Revisiting Few-Shot Learning from a Causal Perspective

Few-shot learning with N-way K-shot scheme is an open challenge in machine learning. Many approaches have been proposed to tackle this problem, e.g., the Matching Networks and CLIP-Adapter. Despite that these approaches have shown significant progress, the mechanism of why these methods succeed has not been well explored. In this paper, we interpret these few-shot learning methods via causal mechanism. We show that the existing approaches can be viewed as specific forms of front-door adjustment, which is to remove the effects of confounders. Based on this, we introduce a general causal method for few-shot learning, which considers not only the relationship between examples but also the diversity of representations. Experimental results demonstrate the superiority of our proposed method in few-shot classification on various benchmark datasets. Code is available in the supplementary material.