NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5999
Title:Gaussian-Based Pooling for Convolutional Neural Networks

Reviewers found this paper well explained and to contain interesting ideas for pooling layers in CNNs. The method was applied to realistic large-scale datasets. The consensus is the proposed method adds an interesting tool to the CNN training kit.