Russ Salakhutdinov: Capsules may replace ConvNets one day.
1 replies, 304 likes
Nitish Srivastava: Happy to share our recent work, "Geometric Capsule Autoencoders for 3D Point Clouds." The main idea is that instead of finding agreement among parts of an object, we find agreement among different views of the object. https://arxiv.org/abs/1912.03310 https://t.co/FQ9DNVP98Y
4 replies, 269 likes
Russ Salakhutdinov: New work on Geometric Capsules: Learning to group 3D points into parts & parts into the whole object in unsupervised way. Each capsule represents a visual entity consisting of a pose & feature representing "where" & ''what'' it is.
w/t @nitishsr & @Hanlin https://t.co/WP5GBaxI8y
0 replies, 141 likes
Hanlin Goh: Check out our latest work on modeling the "what's" and "where's" of objects and parts with geometric capsule representations from 3D point clouds. http://arxiv.org/abs/1912.03310 (with @nitishsr & @rsalakhu)... I'm at #NeurIPS2019 all week if you want to chat about it!
0 replies, 28 likes
Hanlin Goh: Have a look at our recent work with @nitishsr and @rsalakhu on learning geometric capsule representations of objects from 3D point clouds - http://arxiv.org/abs/1912.03310 ... wanna chat about it? Look me up at @NeurIPSConf
1 replies, 15 likes
Antoine Choppin: Reminds me of @JeffCHawkins 's Thousand Brains Theory as he talked about in @lexfridman AI podcast https://youtu.be/-EVqrDlAqYo
1 replies, 4 likes
Josh Susskind: Capsules are all you need ;-)
0 replies, 1 likes
Found on Dec 11 2019 at https://arxiv.org/pdf/1912.03310.pdf