Few-Shot Image Recognition by Predicting Parameters from Activations

TLDR

The authors aim to address few‑shot learning for many categories by proposing a method that adapts a pre‑trained network to novel classes by predicting class‑specific parameters directly from activations. Their approach predicts the parameters of a neural network for each new class from its activations, requiring no additional training and enabling adaptation in a single forward pass. The method achieves state‑of‑the‑art accuracy on ImageNet novel categories and outperforms prior work on MiniImageNet, while maintaining comparable performance on large‑scale categories and enabling fast inference.

Abstract

In this paper, we are interested in the few-shot learning problem. In particular, we focus on a challenging scenario where the number of categories is large and the number of examples per novel category is very limited, e.g. 1, 2, or 3. Motivated by the close relationship between the parameters and the activations in a neural network associated with the same category, we propose a novel method that can adapt a pre-trained neural network to novel categories by directly predicting the parameters from the activations. Zero training is required in adaptation to novel categories, and fast inference is realized by a single forward pass. We evaluate our method by doing few-shot image recognition on the ImageNet dataset, which achieves the state-of-the-art classification accuracy on novel categories by a significant margin while keeping comparable performance on the large-scale categories. We also test our method on the MiniImageNet dataset and it strongly outperforms the previous state-of-the-art methods.