Published on Fri Aug 27 2021

Binocular Mutual Learning for Improving Few-shot Classification

Ziqi Zhou, Xi Qiu, Jiangtao Xie, Jianan Wu, Chi Zhang

Binocular Mutual Learning (BML) is a new way to transfer knowledge from labeled data to a global view. BML uses intra-view and cross-view modeling to combine the global view and the local view.

1
0
0
Abstract

Most of the few-shot learning methods learn to transfer knowledge from datasets with abundant labeled data (i.e., the base set). From the perspective of class space on base set, existing methods either focus on utilizing all classes under a global view by normal pretraining, or pay more attention to adopt an episodic manner to train meta-tasks within few classes in a local view. However, the interaction of the two views is rarely explored. As the two views capture complementary information, we naturally think of the compatibility of them for achieving further performance gains. Inspired by the mutual learning paradigm and binocular parallax, we propose a unified framework, namely Binocular Mutual Learning (BML), which achieves the compatibility of the global view and the local view through both intra-view and cross-view modeling. Concretely, the global view learns in the whole class space to capture rich inter-class relationships. Meanwhile, the local view learns in the local class space within each episode, focusing on matching positive pairs correctly. In addition, cross-view mutual interaction further promotes the collaborative learning and the implicit exploration of useful knowledge from each other. During meta-test, binocular embeddings are aggregated together to support decision-making, which greatly improve the accuracy of classification. Extensive experiments conducted on multiple benchmarks including cross-domain validation confirm the effectiveness of our method.

Fri Sep 13 2019
Computer Vision
Rethinking Zero-Shot Learning: A Conditional Visual Classification Perspective
Zero-shot learning (ZSL) aims to recognize instances of unseen classes solely based on the semantic descriptions of the classes. Existing algorithms usually formulate ZSL as a semantic-visual correspondence problem. We propose to train a deep neural network that directly generates visual feature classifiers from the semantic attributes.
0
0
0
Tue Jul 20 2021
Computer Vision
Boosting few-shot classification with view-learnable contrastive learning
The goal of few-shot classification is to classify new categories with few labeling examples within each class. It is very hard for previous methods to discriminate fine-grained sub-categories in the embedding space without labels. This may lead to unsatisfactory generalization to other subcategories.
0
0
0
Thu May 28 2020
Machine Learning
Boosting Few-Shot Learning With Adaptive Margin Loss
0
0
0
Sun Jun 21 2020
Machine Learning
A Universal Representation Transformer Layer for Few-Shot Image Classification
Few-shot classification aims to recognize unseen classes when presented with only a small number of samples. The problem has inspired the development of benchmarks such as Meta-Dataset. We propose a Universal Representation Transformer (URT) layer, that meta-learns to leverage universal features.
0
0
0
Wed Sep 09 2020
Machine Learning
Proxy Network for Few Shot Learning
The use of a few examples for each class to train a predictive model is a crucial and valuable research direction. We propose a simple but effective end-to-end model that directly learns for class representative and distance metric from data simultaneously.
0
0
0
Wed Apr 01 2020
Machine Learning
Self-Augmentation: Generalizing Deep Networks to Unseen Classes for Few-Shot Learning
Few-shot learning aims to classify unseen classes with a few training examples. We exploit a regional dropout technique called self-mix. We employ a backbone network that has auxiliary branches with its own classifier to enforce knowledge sharing.
0
0
0
Mon Jun 12 2017
NLP
Attention Is All You Need
The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms. Experiments on two machine translation tasks show these models to be superior in
50
215
883
Mon Jun 13 2016
Machine Learning
Matching Networks for One Shot Learning
The standard supervised deep learning paradigm does not offer a satisfactory solution for learning new concepts rapidly from little data. We employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories.
1
0
9
Thu Mar 09 2017
Neural Networks
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
The goal of meta-learning is to train a model on a variety of learning tasks. The parameters of the model are explicitly trained such that a small number of gradient steps with a small amount of training data from a new task will produce good generalization performance.
1
0
3
Mon Mar 30 2020
Computer Vision
Adversarial Feature Hallucination Networks for Few-Shot Learning
Adversarial Feature Hallucination Networks (AFHN) is based on conditional Wasserstein Generative Adversarial networks. Two novel regularizers are incorporated into AFHN to encourage discriminability and diversity of the synthesized features.
0
0
0
Wed Mar 15 2017
Machine Learning
Prototypical Networks for Few-shot Learning
We propose prototypical networks for the problem of few-shot classification. We further extend prototypical networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.
0
0
0
Tue Aug 25 2020
Machine Learning
Transductive Information Maximization For Few-Shot Learning
Transductive Infomation Maximization (TIM) maximizes the mutual information between the query and the label predictions. TIM can be used on top of any base-training feature extractor. It consistently brings between 2% and 5% improvement in accuracy.
0
0
0