Capsule Networks Insight
Tim and Keith discuss the value of capsule networks in learning rich semantic relationships and the trade-off between sample efficiency and speed due to the routing algorithm. Keith highlights the importance of circular harmonics and pose matrices in capsules, suggesting an extension to current models.In this clip
From this podcast
Machine Learning Street Talk (MLST)
Capsule Networks and Education Targets
Related Questions
How are convolutional neural networks (CNNs) used for image recognition in the episode Expectation Maximization, Gaussian Mixtures & Belief Propagation, OH MY! with Inmar Givoni - #101 and the clip Graphical Models Explained?
How does the architecture of neural networks relate to their ability to reason in the episode Ilya Sutskever: Deep Learning | Lex Fridman Podcast #94 and the clip Neural Networks and Reasoning?