TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning

By Sung Whan Yoon, Jun Seo, Jaekyun Moon. 2019

302 citations recorded by the source corpus.

Handling previously unseen tasks after given only a few training examples continues to be a tough challenge in machine learning. We propose TapNets, neural networks augmented with task-adaptive projection for improved few-shot learning. Here, employing a meta-learning strategy with episode-based training, a network and a set of per-class reference vectors are learned across widely varying tasks. At the same time, for every episode, features in the embedding space are linearly projected into a new space as a form of quick task-specific conditioning. The training loss is obtained based on a distance metric between the query and the reference vectors in the projection space. Excellent generalization results in this way. When tested on the Omniglot, miniImageNet and tieredImageNet datasets, we obtain state of the art classification accuracies under various few-shot scenarios.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

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

Citations in the corpus

Subspace Regularizers for Few-Shot Class Incremental Learning (2021, background): “Most effective approaches to few-shot learning rely on additional data for pre-training (Tian et al., 2020) or meta-learning (Vinyals et al., 2016; Finn et al., 2017; Snell et al., 2017; Yoon et al., 2019).”
Meta-Learning in Neural Networks: A Survey (2020, background): “The most common application of meta-learning is few-shot multi-class image recognition, where the inner and outer loss functions are typically the cross entropy over training and validation data respectively [14], [23], [75], [77], [78], [88], [90], [98], [99], [102], [105], [201], [202], [203], [20”
Learning Adaptive Classifiers Synthesis for Generalized Few-Shot Learning (2019, neutral): “20 TapNet [71] ResNet-12 61.”
XtarNet: Learning to Extract Task-Adaptive Representation for Incremental Few-Shot Learning (2020, methodology): “…upon the episodic training, successful few-shot learning algorithms attempt to build proper inductive bias over varying tasks while employing effective task-dependent adaptation for each task (Snell et al., 2017; Finn et al., 2017; Mishra et al., 2017; Oreshkin et al., 2018; Yoon et al., 2019).”
Document-Level Event Argument Extraction by Conditional Generation (2021, methodology): “Speciﬁcally, M satisﬁes the following equation: We refer to the TapNet (Yoon et al., 2019) paper for details and also provide a simpliﬁed derivation in Appendix A.”
A Comprehensive Survey of Few-shot Learning: Evolution, Applications, Challenges, and Opportunities (2022, unknown): “TapNet [240] PMLR’19 Meta ResNet-12 ✓ ✓ 61.”
Few-shot Slot Tagging with Collapsed Dependency Transfer and Label-enhanced Task-adaptive Projection Network (2020, methodology): “To remedy this, Yoon et al. (2019) propose TapNet that learns to project embedding to a space where words of different labels are well-separated.”
Online Structured Meta-learning (2020, unknown): “Under this setting, there are two representative lines of meta-learning algorithm, including optimization-based meta-learning [7–9, 13, 11, 17, 19, 26, 28, 31] and non-parametric meta-learning [12, 22, 24, 34, 35, 38, 43].”
On Episodes, Prototypical Networks, and Few-shot Learning (2020, methodology): “We chose to analyse Prototypical Networks not only for their simplicity, but also because they often appear as important building blocks of newly-proposed methods (e.g. Oreshkin et al. (2018); Cao et al. (2020); Gidaris et al. (2019); Yoon et al. (2019)).”
Curriculum-Based Meta-learning (2021, background): “Meta-learning offers an elegant and effective solution to this problem[21, 36, 39, 41, 47].”
Boosting Few-Shot Learning With Adaptive Margin Loss (2020, unknown): “13 TapNet [36] ResNet12 Metric 61.”
Adaptive Subspaces for Few-Shot Learning (2020, methodology): “To the best of our knowledge, subspaces have been used to address FSL in [40, 41] and our preliminary study [42].”