Dynamic Routing Between Capsules

10/26/2017
by   Sara Sabour, et al.
0

A capsule is a group of neurons whose activity vector represents the instantiation parameters of a specific type of entity such as an object or an object part. We use the length of the activity vector to represent the probability that the entity exists and its orientation to represent the instantiation parameters. Active capsules at one level make predictions, via transformation matrices, for the instantiation parameters of higher-level capsules. When multiple predictions agree, a higher level capsule becomes active. We show that a discrimininatively trained, multi-layer capsule system achieves state-of-the-art performance on MNIST and is considerably better than a convolutional net at recognizing highly overlapping digits. To achieve these results we use an iterative routing-by-agreement mechanism: A lower-level capsule prefers to send its output to higher level capsules whose activity vectors have a big scalar product with the prediction coming from the lower-level capsule.

READ FULL TEXT

page 5

page 6

page 7

research
12/23/2018

Training Deep Capsule Networks

The capsules of Capsule Networks are collections of neurons that represe...
research
05/27/2019

Capsule Routing via Variational Bayes

Capsule Networks are a recently proposed alternative for constructing Ne...
research
07/08/2020

Quaternion Capsule Networks

Capsules are grouping of neurons that allow to represent sophisticated i...
research
12/29/2018

Attention-Based Capsule Networks with Dynamic Routing for Relation Extraction

A capsule is a group of neurons, whose activity vector represents the in...
research
08/11/2018

Neural Network Encapsulation

A capsule is a collection of neurons which represents different variants...
research
10/01/2021

Unsupervised Motion Representation Learning with Capsule Autoencoders

We propose the Motion Capsule Autoencoder (MCAE), which addresses a key ...
research
03/15/2021

Pushing the Limits of Capsule Networks

Convolutional neural networks use pooling and other downscaling operatio...

Please sign up or login with your details

Forgot password? Click here to reset