Latent Weights Generating for Few Shot Learning Using Information Theory

Abstract

Few shot image classification aims at learning a classifier from limited labeled data. Generating the classification weights has been applied in many metalearning approaches for few shot image classification due to its simplicity and effectiveness. However, fixed classification weights for different query samples within one task might be sub-optimal, due to the few shot challenge, and it is difficult to generate the exact and universal classification weights for all the diverse query samples from very few training samples. In this work, we introduce latent weights generating using information theory (LWGIT) for few shot learning which addresses current issues by generating different classification weights for different query samples by letting each of query samples attends to the whole support set. The experiment results demonstrate the effectiveness of LWGIT, thereby contributing to exceed the performances of the existing state-of-the-art models.