NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:1135
Title:Transferable Normalization: Towards Improving Transferability of Deep Neural Networks


		
The paper proposes a novel normalization strategies in the line of AdaBN and AutoDIAL paper. The method is simple, easy to use and shows better results than other normalization strategies. The paper is well written, the experimental evaluation is large and solid with the help of the complementary information provided in the rebuttal. Authors should precise better f distances are used in the probabilities of softmax and Gaussian