Capsule Network Lab

Watch dynamic routing by agreement: the predictions, the coupling coefficients, the squash function, and how the lengths of the higher capsules change with each iteration.

Loading the lab… If it does not appear, make sure JavaScript is enabled in your browser.

The predictions û are generated randomly with a fixed seed, and the algorithm matches the dynamic routing in the paper by Sabour et al. (2017).

Architectures related to this lab

Intermediate2017

Capsule Networks

Capsule Networks (CapsNet)

Networks that represent entities as capsules (vectors) whose length expresses the probability that the entity is present and whose direction expresses its properties (po…

Advanced Neural Architectures

Intermediate2017

Hybrid Capsule Networks

Hybrid Capsule Networks (CNN-CapsNet)

Designs that use a strong convolutional backbone (such as ResNet) to extract features, then a capsule layer with dynamic routing on top to represent part–whole relations…

Hybrid Architectures