This project is by David Bachmann (bacdavid). It is not published or affiliated.

Reinforcement learning agent based on Hindsight Experience Replay (Andrychowicz et al., https://arxiv.org/abs/1707.01495) and Deep Deterministic Policy Gradients (Lilicrap et al., https://arxiv.org/abs/1509.02971). The agent was used in the Perception and Learning for Robotics project at ETH Zürich with the title Let a Robot Solve Tiling Puzzles. Please note that this is a strapt version which excludes the PyBullet environment, a toy example based on OpenAI Gym (https://gym.openai.com/) has been added instead.