Iterative label cleaning for transductive and semi-supervised few-shot learning - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Iterative label cleaning for transductive and semi-supervised few-shot learning

Abstract

Few-shot learning amounts to learning representations and acquiring knowledge such that novel tasks may be solved with both supervision and data being limited. Improved performance is possible by transductive inference, where the entire test set is available concurrently, and semi-supervised learning, where more unlabeled data is available. Focusing on these two settings, we introduce a new algorithm that leverages the manifold structure of the labeled and unlabeled data distribution to predict pseudo-labels, while balancing over classes and using the loss value distribution of a limited-capacity classifier to select the cleanest labels, iteratively improving the quality of pseudo-labels. Our solution surpasses or matches the state of the art results on four benchmark datasets, namely miniImageNet, tieredImageNet, CUB and CIFAR-FS, while being robust over feature space pre-processing and the quantity of available data. The publicly available source code can be found in https: //github.com/MichalisLazarou/iLPC
Fichier principal
Vignette du fichier
C118.iccv21.few-ss.pdf (782.53 Ko) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03530170 , version 1 (17-01-2022)

Identifiers

  • HAL Id : hal-03530170 , version 1

Cite

Michalis Lazarou, Tania Stathaki, Yannis Avrithis. Iterative label cleaning for transductive and semi-supervised few-shot learning. ICCV 2021 - IEEE/CVF International Conference on Computer Vision, Oct 2021, Virtual, France. pp.1-10. ⟨hal-03530170⟩
28 View
50 Download

Share

Gmail Facebook X LinkedIn More