Skip to content

Latest commit

 

History

History
10 lines (8 loc) · 1.08 KB

README.md

File metadata and controls

10 lines (8 loc) · 1.08 KB

RouteNet

Goal: Feedforward neural network that has hard gates such that features are passed from one layer to the next in a content-aware fashion, reducing the overall number of computations needed for inference. There is some relation to CapsuleNets--however, those networks use iterations to perform dynamic routing from a capsule in layer i to a single target capsule in layer i+1, whereas the desired network here has no recurrence/iteration, and routing that can send features in one layer to more than one destination in subsequent layers.

Notes

Similar ideas are finally starting to show up in the literature. E.g.,