论文标题
介入的几次学习
Interventional Few-Shot Learning
论文作者
论文摘要
我们揭示了普遍的几次学习方法（FSL）方法的不断忽视：预训练的知识确实是限制性能的混杂因素。这一发现源于我们的因果假设：用于预先训练的知识，样本特征和标签之间因果关系的结构性因果模型（SCM）。多亏了这一点，我们提出了一种新颖的FSL范式：介入的几次学习（IFSL）。具体而言，我们基于后门调整开发了三种有效的IFSL算法实现，这实质上是对许多射击学习的SCM的因果干预：在因果观点中FSL的上限。值得注意的是，IFSL对现有的基于微调和基于元学习的FSL方法的贡献是正交的，因此IFSL可以改善所有这些方法，从而在\ textit {mini} imagenet，\ textit imageNet，\ textIt {textiT {tiere {tiere {tiereed} Imagenet和Cross-fimainet和Cross-fimain和Cub上获得了新的1-/5-Shot Hot-shot-thot-the-the-theart。代码在https://github.com/yue-zhongqi/ifsl上发布。
We uncover an ever-overlooked deficiency in the prevailing Few-Shot Learning (FSL) methods: the pre-trained knowledge is indeed a confounder that limits the performance. This finding is rooted from our causal assumption: a Structural Causal Model (SCM) for the causalities among the pre-trained knowledge, sample features, and labels. Thanks to it, we propose a novel FSL paradigm: Interventional Few-Shot Learning (IFSL). Specifically, we develop three effective IFSL algorithmic implementations based on the backdoor adjustment, which is essentially a causal intervention towards the SCM of many-shot learning: the upper-bound of FSL in a causal view. It is worth noting that the contribution of IFSL is orthogonal to existing fine-tuning and meta-learning based FSL methods, hence IFSL can improve all of them, achieving a new 1-/5-shot state-of-the-art on \textit{mini}ImageNet, \textit{tiered}ImageNet, and cross-domain CUB. Code is released at https://github.com/yue-zhongqi/ifsl.
