继续讲Open World Vision Workshop。这里也介绍一下UIUC新锐教授Yuxiong的报告“learn to learn more with less”。 题目一看就明白两个点:few shot learning + meta learning。在一个开放和长尾数据发布的世界,数据搜集无法规模化,贵而且难。所以提出,learn how-to-learn on base(common)classes,然后转移到novel ...
ICLR '24. [paper] [code] Known for their impressive performance in generative modeling, diffusion models are attractive candidates for density-based anomaly detection. This paper investigates different variations of diffusion modeling for unsupervised and semi-supervised anomaly detection. In particular,...