用分层语言指导弥补视觉数据不足，提升长尾类别增量学习效果
Compensating Visual Insufficiency with Stratified Language Guidance for Long-Tail Class Incremental Learning
- 构建分层语言树，从粗到细组织语义信息
- 动态调整尾部类别的监督信号，缓解数据不平衡问题
- 利用语言结构稳定约束优化，减轻灾难性遗忘
长尾类别增量学习（LT CIL）因尾部类别样本稀少，不仅难以学习，还加剧了在持续演化和不平衡数据分布下的灾难性遗忘。为此，我们利用语言知识的丰富性和可扩展性，分析数据分布，引导大语言模型（LLMs）生成分层语言树，从粗粒度到细粒度组织语义信息。基于此结构，提出分层自适应语言引导，通过可学习权重融合多尺度语义表示，实现对尾部类别的动态监督调节，缓解数据不平衡影响。同时引入分层对齐语言引导，利用语言树的结构稳定性约束优化过程，强化语义与视觉对齐，缓解灾难性遗忘。多个基准上的实验表明，该方法达到当前最优性能。
原文摘要 · Abstract (English)
Long-tail class incremental learning (LT CIL) remains highly challenging because the scarcity of samples in tail classes not only hampers their learning but also exacerbates catastrophic forgetting under continuously evolving and imbalanced data distributions. To tackle these issues, we exploit the informativeness and scalability of language knowledge. Specifically, we analyze the LT CIL data distribution to guide large language models (LLMs) in generating a stratified language tree that hierarchically organizes semantic information from coarse to fine grained granularity. Building upon this structure, we introduce stratified adaptive language guidance, which leverages learnable weights to merge multi-scale semantic representations, thereby enabling dynamic supervisory adjustment for tail classes and alleviating the impact of data imbalance. Furthermore, we introduce stratified alignment language guidance, which exploits the structural stability of the language tree to constrain optimization and reinforce semantic visual alignment, thereby alleviating catastrophic forgetting. Extensive experiments on multiple benchmarks demonstrate that our method achieves state of the art performance.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。