Capsule networks with dynamic routing are shown to be equivalent to a routing-weighted product of expert neurons, supporting a bottom-up unsupervised learning algorithm via alternating routing and contrastive divergence.
Dropout: a simple way to prevent neural networks from overfitting
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Training capsules as a routing-weighted product of expert neurons
Capsule networks with dynamic routing are shown to be equivalent to a routing-weighted product of expert neurons, supporting a bottom-up unsupervised learning algorithm via alternating routing and contrastive divergence.