개요
강화학습 공부를 시작하기에 앞서
앞서 공부한 사람들의 의견을 듣고,
가장 효율적인 방법을 찾고 싶었다.
내가 searching하면서 찾은 자료들의 중복된 결과는 제거하고
이미 많이 공개된 RL curriculum을 요약한 자료라고 생각하면 될 것 같다.
Media / Lecture
Book
- Sutton & Barto's Introduction to Reinforcement Learning
- 3분 딥러닝
- 케라스로 구현하는 딥러닝과 강화학습
- 파이썬과 케라스로 배우는 강화학습
- 이웅원 gitbook
- Spinning up
Paper
- DQN
- double DQN
- dueling DQN
- Policy Gradient Methods
- Actor-Critic Methods
- DDPG
- TD3
- SAC
- TRPO
- PPO
- ACER
- Prioritized Experience Replay
- Deep Recurrent Q-Learning for Partially Observable MDPs
- Noisy Networks for Exploration
- A Distributional Perspective on Reinforcement Learning
- QR-DQN
- IQN
- GA3C
- Hierarchical Reinforcement Learning
- Deep Determinstic Policy Gradient
- Addressing Function Approximation Error in Actor-Critic Methods
RL Frameworks & Reference Implementations
- OpenAI Baselines
- Stable Baselines
- SLM Lab: Modular Deep Reinforcement Learning framework in PyTorch
- Reinforcement Learning Coach
- Framework for training, evaluating, and deploying robust trading agents
- RLBench: Robot Learning Benchmark
- Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning
- Denny Britz's implementations of basic RL algorithms.
- Garage: A toolkit for evaluating meta-RL and multi-task agents
RL Environments
- OpenAI Gym
- Unity ML-Agents
- PyBullet Gymperium: Open-source implementations of OpenAI Gym MuJoCo environments
- osim-rl: Reinforcement learning with musculoskeletal models in OpenSim
- Project Malmö: platform for Artificial Intelligence experimentation and research built on top of Minecraft
- Multi-agent particle environment used in the paper Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
- Environment for reinforcement-learning algorithmic trading models
- Framework for training, evaluating, and deploying robust trading agents
- Doom-based AI Research Platform for Reinforcement Learning from Raw Visual Information
- OpenSpiel: Collection of RL environments and algorithms
- ns3-gym: The Playground for Reinforcement Learning
- Reco-gym: A Reinforcement Learning Environment for the problem of Product Recommendation in Online Advertising
- RLBench: Robot Learning Benchmark
- OpenAI Gym trading environment
- Sairen: OpenAI Gym Reinforcement Learning Environment for the Stock Market
- Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learning
- Voyage Deepdrive: self-driving simulator
- CARLA: Open-source simulator for autonomous driving research
- Habitat: A simulation platform for research in embodied AI
- MineRL: Towards AI in Minecraft
- D4RL: Datasets for Deep Data-Driven Reinforcement Learning
- Multi-Agent-Learning-Environments
- PettingZoo: Multi-agent RL environments
- MiniGrid: Minimalistic Gridworld Environment
- endtoendAI: compilation of RL environments
- List of environments and competitions for RL and AI training
- Official OpenAI environments, and third party environments
- Awesome RL environment list
Training
- DQN 슈퍼마리오 적용
- DDPG 자동차 시뮬레이션, skeleton 걷는 것에 적용
Data Scientist, Data Analyst
Thanks for posting