NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:4983
Title:Finding the Needle in the Haystack with Convolutions: on the benefits of architectural bias

This paper examines the role of the architectural bias of CNNs, showing that the structural prior of locality and weight-sharing can improve performance of FCNs. The reviewers found the analysis novel and interesting and I agree with their consensus that this paper should be accepted.