3 ms·
Feels like an historical moment. DeepMimic is a great companion to this: simulated robots using self-play to "discover" the art of Kung Fu ;) https://www.youtu
by ArtWomb 8y ago
Feels like an historical moment. DeepMimic is a great companion to this: simulated robots using self-play to "discover" the art of Kung Fu ;)
https://www.youtube.com/watch?v=vppFvq2quQ0 https://www.youtube.com/watch?v=vppFvq2quQ0
The quote from Lee about creativity is haunting. If the current debate centers on using human reinforcement for robot meta-learning. Particularly using one-shot learning from a single video demonstration. It's possible the main competitor to emerge. Given advances in near future TPU Cloud performance and AI accelerators. An emergence of end-to-end transfer learning directly from simulation. With the unexpected result of novel strategies beyond the capacity of expert level humans to devise.
Ilya Sutskever at AI Frontiers 2018: Recent Advances in Deep Learning and AI from OpenAI
https://www.youtube.com/watch?v=ElyFDUab30A https://www.youtube.com/watch?v=ElyFDUab30A
- dmoy 8y agoSo what's up with throwing boxes at everything all the time? Is that just an easy way to put varying amounts of external complications in to see how it handles?
- machiaweliczny 8y agoChecking robustness to noise I think. These algorithms are meant for stochastic, not entirely predictable environments (so your action might not always work and you should be able to carry on task even after some action failure (eg. in case of walking - you could slip, wind blows harder or something hits you etc.))