BriefGPT.xyz
Dec, 2018
元转移学习用于少样本学习
Meta-Transfer Learning for Few-Shot Learning
HTML
PDF
Qianru Sun, Yaoyao Liu, Tat-Seng Chua, Bernt Schiele
TL;DR
本篇论文提出了一种名为元转移学习的新型少样本学习方法,通过训练多个任务以及为每个任务学习深度神经网络权重的缩放和移位函数来实现迁移。作者使用提出的HT元批处理方案对MiniImagenet和Fewshot-CIFAR100这两个具有挑战性的少样本学习基准进行了实验,并将其与相关工作进行了广泛比较,结果验证了元转移学习方法的优越性和高准确性。
Abstract
meta-learning
has been proposed as a framework to address the challenging
few-shot learning
setting. The key idea is to leverage a large number of similar few-shot tasks in order to learn how to adapt a base-lear
→