Learning-a-Few-shot-Embedding-Model-with-Contrastive-Learning

This repo contains codes for Learning a Few-shot Embedding Model with Contrastive Learning(AAAI2021)

Abstract

Few-shot learning (FSL) aims to recognize target classes by adapting the prior knowledge learned from source classes. Such knowledge usually resides in a deep embedding model for a general matching purpose of the support and query image pairs. The objective of this paper is to repurpose the contrastive learning for such matching to learn a fewshot embedding model. We make the following contributions: (i) We investigate the contrastive learning with Noise Contrastive Estimation (NCE) in a supervised manner for training a few-shot embedding model; (ii) We propose a novel contrastive training scheme dubbed infoPatch, exploiting the patch-wise relationship to substantially improve the popular infoNCE. (iii) We show that the embedding learned by the proposed infoPatch is more effective. (iv) Our model is thoroughly evaluated on few-shot recognition task; and demonstrates state-of-the-art results on miniImageNet and appealing performance on tieredImageNet, FewshotCIFAR100 (FC-100).

Paper

AAAI 2021

Citation

@inproceedings{liu2021learning,
  title={Learning a Few-shot Embedding Model with Contrastive Learning},
  author={Liu, Chen and Fu, Yanwei and Xu, Chengming and Yang, Siqian and Li, Jilin and Wang, Chengjie and Zhang, Li},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={35},
  number={10},
  pages={8635--8643},
  year={2021}
}

Pretrain Weight

ResNet-12