ABO-Med: Accelerated Bilevel Optimization for Few-Shot Medical Image Classification

2026-09-27

AI Summary

This study addresses the high computational overhead of second-order optimization and the limited accuracy in few-shot medical image classification by proposing an efficient meta-learning method that integrates first-order bilevel optimization with the MAML framework. Theoretically, this work establishes, for the first time, the theoretical optimality of MAML-style meta-learning. Methodologically, a modality-aware adaptive data augmentation strategy, MedRAug, is designed to enhance model generalization. Experimental results demonstrate that the proposed approach improves classification accuracy by 1.99% to 18.76% across multiple public medical datasets, while the augmentation strategy yields additional performance gains of 2.20% to 6.34%. Overall, this research achieves efficient and scalable few-shot learning for medical image analysis.

Abstract

In recent years, bilevel optimization has been widely used in a variety of machine learning tasks. However, prior bilevel optimization algorithms generally require the computation of second-order information, which limits their practical scalability. Only recently has a first-order paradigm for bilevel optimization been established, attaining near-optimal theoretical guarantees for solving bilevel optimization problems. In this paper, we propose ABO-Med, a scalable instantiation of this paradigm for few-shot learning, by incorporating it into the model-agnostic meta-learning (MAML) framework and tailoring it to medical image classification. We also introduce Medical Adaptive RandomAugment (MedRAug), a modality-aware augmentation strategy designed for medical images. Theoretically, ABO-Med establishes the optimality of MAML-type meta-learning approaches. Empirically, ABO-Med outperforms prior baselines on several public medical datasets, with gains of 1.99% to 18.76%, while MedRAug further improves the average accuracy by 2.20% to 6.34%. Additional cross-domain experiments, augmentation ablation studies, backbone ablation studies, and training efficiency analysis further validate the effectiveness and efficiency of the proposed method.

Problem

• Bilevel Optimization
• Few-Shot Learning
• Medical Image Classification
• Scalability

Innovation

• Bilevel Optimization
• Few-Shot Learning
• Model-Agnostic Meta-Learning (MAML)
• Medical Image Classification
• Data Augmentation