Awesome Decision Transformer

This is a collection of research papers for Decision Transformer (DT). And the repository will be continuously updated to track the frontier of DT.

Welcome to follow and star!

Table of Contents

• A Taxonomy of DT Algorithms

• Papers

  • Arxiv
  • ICML 2022 (New!!!)
  • AAAI 2022
  • ICLR 2022
  • NeurIPS 2021
  • ICML 2021

• Contributing

Overview of Transformer

The Decision Transformer was proposed by “Decision Transformer: Reinforcement Learning via Sequence Modeling” by Chen L. et al. It casts (offline) Reinforcement Learning as a conditional-sequence modeling problem.

Specifically, DT model is a causal transformer model conditioned on the desired return, (past) states, and actions to generate future actions in an autoregressive manner.

Advantage

1. Bypass the need for bootstrapping for long term credit assignment
2. Avoid undesirable short-sighted behaviors due to the discounting future rewards.
3. Enjoy the transformer models widely used in language and vision, which are easy to scale and adapt to multi-modal data.

Papers

format:
- [title](paper link) [links]
  - author1, author2, and author3...
  - publisher
  - key 
  - code 
  - experiment environment

Arxiv

• Multi-Game Decision Transformers

  • Kuang-Huei Lee, Ofir Nachum, Mengjiao Yang, Lisa Lee, Daniel Freeman, Winnie Xu, Sergio Guadarrama, Ian Fischer, Eric Jang, Henryk Michalewski, Igor Mordatch
  • Key: Multi-Task, Finetuning
  • Code: official
  • ExpEnv: Atari, REM

• Offline Pre-trained Multi-Agent Decision Transformer: One Big Sequence Model Tackles All SMAC Tasks

  • Linghui Meng, Muning Wen, Yaodong Yang, Chenyang Le, Xiyun Li, Weinan Zhang, Ying Wen, Haifeng Zhang, Jun Wang, Bo Xu
  • Key: Multi-Agent RL
  • Code: official
  • ExpEnv: SMAC

• Transfer learning with causal counterfactual reasoning in Decision Transformers

  • Ayman Boustati, Hana Chockler, Daniel C. McNamee
  • Key: Causal reasoning, Transfer Learning
  • ExpEnv: MINIGRID

• Pretraining for Language Conditioned Imitation with Transformers

  • Aaron L Putterman, Kevin Lu, Igor Mordatch, Pieter Abbeel
  • Key: Text-Conditioned Decision
  • ExpEnv: Text-Conditioned Frostbite (MultiModal Benchmark)

• An Offline Deep Reinforcement Learning for Maintenance Decision-Making

  • Hamed Khorasgani, Haiyan Wang, Chetan Gupta, Ahmed Farahat
  • Publisher: Annual Conference of the PHM Society 2021
  • Key: Offline Supervised RL, Remaining Useful Life Estimation
  • ExpEnv: NASA C-MAPSS

• A Sequence Modelling Approach to Question Answering in Text-Based Games

  • Gregory Furman, Edan Toledo, Jonathan Shock, Jan Buys
  • Publisher: Proceedings of the 3rd Wordplay: When Language Meets Games Workshop (Wordplay 2022)
  • Key: VQA
  • ExpEnv: QAIT

• Can Wikipedia Help Offline Reinforcement Learning?

  • Machel Reid, Yutaro Yamada, Shixiang Shane Gu
  • Key: VLN, Transfer Learning
  • Code: official
  • ExpEnv: MuJoco, D4RL, Atari

• Switch Trajectory Transformer with Distributional Value Approximation for Multi-Task Reinforcement Learning

  • Qinjie Lin, Han Liu, Biswa Sengupta
  • Key: Multi-Task RL, Sparse Reward
  • ExpEnv: MINIGRID

• Bootstrapped Transformer for Offline Reinforcement Learning

  • Kerong Wang, Hanye Zhao, Xufang Luo, Kan Ren, Weinan Zhang, Dongsheng Li
  • Key: Generation model
  • Code: official
  • ExpEnv: MuJoco, D4RL, Adroit

• Deep Transformer Q-Networks for Partially Observable Reinforcement Learning

  • Kevin Esslinger, Robert Platt, Christopher Amato
  • Key: POMDP, Transformer Q-Networks
  • ExpEnv: GV, Car Flag

• Multi-Agent Reinforcement Learning is a Sequence Modeling Problem

  • Muning Wen, Jakub Grudzien Kuba, Runji Lin, Weinan Zhang, Ying Wen, Jun Wang, Yaodong Yang
  • Key: Multi-Agent RL
  • ExpEnv: SMAC, MA MuJoco

• Transformers are Adaptable Task Planners

  • Vidhi Jain, Yixin Lin, Eric Undersander, Yonatan Bisk, Akshara Rai
  • Key: Task Planning, Prompt, Control, Generalization
  • Code: official
  • ExpEnv: Dishwasher Loading

• You Can't Count on Luck: Why Decision Transformers Fail in Stochastic Environments

  • Keiran Paster, Sheila McIlraith, Jimmy Ba
  • Key: Stochastic Environments
  • ExpEnv: Gambling, Connect Four, 2048

• When does return-conditioned supervised learning work for offline reinforcement learning?

  • David Brandfonbrener, Alberto Bietti, Jacob Buckman, Romain Laroche, Joan Bruna
  • Key: Theoretical analysis
  • ExpEnv: MuJoco, D4RL

• SimStu-Transformer: A Transformer-Based Approach to Simulating Student Behaviour

  • Zhaoxing Li, Lei Shi, Alexandra Cristea, Yunzhan Zhou, Chenghao Xiao, Ziqi Pan
  • Key: Intelligent Tutoring System

ICML 2022

• Online Decision Transformer

  • Qinqing Zheng, Amy Zhang, Aditya Grover
  • Publisher: ICML 2022 (Oral)
  • Key: Online finetuning, Max-entropy, Exploration
  • Code: unofficial
  • ExpEnv: MuJoco, D4RL

• Prompting Decision Transformer for Few-Shot Policy Generalization

  • Mengdi Xu, Yikang Shen, Shun Zhang, Yuchen Lu, Ding Zhao, Joshua B. Tenenbaum, Chuang Gan
  • Publisher: ICML 2022 (Poster)
  • Key: Prompt, Few-shot, Generalization
  • Code: official (released soon)
  • ExpEnv: DMC

• Addressing Optimism Bias in Sequence Modeling for Reinforcement Learning

  • Adam R Villaflor, Zhe Huang, Swapnil Pande, John M Dolan, Jeff Schneider
  • Publisher: ICML 2022 (Poster)
  • Key: World model
  • Code: official (released soon)
  • ExpEnv: CARLA

• AnyMorph: Learning Transferable Polices By Inferring Agent Morphology

  • Brandon Trabucco, Mariano Phielipp, Glen Berseth
  • Publisher: ICML 2022 (Poster)
  • Key: Morphology, Transfer Learning, Zero Shot
  • ExpEnv: Modular-RL

AAAI 2022

• Dreaming with Transformers
  • Catherine Zeng, Jordan Docter, Alexander Amini, Igor Gilitschenski, Ramin Hasani, Daniela Rus
  • Publisher: AAAI 2022 (RLG Workshop)
  • Key: Dreamer, World Model
  • ExpEnv: Deepmind Lab, VISTA

ICLR 2022

• Learning Transferable Policies By Inferring Agent Morphology

  • Brandon Trabucco, Mariano Phielipp, Glen Berseth
  • Publisher: ICLR 2022 (GPL Workshop Poster)
  • Key: Morphology, Transfer Learning, Zero Shot
  • ExpEnv: Modular-RL

• Silver-Bullet-3D at ManiSkill 2021: Learning-from-Demonstrations and Heuristic Rule-based Methods for Object Manipulation

  • Yingwei Pan, Yehao Li, Yiheng Zhang, Qi Cai, Fuchen Long, Zhaofan Qiu, Ting Yao, Tao Mei
  • Publisher: ICLR 2022 (GPL Workshop Poster)
  • Key: Object Manipulation
  • Code: official
  • ExpEnv: ManiSkill

• Generalized Decision Transformer for Offline Hindsight Information Matching

  • Hiroki Furuta, Yutaka Matsuo, Shixiang Shane Gu
  • Publisher: ICLR 2021 (Spotlight)
  • Key: HIM, SMM
  • Code: official
  • ExpEnv: MuJoco, D4RL

NeurIPS 2021

• Decision Transformer: Reinforcement Learning via Sequence Modeling

  • Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch
  • Publisher: NeurIPS 2021 (Poster)
  • Key: Conditional sequence modeling
  • Code: official, DI-engine
  • ExpEnv: MuJoco, D4RL, Atari

• Offline Reinforcement Learning as One Big Sequence Modeling Problem

  • Michael Janner, Qiyang Li, Sergey Levine
  • Publisher: NeurIPS 2021 (Spotlight)
  • Key: Conditional sequence modeling, Discretization
  • Code: official
  • ExpEnv: MuJoco, D4RL

• TransDreamer: Reinforcement Learning with Transformer World Models

  • Chang Chen, Yi-Fu Wu, Jaesik Yoon, Sungjin Ahn
  • Publisher: NeurIPS 2021 (Deep RL Workshop)
  • Key: Dreamer, World Model
  • ExpEnv: Hidden Order Discovery, DMC, Atari

ICML 2021

• Reinforcement learning as one big sequence modeling problem
  • Michael Janner, Qiyang Li, Sergey Levine
  • Publisher: ICML workshop
  • Key: Conditional sequence modeling, Discretization
  • Code: official
  • ExpEnv: MuJoco, D4RL

Contributing

Our purpose is to make this repo even better. If you are interested in contributing, please refer to HERE for instructions in contribution.

License

Awesome Decision Transformer is released under the Apache 2.0 license.