Awesome My Interests

A curated list of my interested resources.

Contents

• Game AI
• Human-Centered AI
• Reinforcement Learning
  • Decision Transformer
  • Big Models for RL

Game AI

Industrial works

• DeepMind AlphaStar, 2019.
• OpenAI Five, 2019.
• Tencent Juewu, 2020.
• Learning Diverse Policies in MOBA Games via Macro-Goals, 2021.

Others

• awesome-game-ai - An awesome library.
• Game AI Pro - Game AI Pro book series.
• GDC Vault - A trove of in-depth design, technical and inspirational talks and slides from the influencers of the game development industry, taken from over 20 years of the worldwide Game Developers Conferences.
• ✅ AI in Games: Techniques, Challenges and Opportunities, 2021.
• Rethinking of AlphaStar, 2021

Human-Centered AI

• Awesome Human-Centered AI (HCAI) - An awesome library.

Reinforcement Learning

Learning Transferable Policies and Representations with Transformer

• AnyMorph: Learning Transferable Polices By Inferring Agent Morphology, ICML 2022 (Spotlight).
  • Related work: One Policy to Control Them All: Shared Modular Policies for Agent-Agnostic Control, ICML 2020.
• MetaMorph: Learning Universal Controllers with Transformers, 2022.
• Can Wikipedia Help Offline Reinforcement Learning?, 2022.

Decision Transformer

• ✅ Decision Transformer: Reinforcement Learning via Sequence Modeling, 2021.
• ✅ Offline Reinforcement Learning as One Big Sequence Modeling Problem, 2021.
• ✅ A Generalist Agent (Gato), 2022.
  -> Use prompt conditioning for multi-task.
  • ✅ GPT，GPT-2，GPT-3 论文精读
• Online Decision Transformer, ICML 2022 (Oral).
• ✅ Multi-Game Decision Transformers, 2022.
• You Can’t Count on Luck: Why Decision Transformers Fail in Stochastic Environments, 2022.
• Action-Conditioned Contrastive Policy Pretraining, 2022.
• Multi-agent
  • Offline Pre-trained Multi-Agent Decision Transformer: One Big Sequence Model Tackles All SMAC Tasks, 2021.
  • Multi-Agent Reinforcement Learning is a Sequence Modeling Problem, 2022.
• Generalized Decision Transformer for Offline Hindsight Information Matching, ICLR 2021 (Spotlight).
• Prompting Decision Transformer for Few-Shot Policy Generalization, ICML 2022 (Poster).
• Addressing Optimism Bias in Sequence Modeling for Reinforcement Learning, ICML 2022 (Poster).
• RvS: What is Essential for Offline RL via Supervised Learning?, ICLR 2022.

Offline RL with Online Finetuning

• Acceleratingonline reinforcement learning with offline datasets, CORR 2022.
• Offline Reinforcement Learning with Implicit Q-Learning, Arxiv 2021.
• Offline-to-Online Reinforcement Learning via Balanced Replay and Pessimistic Q-Ensemble, CoRL 2021.
• AW-Opt: Learning Robotic Skills with Imitation and Reinforcement at Scale, CoRL 2021.
• Online Decision Transformer, ICML 2022 (Oral).

Big Models for RL

• ✅ OpenAI VPT Learning to Play Minecraft with Video PreTraining (VPT), 2022.
  • 用到的Transformer模型：Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context, 2022.
• NVIDIA MineDojo MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge, 2022.
• LM-Nav: Robotic Navigation with Large Pre-Trained Models of Language, Vision, and Action, 2022.
• A Dataset Perspective on Offline Reinforcement Learning, 2022.

Real-World RL

• Challenges of Real-World Reinforcement Learning, 2019.
• Reinforcement Learning In Practice: Oppotunities And Challenges, 2022.

Autonomous Driving

• Deep Reinforcement Learning for Autonomous Driving: A Survey, 2020.

Environments

• DIAMBRA Arena: An interface towards popular arcade emulated video games.
• Bomberland: A multi-agent AI competition inspired by Bomberman.
• Honor of Kings Game Environment: The open environment of Honor of kings 1V1.

Empirical study

• What Matters In On-Policy Reinforcement Learning? A Large-Scale Empirical Study, 2020.
• Distilling Policy Distillation, 2019.