For Prospective Undergraduates

Welcome! 🎉

Preliminaries

if you are wondering whether you are …

capable:

1. PyTorch
2. GitHub
3. Weights & Biases
4. Advantage Actor Critic (A2C)
5. You have a relatively large amount of time.

A straightforward test: Reproduce A2C to complete any RL task (development needs to take place within a GitHub repository), and utilize Weights & Biases for experiment visualization and hyperparameter tuning.

interested in our topic:

Currently, we are focusing on information design (a subfield of computational economics and game theory) and multi-agent reinforcement learning (MARL). Our next few works will be based on:

• Information Design in Multi-Agent Reinforcement Learning.
  Yue Lin, Wenhao Li, Hongyuan Zha, Baoxiang Wang.
  arXiv 2023.

For more specific details, please contact me so that we can discuss in person.

Recommended Readings

Our communication will be smoother if you have read some of the following papers.

MARL in Sequential Social Dilemmas

1. MARL in Sequential Social Dilemmas
2. Melting Pot

Information Design

1. Bayesian Persuasion
2. Bayes Correlated Equilibrium and the Comparison of Information Structures in Games
3. Surveys:
   1. Bayesian Persuasion and Information Design
   2. Algorithmic Information Structure Design: A Survey

Mechanism Design in MARL

1. LIO
2. The AI Economist: Taxation

MARL Algorithms

1. MADDPG
2. COMA
3. QMIX

MARL Communication

1. RIAL,DIAL
2. CommNet