Rethinking the Sample Relations for Few-Shot Classification

2025-01-23

AI Summary

To address the weak discriminability and poor semantic consistency of pretrained features, as well as coarse-grained modeling of sample relationships in few-shot learning (FSL), this paper proposes Multi-Granularity Relational Contrastive Learning (MGRCL). MGRCL is the first method to systematically distinguish and jointly model three relational granularities: intra-sample, intra-class, and inter-class. It introduces Transformation Consistency Learning (TCL) to enhance semantic robustness and integrates Class Contrastive Learning (CCL) to improve inter-class separability. MGRCL requires no architectural modification to the backbone network and is readily deployable within the pretrain-fine-tune paradigm. It achieves state-of-the-art performance on four major FSL benchmarks—MiniImageNet, TieredImageNet, CUB, and FC100—and serves as a plug-in feature extractor that consistently boosts the performance of diverse FSL algorithms.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Multi-modal Vision
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Feature quality is paramount for classification performance, particularly in few-shot scenarios. Contrastive learning, a widely adopted technique for enhancing feature quality, leverages sample relations to extract intrinsic features that capture semantic information and has achieved remarkable success in Few-Shot Learning (FSL). Nevertheless, current few-shot contrastive learning approaches often overlook the semantic similarity discrepancies at different granularities when employing the same modeling approach for different sample relations, which limits the potential of few-shot contrastive learning. In this paper, we introduce a straightforward yet effective contrastive learning approach, Multi-Grained Relation Contrastive Learning (MGRCL), as a pre-training feature learning model to boost few-shot learning by meticulously modeling sample relations at different granularities. MGRCL categorizes sample relations into three types: intra-sample relation of the same sample under different transformations, intra-class relation of homogenous samples, and inter-class relation of inhomogeneous samples. In MGRCL, we design Transformation Consistency Learning (TCL) to ensure the rigorous semantic consistency of a sample under different transformations by aligning predictions of input pairs. Furthermore, to preserve discriminative information, we employ Class Contrastive Learning (CCL) to ensure that a sample is always closer to its homogenous samples than its inhomogeneous ones, as homogenous samples share similar semantic content while inhomogeneous samples have different semantic content. Our method is assessed across four popular FSL benchmarks, showing that such a simple pre-training feature learning method surpasses a majority of leading FSL methods. Moreover, our method can be incorporated into other FSL methods as the pre-trained model and help them obtain significant performance gains.

Problem

• small sample size
• classification performance
• semantic similarity

Innovation

• Multi-granularity Relational Contrastive Learning
• Transformation Consistency Learning
• Category Contrastive Learning