In Hintons talk "What's wrong about convolutional nets" (Late 2016 or early 2015, I guess) he talks about capsules to make a modular CNN.
Is there any publicly available implementation or papers about this idea?
Data Science Stack Exchange is a question and answer site for Data science professionals, Machine Learning specialists, and those interested in learning more about the field. It only takes a minute to sign up.
Sign up to join this communityIn Hintons talk "What's wrong about convolutional nets" (Late 2016 or early 2015, I guess) he talks about capsules to make a modular CNN.
Is there any publicly available implementation or papers about this idea?
Following the publication of Dynamic Routing of Capsules (https://arxiv.org/abs/1710.09829), there have been open source implementations of this idea released. See https://github.com/XifengGuo/CapsNet-Keras and https://github.com/soskek/dynamic_routing_between_capsules for two such examples.
Thought I'd add a selection of what's of some of the 'best' stuff out there to suit different learning styles :)
Literature - Original paper by Sara Sabour, Nicholas Frosst, Geoffrey E Hinton
Blog post - Understanding Hinton’s Capsule Networks. Part I: Intuition by Max Pechyonkin
Youtube video - Capsule Networks (CapsNets) – Tutorial by Aurélien Géron
Implementation - Keras implementation by Xifeng Guo
More reading - Github repository containing comprehensive list of CapsNet info