對抗學習之Learning from Simulated and Unsupervised Images through Adversarial Training

今天講一篇cvpr2017的best paper,不是大家熟知的densely connect,是蘋果的Learning from Simulated and Unsupervised Images through Adversarial Training,用GAN來做對抗學習的。 Motivation:       搞數據一直是深度學習的老大難問題,也是工業界優勢於學術界的地方,需要花錢標很多數
相關文章
相關標籤/搜索