• Corpus ID: 309759

Prototypical Networks for Few-shot Learning

@article{Snell2017PrototypicalNF,
  title={Prototypical Networks for Few-shot Learning},
  author={Jake Snell and Kevin Swersky and Richard S. Zemel},
  journal={ArXiv},
  year={2017},
  volume={abs/1703.05175}
}
A recent approach to few-shot classification called matching networks has demonstrated the benefits of coupling metric learning with a training procedure that mimics test. [] Key Method Our method is competitive with state-of-the-art one-shot classification approaches while being much simpler and more scalable with the size of the support set. We empirically demonstrate the performance of our approach on the Omniglot and mini-ImageNet datasets. We further demonstrate that a similar idea can be used for zero…

Figures and Tables from this paper

Prototypical Siamese Networks for Few-shot Learning

  • Junhua WangYongping Zhai
  • Computer Science
    2020 IEEE 10th International Conference on Electronics Information and Emergency Communication (ICEIEC)
  • 2020
We propose a novel architecture, called Prototypical Siamese Networks, for few-shot learning, where a classifier must generalize to new classes not seen in the training set, given only a few examples

Regression Networks for Meta-Learning Few-Shot Classification

This work proposes to meta-learn classification of embedded points by regressing the closest approximation in every class subspace while using the regression error as a distance metric.

Transductive Prototypical Network For Few-Shot Classification

This paper proposes Transductive Prototypical Network (Td-PN), a universal transductive approach that refines the class representations by merging scarce labeled samples and high-confidence ones of target set in a classifying-friendly embedding space.

Learning to Compare: Relation Network for Few-Shot Learning

A conceptually simple, flexible, and general framework for few-shot learning, where a classifier must learn to recognise new classes given only few examples from each, which is easily extended to zero- shot learning.

IPNET: Influential Prototypical Networks for Few Shot Learning

This work proposes a novel version of PN that attributes weights to support samples corresponding to their influence on the support sample distribution based on maximum mean discrepancy (MMD) between the mean embeddings of sample distributions including and excluding the sample.

Meta-Relation Networks for Few Shot Learning

A meta-relation network is proposed to solve the few shot learning problem, where the classifier must learn to recognize new classes given only few examples from each, based on relation networks and Model-Agnostic Meta-Learning training methods.

Few-Shot Learning with Metric-Agnostic Conditional Embeddings

This work introduces a novel architecture where class representations are conditioned for each few-shot trial based on a target image, and deviates from traditional metric-learning approaches by training a network to perform comparisons between classes rather than relying on a static metric comparison.

One-Way Prototypical Networks

A new way of training prototypical few-shot models for just a single class is shown, and a novel Gaussian layer for distance calculation in a prototypical network is proposed, which takes the support examples' distribution rather than just their centroid into account.

Adaptive Prototypical Networks

This paper proposes an approach that intuitively pushes the embeddings of each of the classes away from the others in the meta-testing phase, thereby grouping them closely based on the distinct class labels rather than only the similarity of spatial features.

Prototypical Bregman Networks

This work proposes a metric learner that learns a Bregman divergence by learning its underlying convex function, and proposes a flexible extension to prototypical networks to enable joint learning of the embedding and the divergence, while preserving computational efficiency.
...

References

SHOWING 1-10 OF 38 REFERENCES

Siamese Neural Networks for One-Shot Image Recognition

A method for learning siamese neural networks which employ a unique structure to naturally rank similarity between inputs and is able to achieve strong results which exceed those of other deep learning models with near state-of-the-art performance on one-shot classification tasks.

Matching Networks for One Shot Learning

This work employs ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories to learn a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types.

Write a Classifier: Zero-Shot Learning Using Purely Textual Descriptions

An approach for zero-shot learning of object categories where the description of unseen categories comes in the form of typical text such as an encyclopedia entry, without the need to explicitly defined attributes is proposed.

Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks

We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning

Evaluation of output embeddings for fine-grained image classification

This project shows that compelling classification performance can be achieved on fine-grained categories even without labeled training data, and establishes a substantially improved state-of-the-art on the Animals with Attributes and Caltech-UCSD Birds datasets.

Predicting Deep Zero-Shot Convolutional Neural Networks Using Textual Descriptions

A new model is presented that can classify unseen categories from their textual description and takes advantage of the architecture of CNNs and learn features at different layers, rather than just learning an embedding space for both modalities, as is common with existing approaches.

Towards a Neural Statistician

An extension of a variational autoencoder that can learn a method for computing representations, or statistics, of datasets in an unsupervised fashion is demonstrated that is able to learn statistics that can be used for clustering datasets, transferring generative models to new datasets, selecting representative samples of datasets and classifying previously unseen classes.

One shot learning of simple visual concepts

A generative model of how characters are composed from strokes is introduced, where knowledge from previous characters helps to infer the latent strokes in novel characters, using a massive new dataset of handwritten characters.

Zero-Shot Recognition via Structured Prediction

A novel structured prediction method for maximum a posteriori (MAP) estimation, where parameters account for test-time domain shift from what is predicted primarily using source domain information, is developed and proposed.