基于噪声容忍核心集的类增量持续学习
Noise-Tolerant Coreset-Based Class Incremental Continual Learning
AI总结：
针对类增量持续学习中的标签与实例噪声问题，提出基于核心集重放缓冲区的两种噪声容忍算法，理论推导鲁棒性界限，实验证明其显著提升分类精度并减少遗忘。
摘要：
计算机视觉的许多应用要求在部署后能够适应新的数据分布。这种适应性需要能够进行持续学习（CL）的算法。持续学习者必须具有可塑性，以适应新任务，同时最小化对先前任务的遗忘。然而，持续学习为噪声进入训练流程并干扰持续学习提供了途径。本研究聚焦于类增量学习（CIL）背景下的标签噪声和实例噪声，其中新类别随时间添加到分类器中，且无法访问过去类别的外部数据。我们旨在理解通过从基于核心集思想构建的记忆中重放项目来工作的持续学习方法的敏感性。我们推导了在一般加性噪声威胁模型下，此类方法对不相关实例噪声鲁棒性的新界限，揭示了若干见解。将理论付诸实践，我们创建了两种持续学习算法来构建噪声容忍的重放缓冲区。我们在五个不同数据集上，在标签噪声和不相关实例噪声下，实证比较了先前基于记忆的持续学习者和所提出算法的有效性。我们表明，现有的基于记忆的持续学习并不鲁棒，而所提出的方法在最大化分类准确性和最小化噪声CIL环境中的遗忘方面表现出显著改进。
英文摘要：
Many applications of computer vision require the ability to adapt to novel data distributions after deployment. Adaptation requires algorithms capable of continual learning (CL). Continual learners must be plastic to adapt to novel tasks while minimizing forgetting of previous tasks.However, CL opens up avenues for noise to enter the training pipeline and disrupt the CL. This work focuses on label noise and instance noise in the context of class-incremental learning (CIL), where new classes are added to a classifier over time, and there is no access to external data from past classes. We aim to understand the sensitivity of CL methods that work by replaying items from a memory constructed using the idea of Coresets. We derive a new bound for the robustness of such a method to uncorrelated instance noise under a general additive noise threat model, revealing several insights. Putting the theory into practice, we create two continual learning algorithms to construct noise-tolerant replay buffers. We empirically compare the effectiveness of prior memory-based continual learners and the proposed algorithms under label and uncorrelated instance noise on five diverse datasets. We show that existing memory-based CL are not robust whereas the proposed methods exhibit significant improvements in maximizing classification accuracy and minimizing forgetting in the noisy CIL setting.
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。
