元学习方向 optimization based meta learning 之 MAML论文详细解读

元学习系列文章

1. optimization based meta-learning
   1. 《Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks》 论文翻译笔记
   2. 元学习方向 optimization based meta learning 之 MAML论文详细解读：本篇博客
   3. MAML 源代码解释说明 (一)
   4. MAML 源代码解释说明 (二)
   5. 元学习之《On First-Order Meta-Learning Algorithms》论文详细解读
   6. 元学习之《OPTIMIZATION AS A MODEL FOR FEW-SHOT LEARNING》论文详细解读
2. metric based meta-learning: 待更新…
3. model based meta-learning: 待更新…

目前 meta-learning 的研究一共有三个方向：

1. optimization based
2. metric based
3. model based

本篇要讲的论文是第一个方向，optimization based 中的开篇之作，《Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks》 简称 MAML