提出数学框架证明判别模型无法避免类别混淆，生成模型才可实现最优增量学习。
Task Confusion and Catastrophic Forgetting in Class-Incremental Learning: A Mathematical Framework for Discriminative and Generative Modelings
- 构建理论框架，揭示判别模型因任务混淆导致性能极限
- 证明生成模型能克服混淆，理论上可实现最优增量学习
- 建议采用生成模型策略，如生成回放或生成分类器
在类别增量学习（class-IL）中，模型需在测试时对所有已见过的类别进行分类，且不依赖任务编号，由此引发任务混淆问题。尽管这是关键挑战，但缺乏理论解释。本文提出一种新的数学框架，证明了不可行定理：由于任务混淆，判别式建模无法实现最优类增量学习。然而，我们建立了可行性定理，表明生成式建模可通过克服任务混淆实现最优类增量学习。随后，我们基于该框架评估了多种主流策略，包括正则化、偏置校正、回放和生成分类器。分析表明，采用生成式建模（无论是用于生成回放还是直接分类）是实现最优类增量学习的关键。
原文摘要 · Abstract (English)
In class-incremental learning (class-IL), models must classify all previously seen classes at test time without task-IDs, leading to task confusion. Despite being a key challenge, task confusion lacks a theoretical understanding. We present a novel mathematical framework for class-IL and prove the Infeasibility Theorem, showing optimal class-IL is impossible with discriminative modeling due to task confusion. However, we establish the Feasibility Theorem, demonstrating that generative modeling can achieve optimal class-IL by overcoming task confusion. We then assess popular class-IL strategies, including regularization, bias-correction, replay, and generative classifier, using our framework. Our analysis suggests that adopting generative modeling, either for generative replay or direct classification (generative classifier), is essential for optimal class-IL.
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。