Compensating Visual Insufficiency with Stratified Language Guidance for Long-Tail Class Incremental Learning

2026-03-23

AI Summary

This work addresses the challenges of learning bias and catastrophic forgetting in long-tailed class-incremental learning, where tail classes suffer from severe data scarcity. To mitigate these issues, the authors propose a hierarchical language-guided framework that leverages large language models to construct a structurally stable hierarchical language tree, generating multi-granularity semantic priors. An adaptive weighting mechanism is introduced to fuse multi-scale semantic representations and dynamically modulate supervision signals. This approach effectively compensates for insufficient visual information and enhances semantic-visual alignment. Extensive experiments demonstrate that the method significantly alleviates both data imbalance and catastrophic forgetting, achieving state-of-the-art performance across multiple benchmarks.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Life-Long and Continual Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

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.

Problem

• long-tail class incremental learning
• catastrophic forgetting
• data imbalance
• tail classes
• visual insufficiency

Innovation

• stratified language guidance
• long-tail class incremental learning
• large language models
• catastrophic forgetting
• semantic visual alignment