Sim-to-real — main techniques.
hardAnswer
- (1) Domain randomization: randomize physics params (mass, friction, latency) in sim → policy robust to real variation.
- (2) Domain adaptation: fine-tune on real data.
- (3) System identification: fit sim params to match real data.
- (4) Adversarial DR / meta-learning: adaptive randomization.
- (5) Photorealistic sim + vision randomization.
- Standard in robotics locomotion (Tan et al. 2018, Rudin et al. 2022) + manipulation (OpenAI's Rubik's cube).
Check yourself — multiple choice
- Random
- Domain randomization (physics + visuals) + domain adaptation on real data + system ID + meta-learning + photorealistic sim; standard in robotics locomotion + manipulation
- Same as fine-tune
- Not real
Sim2real: DR + DA + system ID + meta-learning + photorealism.
#deep-rl#applications
Practise Reinforcement Learning
214 interview questions in this topic.