Abstract:In Few-Shot Learning (FSL), the objective is to correctly recognize new samples from novel classes with only a few available samples per class. Existing methods in FSL primarily focus on learning transferable knowledge from base classes by maximizing the information between feature representations and their corresponding labels. However, this approach may suffer from the "supervision collapse" issue, which arises due to a bias towards the base classes. In this paper, we propose a solution to address this issue by preserving the intrinsic structure of the data and enabling the learning of a generalized model for the novel classes. Following the InfoMax principle, our approach maximizes two types of mutual information (MI): between the samples and their feature representations, and between the feature representations and their class labels. This allows us to strike a balance between discrimination (capturing class-specific information) and generalization (capturing common characteristics across different classes) in the feature representations. To achieve this, we adopt a unified framework that perturbs the feature embedding space using two low-bias estimators. The first estimator maximizes the MI between a pair of intra-class samples, while the second estimator maximizes the MI between a sample and its augmented views. This framework effectively combines knowledge distillation between class-wise pairs and enlarges the diversity in feature representations. By conducting extensive experiments on popular FSL benchmarks, our proposed approach achieves comparable performances with state-of-the-art competitors. For example, we achieved an accuracy of 69.53% on the miniImageNet dataset and 77.06% on the CIFAR-FS dataset for the 5-way 1-shot task.

Adaptive few-shot learning with a fair priori distribution

Feature Transformation for Few-Shot Learning

Reweighting and Information-Guidance Networks for Few-Shot Learning

Adaptive Attribute Distribution Similarity for Few-Shot Learning

Adaptive Metric-weight and Bias Network for Few-shot Learning.

Fair Few-shot Learning with Auxiliary Sets

Recent Advances of Few-Shot Learning Methods and Applications

Adaptive Fine-Tuning Strategy for Few-Shot Learning

A Few-shot-learning-based Method to Object Recognition in Multiple Scenarios.

LibFewShot: A Comprehensive Library for Few-Shot Learning

Domain-Adaptive Few-Shot Learning

Less is More: A Closer Look at Semantic-based Few-Shot Learning

Adaptive Prototype Few-Shot Image Classification Method Based on Feature Pyramid

Low-resolution Few-Shot Learning Via Multi-Space Knowledge Distillation

Autonomous Perception and Adaptive Standardization for Few-Shot Learning

ENHANCING PROTOTYPICAL FEW-SHOT LEARNING BY LEVERAGING THE LOCAL-LEVEL STRATEGY

Instance-Level Embedding Adaptation for Few-Shot Learning

Balancing Feature Alignment and Uniformity for Few-Shot Classification.

Federated Few-Shot Learning with Adversarial Learning

Generalizing from a Few Examples: A Survey on Few-Shot Learning