Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning — mentioned by hcrisp

Sat, 10 Dec 2022 00:09:16 GMT

It's not new. A paper in 2021 [0] showed you can train a quadruped robot to walk in minutes using parallel simulation in GPU, and then deploy it on the physical robot. Being parallel it is faster, and more so if on GPU. But sim-to-reality transfer is still a concern, and the architecture doesn't help if there are sparse rewards. [0] https://arxiv.org/abs/2109.11978?context=cs.LG

Paper and discussion history. A mention is not an endorsement.

Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning · HackerPapers discussions

Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learning — mentioned by hcrisp