Hybrid Simulator Identification for Domain Adaptation via Adversarial Reinforcement Learning

Reinforcement Learning • Transfer RL

21 Nov 2022

Comments

Transfer RL과 관련하여 공부하고 있다. GARAT, MPGE는 action transfer policy를 이용하여 transfer한다면, SimGAN은 simulator 자체를 학습하여 transfer 한다.

---

Reference

1. SimGAN: Hybrid Simulator Identification for Domain Adaptation via Adversarial Reinforcement Learning (paper)
2. SimGAN (youtube)
3. SimGAN (official blog post)