追随梦想，实现梦想

最新必读的8篇「小样本学习（few-shot learning）」2020顶会论文和代码

Few-Shot Papers

Generalizing from a Few Examples: A Survey on Few-Shot LearningGeneralizing from a Few Examples: A Survey on Few-Shot Learning

To tackle this problem, Few-Shot Learning (FSL) is proposed. It can rapidly generalize from new tasks of limited supervised experience using prior knowledge.

小样本学习年度进展|VALSE2018小样本学习，路在何方？从 CVPR 2019 一览小样本学习研究进展

PPT链接地址

阅读原文点我 （里面是一个59页的PPT）

小样本学习数据集 （阅读原文点击左侧）Few-shot regression(toy problems):Few-shot classification(image classification):Omniglot (Lake et al., 2015, 2019)Mini-ImageNet (Vinyals et al., 2016,Ravi et al., 2017)...

Generalizing from a Few Examples: A Survey on Few-Shot Learning下载A Closer Look at Few-shot Classification下载

小样本学习综述.….. 

Prototypical networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class.  

阅读原文点我AbstractIn this work, we employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories. Our framework learns...

当小样本遇上机器学习 fewshot learning  

最近看论文的时候，看到一个词 meta learning--Learning to Learn ，有点不太明白，所以就找了下面这篇文章。（仅供学习交流使用）Learning to LearnChelsea Finn    Jul 18, 2017A key aspect of intelligence is versatility – the capability of doing m...

CVPR 2018 最前沿：让神经网络学习比较来实现少样本学习下文转载自雷锋网 （仅供学校交流使用）雷锋网(公众号：雷锋网) AI 科技评论按：本文作者 Flood Sung，他为雷锋网 AI 科技评论撰写了基于 CVPR 录用论文《Learning to Compare: Relation Network for Few-Shot Learning》的独家解读稿件，未经许可不得转载。...

知乎大佬总结的一篇文章，内容还是比较全的。小样本学习（Few-shot Learning）综述（知乎）小样本学习综述（微信）两者是同一个作者写的文章。...

Few-Shot Learning in CVPR 2019