Extended Few-Shot Learning: Exploiting Existing Resources for Novel Tasks
This work addresses the problem of leveraging abundant auxiliary data for few-shot image classification, which is an incremental improvement for researchers in few-shot learning.
This paper introduces extended few-shot learning, a new problem setting where auxiliary datasets are available alongside scarce labeled examples for novel tasks. The authors propose a masking module that adjusts features of auxiliary data to be more similar to target classes, outperforming naive approaches by 4.68 and 6.03 percentage points.
In many practical few-shot learning problems, even though labeled examples are scarce, there are abundant auxiliary datasets that potentially contain useful information. We propose the problem of extended few-shot learning to study these scenarios. We then introduce a framework to address the challenges of efficiently selecting and effectively using auxiliary data in few-shot image classification. Given a large auxiliary dataset and a notion of semantic similarity among classes, we automatically select pseudo shots, which are labeled examples from other classes related to the target task. We show that naive approaches, such as (1) modeling these additional examples the same as the target task examples or (2) using them to learn features via transfer learning, only increase accuracy by a modest amount. Instead, we propose a masking module that adjusts the features of auxiliary data to be more similar to those of the target classes. We show that this masking module performs better than naively modeling the support examples and transfer learning by 4.68 and 6.03 percentage points, respectively.