Adaptive Decision Boundary for Few-Shot Class-Incremental Learning

2025-04-11

Proceedings of the AAAI Conference on Artificial Intelligence

AI Summary

Few-Shot Class-Incremental Learning (FSCIL) confronts three key challenges: rigid decision boundaries, weak inter-class discriminability, and catastrophic forgetting. To address these, we propose the Adaptive Decision Boundary Strategy (ADBS), which learns category-specific decision boundaries that dynamically adapt during incremental learning. A novel inter-class constraint loss is further introduced to jointly optimize boundary parameters and class prototypes. Our method adopts a plug-and-play architecture—requiring no modification to the backbone network—ensuring seamless integration with existing FSCIL frameworks. Extensive experiments on CIFAR-100, miniImageNet, and CUB-200 demonstrate consistent and significant performance gains over state-of-the-art FSCIL approaches, achieving new SOTA results. Notably, ADBS exhibits strong generalization under low-shot settings and multi-stage incremental scenarios, validating its robustness and scalability across diverse data regimes and incremental protocols.

Technology Category

Machine Learning: Transfer, Domain Adaptation, Multi-Task LearningSearch and Optimization: Learning to SearchMultiagent Systems: Adversarial Agents

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingGraph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphsUser Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Few-Shot Class-Incremental Learning (FSCIL) aims to continuously learn new classes from a limited set of training samples without forgetting knowledge of previously learned classes. Conventional FSCIL methods typically build a robust feature extractor during the base training session with abundant training samples and subsequently freeze this extractor, only fine-tuning the classifier in subsequent incremental phases. However, current strategies primarily focus on preventing catastrophic forgetting, considering only the relationship between novel and base classes, without paying attention to the specific decision spaces of each class. To address this challenge, we propose a plug-and-play Adaptive Decision Boundary Strategy (ADBS), which is compatible with most FSCIL methods. Specifically, we assign a specific decision boundary to each class and adaptively adjust these boundaries during training to optimally refine the decision spaces for the classes in each session. Furthermore, to amplify the distinctiveness between classes, we employ a novel inter-class constraint loss that optimizes the decision boundaries and prototypes for each class. Extensive experiments on three benchmarks, namely CIFAR100, miniImageNet, and CUB200, demonstrate that incorporating our ADBS method with existing FSCIL techniques significantly improves performance, achieving overall state-of-the-art results.

Problem

• Adaptive decision boundaries for few-shot class-incremental learning
• Preventing catastrophic forgetting while learning new classes
• Optimizing decision spaces and inter-class distinctiveness

Innovation

• Adaptive Decision Boundary Strategy for FSCIL
• Inter-class constraint loss enhances distinctiveness
• Plug-and-play compatibility with existing methods