arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.15944cs.LGcs.AI
收藏

面向多模态概念漂移的李雅普诺夫稳定自适应控制

Lyapunov-Stable Adaptive Control for Multimodal Concept Drift

Department of Electrical and Computer Engineering(电气与计算机工程系);Florida Institute of National Security(佛罗里达国家安全研究所);Applied Artificial Intelligence Group(应用人工智能组);University of Florida(佛罗里达大学)

Tianyu Bell Pan, Mengdi Zhu, Alexa Jordyn Cole, Ronald Wilson, Damon L. Woodard

AI总结:

针对多模态学习中的概念漂移与模态失效问题,提出 LS-OGD 在线自适应控制框架,通过动态调节学习率和模态融合权重获得李雅普诺夫稳定性保证。

摘要:

多模态学习系统在非平稳环境中常因概念漂移而表现不佳,不断变化的数据分布会降低性能。模态特定漂移以及缺乏持续、稳定的适应机制,使这一挑战进一步加剧。本文提出 LS-OGD,这是一种用于在概念漂移存在时实现鲁棒多模态学习的新型自适应控制框架。LS-OGD 使用在线控制器,根据检测到的漂移和不断演变的预测误差,动态调整模型的学习率以及不同数据模态之间的融合权重。我们证明,在有界漂移条件下,LS-OGD 系统的预测误差一致最终有界,并且如果漂移停止,误差将收敛至零。此外,我们证明自适应融合策略能够有效隔离并缓解严重模态特定漂移的影响,从而确保系统弹性和容错能力。这些理论保证为开发可靠且持续适应的多模态学习系统奠定了有原则的基础。

英文摘要:

Multimodal learning systems often struggle in non-stationary environments due to concept drift, where changing data distributions can degrade performance. Modality-specific drifts and the lack of mechanisms for continuous, stable adaptation compound this challenge. This paper introduces LS-OGD, a novel adaptive control framework for robust multimodal learning in the presence of concept drift. LS-OGD uses an online controller that dynamically adjusts the model's learning rate and the fusion weights between different data modalities in response to detected drift and evolving prediction errors. We prove that under bounded drift conditions, the LS-OGD system's prediction error is uniformly ultimately bounded and converges to zero if the drift ceases. Additionally, we demonstrate that the adaptive fusion strategy effectively isolates and mitigates the impact of severe modality-specific drift, thereby ensuring system resilience and fault tolerance. These theoretical guarantees establish a principled foundation for developing reliable and continuously adapting multimodal learning systems.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑