex mechanism design problem. A policy designer needs to consider multiple objectives, policy levers, and behavioral responses from strategic actors who optimize for their individual objectives. Moreover, real-world policies should be explainable and robust to simulation-to-reality gaps, e.g., due to calibration issues. Existing… Expand
1 Citations
WarpDrive: Extremely Fast End-to-End Deep Multi-Agent Reinforcement Learning on a GPU
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


The AI Economist: Improving Equality and Productivity with AI-Driven Tax Policies
This work proposes a two-level deep reinforcement learning approach to learn dynamic tax policies, based on economic simulations in which both agents and a government learn and adapt, and shows that AI-driven tax policies perform strongly in the face of emergent tax-gaming strategies learned by AI agents. Expand
The AI Economist: Optimal Economic Policy Design via Two-level Deep Reinforcement Learning
It is demonstrated for the first time that two-level, deep RL can be used for understanding and as a complement to theory for economic design, unlocking a new computational learning-based approach to understanding economic policy. Expand
A Simple Planning Problem for COVID-19 Lockdown
We study the optimal lockdown policy for a planner who wants to control the fatalities of a pandemic while minimizing the output costs of the lockdown. We use the SIR epidemiology model and a linearExpand
Optimal Targeted Lockdowns in a Multi-Group Sir Model
It is found that optimal policies differentially targeting risk/age groups significantly outperform optimal uniform policies and most of the gains can be realized by having stricter lockdown policies on the oldest group. Expand
Agent-Based and Individual-Based Modeling: A Practical Introduction
This innovative textbook gives students and scientists the skills to design, implement, and analyze agent-based models and provides an introduction to NetLogo, an easy-to-use, free, and powerful software platform. Expand
Simulating Business Cash Flow Taxation: An Illustration Based on the “Better Way” Corporate Tax Reform
The U.S., according to some measures, has one of the highest marginal effective corporate tax rates (METRs) of any developed country. Yet the tax collects less than 2 percent of GDP. This paperExpand
The Cost of Staying Open: Voluntary Social Distancing and Lockdowns in the US
In combating the spread of COVID-19, some governments have been reluctant to adopt lockdown policies due to their perceived economic costs. Such costs can, however, arise even in the absence ofExpand
The econometrics of DSGE models
This paper reviews the literature on the formulation and estimation of dynamic stochastic general equilibrium (DSGE) models with a special emphasis on Bayesian methods, and briefly introduces some of the techniques required to compute and estimate these models. Expand
Covid-19 and Remote Work: An Early Look at Us Data
We report the results of a nationally-representative sample of the US population during the COVID-19 pandemic. The survey ran in two waves from April 1-5, 2020 and May 2-8, 2020. Of those employedExpand
Innovations in Multi-Agent Systems and Applications - 1
Reading innovations in multi agent systems and application1 is also a way as one of the collective books that gives many advantages. Expand