Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
capsule_theory [2018/11/19 23:31]
admin
capsule_theory [2018/12/29 12:56]
admin
Line 66: Line 66:
 We introduce, (1) a novel routing weight initialization technique, (2) an improved CapsNet design that exploits semantic relationships between the primary capsule activations using a densely connected Conditional Random Field and (3) a Cholesky transformation based correlation module to learn a general priority scheme. Our proposed design allows CapsNet to scale better to more complex problems, such as the multi-label classification task, where semantically related categories co-exist with various interdependencies. ​ We introduce, (1) a novel routing weight initialization technique, (2) an improved CapsNet design that exploits semantic relationships between the primary capsule activations using a densely connected Conditional Random Field and (3) a Cholesky transformation based correlation module to learn a general priority scheme. Our proposed design allows CapsNet to scale better to more complex problems, such as the multi-label classification task, where semantically related categories co-exist with various interdependencies. ​
  
-https://​arxiv.org/​abs/​1811.06969v1+https://​arxiv.org/​abs/​1811.06969v1 ​DARCCC: Detecting Adversaries by Reconstruction from Class Conditional Capsules 
 + 
 +https://​arxiv.org/​abs/​1812.09707v1 Training Deep Capsule Networks 
 + 
 +To ensure that all active capsules form a parse tree, we introduce a new routing algorithm called dynamic deep routing. We show that this routing algorithm allows the training of deeper capsule networks and is also more robust to white box adversarial attacks than the original routing algorithm.