Awesome Machine Learning for Robotics 
Machine Learning is becoming a common technique to address robotics tasks. This repository intends cover this usage from a broad point of view.
Papers
Learned Approximate Inference
- Auto-Encoding Variational Bayes, by Kingma et Al.
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models, by Rezende et Al.
- Variational Inference with Normalizing Flows, by Rezende et Al.
- Improved Variational Inference with Inverse Autoregressive Flows, By Kingma et Al.
- Gradient Estimation Using Stochastic Computation Graphs by Schulman et Al.
- Deep Variational Bayes Filters: Unsupervised Learning of State Space Models from Raw Data by Karl et Al.
- Deep Kalman Filters by Krishnan et Al.
- Variational Inference: Foundations and Modern Methods, byu Blei et Al.
Reinforcement Learning
Deep Reinforcement Learning
- Deep Q-Learning: Human-level control through deep reinforcement learning, by Mnih et Al.
- DDPG: Continuous control with Deep Reinforcement Learning, by Lillicrap et Al.
- Prioritized Experience Replay, by Schaul et Al.
- Auxiliary tasks: Reinforcement learning with unsupervised auxiliary tasks, by Jaderberg et Al.
- Emergence of Locomotion Behaviours in Rich Environments, by Heess et Al.
Reproducibility of Deep RL experiments
- Deep RL that matters, by Henderson et Al.
- Reproducibility of Benchmarked Deep Reinforcement Learning Tasks for Continuous Control, by Islam et Al.
Reinforcement Learning Theory
- REINFORCE: Simple statistical gradient-following algorithms for connectionist reinforcement learning, Williams et Al.
- Policy Gradient Theorem: Policy Gradient Methods for Reinforcement Learning with Function Approximation, Sutton et Al.
- Deterministic Policy Gradient Algorithms, Silver et Al.
Policy Search
- Policy search suvery: Reinforcement learning of motor skills with policy gradients, by Peters and Schaal.
- Guided policy search, by Levine et Al.
Meta-Learning
Goal Exploration Processes
- Intrinsically Motivated Multi-Task Reinforcement Learning, by Forestier and Oudeyer.
Curriculum learning
- Automated curriculum learning, by Graves et Al.
Multi-task learning
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks, by Finn and Al.
Implementations
- OpenAI Gym: A Python library providing many simulation environments.
- OpenAI baselines: Implementations of Deep Reinforcement Learning algorithms by experts.
- Explauto: A library to perform intrinsically motivated exploration.
- Guided Policy Search: Implementation of the Guided Policy Search algorithm.
- Keras-RL: A keras-compatible Deep Reinforcement Learning framework (DQN, SARSA, DDPG...).
- Deepmind DQN: Deepmind's implementation used for the Nature paper.
- Devsisters DQN: A nice DQN implementation.
Implementing RL algorithms
- Best pratices in implementing Deep RL algorithms, as part of a blog post.
- A note about gradient clipping, by Karpathy. Further explained in a blog post.
Robotic simulator
- MuJoCo: The reference. Closed-source
- OpenAI Roboschool: A Mujoco clone in bullet, open-source.
- Gazebo: A simulator used in the ROS suite.
- V-REP: A simulator used with the Poppy project.
Robotic platforms
- Poppy: An open-source 3D-printed robotic ecosystem (humanoid, torso...)
About
Authors: