SIGGRAPH Asia 2018: Skills from Videos paper (main video)

Main video accompanying the SIGGRAPH Asia 2018 paper:
"SFV: Reinforcement Learning of Physical Skills from Videos".
Supplementary video: • SIGGRAPH Asia 2018: Sk...
Project page: xbpeng.github....
Blog: bair.berkeley....

[@eliadlevy1760]

Amazing. Reminds of Sci-Fi movies I had watched not too long ago and murmured "yeah right".

[@Firewalker124]

Oh man, I would really like to use this for data augmentation in another case. Any possibilities in that direction?

[@5ilver42]

how might one obtain this simulation program to put to use with their own 3d model and reference video?

[@jinchangfan9243]

great work!

[@christopherborchert6951]

Were @pigme123 and @GingerNinjaTrickster informed of the use of their videos and given credit? I'm sure if they're not aware of their participation, they would be psyched to see this.

[@jasonpeng2176]

Yup, we've contacted them about this already.

[@christopherborchert6951]

@@jasonpeng2176 Nice. Glad to hear it.

[@xc5838]

How long did this take to train? I read the last paper from you guys and it stated that it takes a or day or more to learn one pose but you guys also didn't release what kind of hardware you were using. I also read the limitation from this paper and there was no mention of training time.

[@ДжонШепард-ы8й]

Great!

[@fargiofarmani6736]

LMAO 3:15 is too funny

[@martinl6790]

Can/will this be used for game developing?

[@draken5379]

In the future, this could be used with VR, to capture the whole body without advanced sensors.

[@VinBhaskara_]

goddamn it! $*&^%*& awesome Berkeley!!

[@SVAFnemesis]

I better be retired when this thing come out