Robots learning to walk by trial and error

I find it fascinating to watch robots learn to walk by reinforced deep learning.

2 hours to learn to walk

https://youtu.be/G62dyCV4oJg?si=_KYDb1tnRf6LpZhK

10 minutes of reinforcement learning 

https://youtu.be/-jqykPcQ5bs?si=P5pgNcXSs2ufC37M

https://youtu.be/n2gE7n11h1Y?si=kLJsq7j3kdSt-j7k

Dog-style robot reinforcement learning

https://youtu.be/xAXvfVTgqr0?si=icFlBB-Aqj3sWo6P

https://youtu.be/D5nk4mwmHXk?si=ynx0TW2KUKZGriaT

https://youtu.be/v90CWJ_HsnM?si=8xDZNzL_7jTaYWso

https://youtu.be/28gPyUmiFa4?si=9Cr7Dd-mSQMXQT4F

https://www.reddit.com/r/reinforcementlearning/s/DvxA7mkFsq