NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5567
Title:Large Scale Adversarial Representation Learning

A algorithmic scaling-up paper for representation learning with a better generator by combining BiGAN and BigGAN. Well-explored ablation studies showcasing the efficacy of the approach. While the reviewers have raised concerns of compute time, I think this paper needs sharing broadly. I would suggest authors incorporating suggestions from the reviews to strengthen the paper; this includes visualizations of iterative generation and additional results.