https://github.com/skylark0924/reinforcement-learning-in-robotics
This is a private learning repository for reinforcement learning techniques used in robotics.
https://github.com/skylark0924/reinforcement-learning-in-robotics
Last synced: 7 months ago
JSON representation
This is a private learning repository for reinforcement learning techniques used in robotics.
- Host: GitHub
- URL: https://github.com/skylark0924/reinforcement-learning-in-robotics
- Owner: Skylark0924
- License: mit
- Created: 2019-12-08T11:54:37.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2023-08-25T04:55:02.000Z (about 2 years ago)
- Last Synced: 2025-01-28T14:47:35.791Z (8 months ago)
- Language: HTML
- Size: 94.9 MB
- Stars: 398
- Watchers: 7
- Forks: 56
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Funding: .github/FUNDING.yml
- License: LICENSE
Awesome Lists containing this project
README
#! https://zhuanlan.zhihu.com/p/143392167
# **如需转发，烦请邮件告知 [junjialiu@sjtu.edu.cn](junjialiu@sjtu.edu.cn)**
# Reinforcement-Learning-in-Robotics Content 专栏目录
This is a private learning repository about **R**einforcement learning techniques, **R**easoning, and **R**epresentation learning used in **R**obotics, founded for **Real intelligence**.
## Reinforcement Learning Foundation
1. **神经网络基础**：反向传播推导与卷积公式 [[Zhihu](https://zhuanlan.zhihu.com/p/114370969)]
2. **强化学习基础 Ⅰ**：马尔可夫与值函数 [[Zhihu](https://zhuanlan.zhihu.com/p/114377860)]
3. **强化学习基础 Ⅱ**：动态规划，蒙特卡洛，时序差分 [[Zhihu](https://zhuanlan.zhihu.com/p/114482584)]
4. **强化学习基础 Ⅲ**：on-policy, off-policy & Model-based, Model-free & Rollout [[Zhihu](https://zhuanlan.zhihu.com/p/115629505)]
5. **强化学习基础 Ⅳ**：State-of-the-art 强化学习经典算法汇总 [[Zhihu](https://zhuanlan.zhihu.com/p/137208923)]
6. **强化学习基础 Ⅴ**：Q learning 原理与实战 [[Zhihu](https://zhuanlan.zhihu.com/p/141267943)]
7. **强化学习基础 Ⅵ**：DQN 原理与实战 [[Zhihu](https://zhuanlan.zhihu.com/p/141268549)]
8. **强化学习基础 Ⅶ**：Double DQN & Dueling DQN 原理与实战 [[Zhihu](https://zhuanlan.zhihu.com/p/141268851)]
9. **强化学习基础 Ⅷ**：Vanilla Policy Gradient 策略梯度原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/141269134)]
10. **强化学习基础 Ⅸ**：一文读懂 TRPO 原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/141269503)]
11. **强化学习基础 Ⅹ**：一文读懂两种 PPO 原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/141269918)]
12. **强化学习基础 Ⅺ**： Actor-Critic & A2C 原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/145168493)]
13. **强化学习基础 Ⅻ**：DDPG 原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/145181679)]
14. **强化学习基础 XIII**：Twin Delayed DDPG TD3原理与实现 [[Zhihu](https://zhuanlan.zhihu.com/p/145621630)]
## Model-based RL
1. **Model-Based RL Ⅰ**：Dyna, MVE & STEVE [[Zhihu](https://zhuanlan.zhihu.com/p/102197348)]
2. **Model-Based RL Ⅱ**：MBPO原理解读 [[Zhihu](https://zhuanlan.zhihu.com/p/105645139)]
3. **Model-Based RL Ⅲ**：从源码读懂PILCO [[Zhihu](https://zhuanlan.zhihu.com/p/138337983)]
## Probabilistic in Robotics
1. **PR 序**：机器人学的概率方法学习路径 [[Zhihu](https://zhuanlan.zhihu.com/p/150563142)]
2. **PR Ⅰ**：最大似然估计MLE与最大后验概率估计MAP [[Zhihu](https://zhuanlan.zhihu.com/p/138608823)]
3. **PR Ⅱ**：贝叶斯估计/推断及其与MAP的区别 [[Zhihu](https://zhuanlan.zhihu.com/p/139480748)]
4. **PR Ⅲ**：从高斯分布到高斯过程、高斯过程回归、贝叶斯优化 [[Zhihu](https://zhuanlan.zhihu.com/p/139478368)]
5. **PR Ⅳ**：贝叶斯神经网络 Bayesian Neural Network [[Zhihu](https://zhuanlan.zhihu.com/p/139523520)]
6. **PR Ⅴ**：熵、KL散度、交叉熵、JS散度及python实现 [[Zhihu](https://zhuanlan.zhihu.com/p/143105854)]
7. **PR Ⅵ**：多元连续高斯分布的KL散度及python实现 [[Zhihu](https://zhuanlan.zhihu.com/p/143124676)]
8. **PR Sampling Ⅰ**：蒙特卡洛采样、重要性采样及python实现 [[Zhihu](https://zhuanlan.zhihu.com/p/150693309)]
9. **PR Sampling Ⅱ**：马尔可夫链蒙特卡洛 MCMC及python实现 [[Zhihu](https://zhuanlan.zhihu.com/p/150742395)]
10. **PR Sampling Ⅲ**：M-H and Gibbs 采样 [[Zhihu](https://zhuanlan.zhihu.com/p/150946559)]
11. **PR Structured Ⅰ**：Graph Neural Network: An Introduction Ⅰ [[Zhihu](https://zhuanlan.zhihu.com/p/158984343)]
12. **PR Structured Ⅱ**：Structured Probabilistic Model 结构化概率模型 [[Zhihu](https://zhuanlan.zhihu.com/p/161703636)]
13. **PR Structured Ⅲ**：马尔可夫、隐马尔可夫 HMM 、条件随机场 CRF 全解析及其python实现 [[Zhihu](https://zhuanlan.zhihu.com/p/259660645)]
14. **PR Structured Ⅳ**：General / Graph Conditional Random Field (CRF) 及其 python 实现 [[Zhihu](https://zhuanlan.zhihu.com/p/259883878)]
15. **PR Structured Ⅴ**：GraphRNN——将图生成问题转化为序列生成 [[Zhihu](https://zhuanlan.zhihu.com/p/276873641)]
16. **PR Reasoning 序**：Reasoning Robotics 推理机器人学习路线与资源汇总 [[Zhihu](https://zhuanlan.zhihu.com/p/262568794)]
17. **PR Reasoning Ⅰ**：Bandit问题与 UCB / UCT / AlphaGo [[Zhihu](https://zhuanlan.zhihu.com/p/218398647)]
18. **PR Reasoning Ⅱ**：Relational Inductive bias 关系归纳偏置及其在深度学习中的应用 [[Zhihu](https://zhuanlan.zhihu.com/p/261081574)]
19. **PR Reasoning Ⅲ**：基于图表征的关系推理框架 —— Graph Network [[Zhihu](https://zhuanlan.zhihu.com/p/261127145)]
20. **PR Reasoning Ⅳ**：数理逻辑（命题逻辑、谓词逻辑）知识整理 [[Zhihu](https://zhuanlan.zhihu.com/p/262984951)]
21. **PR Memory Ⅰ**：Memory systems 2018 – towards a new paradigm **【重磅综述】记忆系统——神经科学的启示** [[Zhihu](https://zhuanlan.zhihu.com/p/166692908)]
22. **PR Perspective Ⅰ**：Embodied AI 的新浪潮 —— new generation of AI [[Zhihu](https://zhuanlan.zhihu.com/p/260562672)]
23. **PR Perspective Ⅱ**：2021/08/03 近期 Robot Learning 领域大事件及思考 [[Zhihu](https://zhuanlan.zhihu.com/p/395562430)]
24. **PR Efficient Ⅰ**：机器人中的数据高效强化学习 [[Zhihu](https://zhuanlan.zhihu.com/p/358668613)]
25. **PR Efficient Ⅱ**：Bayesian Transfer RL with prior knowledge [[Zhihu](https://zhuanlan.zhihu.com/p/359620737)]
26. **PR Efficient Ⅲ**：像训练狗狗一样高效地训练机器人 [[Zhihu](https://zhuanlan.zhihu.com/p/359776893)]
27. **PR Efficient Ⅳ**：五分钟内让四足机器人自主学会行走 [[Zhihu](https://zhuanlan.zhihu.com/p/360314680)]
28. **PR Efficient Ⅴ**：自预测表征，让RL agent高效地理解世界 [[Zhihu](https://zhuanlan.zhihu.com/p/360526111)]
## Meta-Learning
1. **Meta-Learning:** An Introduction Ⅰ [[Zhihu](https://zhuanlan.zhihu.com/p/99730942)]
2. **Meta-Learning:** An Introduction Ⅱ [[Zhihu](https://zhuanlan.zhihu.com/p/100035717)]
3. **Meta-Learning:** An Introduction Ⅲ [[Zhihu](https://zhuanlan.zhihu.com/p/100266389)]
## Imitation Learning
1. **Imitation Learning Ⅰ：模仿学习** (Imitation Learning) 入门指南 [[Zhihu](https://zhuanlan.zhihu.com/p/140348314)]
2. **Imitation Learning Ⅱ**：DAgger透彻理论分析 [[Zhihu](https://zhuanlan.zhihu.com/p/140939491)]
3. **Imitation Learning Ⅲ**：EnsembleDAgger 一种贝叶斯DAgger [[Zhihu](https://zhuanlan.zhihu.com/p/140952343)]
## RL from Demonstrations
1. **RLfD Ⅰ**：Deep Q-learning from Demonstrations 解读 [[Zhihu](https://zhuanlan.zhihu.com/p/142779768)]
2. **RLfD Ⅱ**：Reinforcement Learning from Imperfect Demonstrations under Soft Expert Guidance [[Zhihu](https://zhuanlan.zhihu.com/p/143282816)]
## Multi-agent Reinforcement Learning
1. **MARL Ⅰ**：A Selective Overview of Theories and Algorithms **【重磅综述】 多智能体强化学习算法理论研究** [[Zhihu](https://zhuanlan.zhihu.com/p/220581474)]
## Paper Reading
**Active Visual Navigation**
1. Reading：**利用物体关系的目标驱动视觉导航** [[Zhihu](https://zhuanlan.zhihu.com/p/153404395)]
2. Reading：**Learning to learn how to learn-Meta自适应视觉导航** [[Zhihu](https://zhuanlan.zhihu.com/p/154184867)]
3. Reading：**Bayesian Relational Memory 在视觉导航中的应用** [[Zhihu](https://zhuanlan.zhihu.com/p/154290529)]
4. Reading：**Attention+3D空间关系图在视觉导航中的应用** [[Zhihu](https://zhuanlan.zhihu.com/p/156787516)]
5. Reading：**机器人导航的半参数化拓扑记忆结构** [[Zhihu](https://zhuanlan.zhihu.com/p/157227332)]
6. Reading：**将Transformer应用到机器人视觉导航中** [[Zhihu](https://zhuanlan.zhihu.com/p/157316200)]
**RL robotics in the physical world with micro-data / data-efficiency**
1. **【重磅综述】如何在少量尝试下学习机器人强化学习控制** [[Zhihu](https://zhuanlan.zhihu.com/p/144544347)]
**Others**
1. End-to-End Robotic Reinforcement Learning without Reward Engineering: [[Medium](https://medium.com/@skylark0924/notes-of-end-to-end-robotic-reinforcement-learning-without-reward-engineering-a6ffcc5c47f3)] [[Zhihu](https://zhuanlan.zhihu.com/p/96839443)]
2. Overcoming Exploration in RL with Demonstrations: [[Medium](https://medium.com/@skylark0924/notes-of-overcoming-exploration-in-reinforcement-learning-with-demonstrations-52dac4e84c58)] [[Zhihu](https://zhuanlan.zhihu.com/p/96841783)]
3. The Predictron: End-To-End Learning and Planning: [[Zhihu](https://zhuanlan.zhihu.com/p/96917057)]
4. **IROS2019 Paper速读(一)** [[Zhihu](https://zhuanlan.zhihu.com/p/97891687)]
5. **IROS2019 Paper速读(二)** [[Zhihu](https://zhuanlan.zhihu.com/p/98365711)]
6. **IROS2019 Paper速读(三)** [[Zhihu](https://zhuanlan.zhihu.com/p/98712344)]
7. **IROS2019 Paper速读(四)** [[Zhihu](https://zhuanlan.zhihu.com/p/98762958)]
## Simulator
1. [MuJoCo自定义机器人建模指南](https://zhuanlan.zhihu.com/p/143983506)
2. [Sim2real in Robotics: An Introduction]()
## Tools
1. **Tools 1**：如何用 PyQt5 和 Qt Designer 在 Pycharm 中愉快地开发软件 [[Zhihu](https://zhuanlan.zhihu.com/p/259564109)]
2. **Tools 2**：Arxiv 论文提交流程——看这篇就够了 [[Zhihu](https://zhuanlan.zhihu.com/p/109405192)]
3. **Tools 3**：Python socket 服务器与客户端双向通信（服务器NAT，文件传输） [[Zhihu](https://zhuanlan.zhihu.com/p/263630359)]
4. **Tools 4**：Python三行转并行——真香！[[Zhihu](https://zhuanlan.zhihu.com/p/109405192)]
5. **Tools 5**：Python三行转并行后续——多进程全局变量 [[Zhihu](https://zhuanlan.zhihu.com/p/273508904)]
## 专栏关联Github代码库
[Reinforcement-Learning-in-Robotics](https://github.com/Skylark0924/Reinforcement-Learning-in-Robotics 'card')\
[Machine-Learning-is-ALL-You-Need](https://github.com/Skylark0924/Machine-Learning-is-ALL-You-Need 'card')
---
If you're interested in reinforcement learning, we encourage you to check out our latest library of reinforcement learning and imitation learning in (humanoid) robotics.
[](https://github.com/Skylark0924/Rofunc)
[](https://pypi.org/project/rofunc/)
[](https://github.com/Skylark0924/Rofunc/issues?q=is%3Aissue+is%3Aclosed)
[](https://github.com/Skylark0924/Rofunc/issues?q=is%3Aopen+is%3Aissue)
[](https://rofunc.readthedocs.io/en/latest/?badge=latest)
[](https://actions-badge.atrox.dev/Skylark0924/Rofunc/goto?ref=main)
> **Repository address: https://github.com/Skylark0924/Rofunc**