4 ms·
Self-supervised learning is very promising. But it is strange to me that so much computer vision research has no mention of the 3d structure of objects. That'
by ilaksh 6y ago
Self-supervised learning is very promising. But it is strange to me that so much computer vision research has no mention of the 3d structure of objects.
That's why I like things like this https://arxiv.org/abs/2102.12627 https://arxiv.org/abs/2102.12627 "How to represent part-whole hierarchies in a neural network" by Hinton.
- jtsiskin 6y agoYeah, 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.