4 ms·
Yeah, humans don’t learn from still pictures, we see things moving through time and space and rotating. This is likely providing us with essentially a lot of “s
by jtsiskin 6y ago
Yeah, humans don’t learn from still pictures, we see things moving through time and space and rotating. This is likely providing us with essentially a lot of “self supervised” data.
For instance https://www.ted.com/talks/pawan_sinha_how_brains_learn_to_see/transcript https://www.ted.com/talks/pawan_sinha_how_brains_learn_to_se... starting at 8:11
- ilaksh 6y agoGreat video. It's interesting to see how far semantic segmentation has progressed since 2009.