Low-Shot Learning with Large-Scale Diffusion

By Matthijs Douze, Arthur Szlam, Bharath Hariharan, Herv'e J'egou. 2017

121 citations recorded by the source corpus.

This paper considers the problem of inferring image labels from images when only a few annotated examples are available at training time. This setup is often referred to as low-shot learning, where a standard approach is to retrain the last few layers of a convolutional neural network learned on separate classes for which training examples are abundant. We consider a semi-supervised setting based on a large collection of images to support label propagation. This is possible by leveraging the recent advances on large-scale similarity graph construction. We show that despite its conceptual simplicity, scaling label propagation up to hundred millions of images leads to state of the art accuracy in the low-shot learning regime.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Ocean Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Ocean Engineering

Citations in the corpus

Memory-Efficient Incremental Learning Through Feature Adaptation (2020, methodology): “These methods can be semi-supervised [24,30,49], self-supervised [5,17,39] or perform few-shot learning [12,13,48].”
Few-Shot Object Detection via Feature Reweighting (2018, background): “[9] assume abundant unlabeled images and adopts label propagation in a semi-supervised setting.”
Generalizing from a Few Examples (2019, background): “us gestures of different people but no clear break between gestures. Then the final gesture classifier is build using these selected samples. Label propagation is used to label Dunlabeled directly in [27]. 3.2.2 Similar Data Set. This strategy augments Dtrain by aggregating samples pairs from other ”
Generalized Many-Way Few-Shot Video Classification (2020, background): “Therefore, the k training examples are often sampled several times and accuracy results are averaged [15,6].”
Label Propagation for Deep Semi-Supervised Learning (2019, background): “[7] perform label propagation on a large image dataset with CNN descriptors for few shot learning.”
Variational Few-Shot Learning (2019, background): “Many prior works [25, 45, 7, 6] stress on learning a well-organized matching mechanism.”
Large-Scale Few-Shot Learning: Knowledge Transfer With Class Hierarchy (2019, methodology): “The more recent works [7, 2, 20, 30] start to pay more attention to large scale FSL.”
Billion-scale semi-supervised learning for image classification (2019, background): “[11] improve classiﬁcation when the number of labelled data for the new classes is very low (1–5), but the gain disappears when more labelled data is available.”
MetaSearch: Incremental Product Search via Deep Meta-Learning (2020, background): “Douze et al. [10] exploited a large number of unlabeled data in combination with tag propagation for few-shot learning in a semisupervised setting.”
Boosting Few-Shot Learning With Adaptive Margin Loss (2020, methodology): “Then, the margin is integrated into the classiﬁcation loss to make similar classes more separable in the embedding space, which beneﬁts FSL. lem [6, 13, 25, 27].”
A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities (2022, background): “Exploit the Unknown Gradually [63] CVPR’18 Data Expansion MARS ✓ Large-scale Difusion [75] CVPR’18 Data Expansion ImageNet ✓ ✓”