Vincent Gripon's Homepage

Research and Teaching Blog

Exploiting Unsupervised Inputs for Accurate Few-Shot Classification

Y. Hu, V. Gripon and S. Pateux, "Exploiting Unsupervised Inputs for Accurate Few-Shot Classification," in ArXiv Preprint: 2001.09849, 2020.

In few-shot classification, the aim is to learn models able to discriminate classes with only a small number of labelled examples. Most of the literature considers the problem of labelling a single unknown input at a time. Instead, it can be beneficial to consider a setting where a batch of unlabelled inputs are treated conjointly and non-independently. In this paper, we propose a method able to exploit three levels of information: a) feature extractors pretrained on generic datasets, b) few labelled examples of classes to discriminate and c) other available unlabelled inputs. If for a), we use state-of-the-art approaches, we introduce the use of simplified graph convolutions to perform b) and c) together. Our proposed model reaches state-of-the-art accuracy with a 6-11% increase compared to available alternatives on standard few-shot vision classification datasets.


Bibtex
@inproceedings{HuGriPat2020,
  author = {Yuqing Hu and Vincent Gripon and Stéphane
Pateux},
  title = {Exploiting Unsupervised Inputs for Accurate
Few-Shot Classification},
  booktitle = {ArXiv Preprint: 2001.09849},
  year = {2020},
}




You are the 984150th visitor

Vincent Gripon's Homepage