RDumb++：漂移感知的持续测试时自适应
RDumb++: Drift-Aware Continual Test-Time Adaptation
AI总结：
针对持续测试时自适应中分布快速变化或长期漂移导致性能崩溃的问题，提出RDumb++方法，通过熵和KL散度漂移检测机制与自适应重置策略，在CCC基准上实现约3%的绝对准确率提升。
摘要：
持续测试时自适应（CTTA）旨在仅使用传入的无标签数据流在部署期间更新预训练模型。尽管先前的方法如Tent、EATA等在短期演化偏移下提供了有意义的改进，但当测试分布快速变化或时间跨度极长时，它们表现不佳。CCC基准测试体现了这一挑战，模型在包含750万样本且不断变化损坏类型和严重程度的数据流上运行。我们提出RDumb++，它是RDumb的合理扩展，引入了两种漂移检测机制，即基于熵的漂移评分和KL散度漂移评分，以及自适应重置策略。这些机制使模型能够检测累积的自适应何时变得有害，并在预测崩溃发生前恢复。在包含三种速度和三种种子的CCC-medium（九次运行，每次包含一百万样本）上，RDumb++始终优于RDumb，在整个数据流中实现约3%的绝对准确率提升，同时保持稳定的自适应。关于漂移阈值和重置强度的消融实验进一步表明，漂移感知重置对于防止崩溃和实现可靠的长期CTTA至关重要。
英文摘要：
Continual Test-Time Adaptation (CTTA) seeks to update a pretrained model during deployment using only the incoming, unlabeled data stream. Although prior approaches such as Tent, EATA etc. provide meaningful improvements under short evolving shifts, they struggle when the test distribution changes rapidly or over extremely long horizons. This challenge is exemplified by the CCC benchmark, where models operate over streams of 7.5M samples with continually changing corruption types and severities. We propose RDumb++, a principled extension of RDumb that introduces two drift-detection mechanisms i.e entropy-based drift scoring and KL-divergence drift scoring, together with adaptive reset strategies. These mechanisms allow the model to detect when accumulated adaptation becomes harmful and to recover before prediction collapse occurs. Across CCC-medium with three speeds and three seeds (nine runs, each containing one million samples), RDumb++ consistently surpasses RDumb, yielding approx 3% absolute accuracy gains while maintaining stable adaptation throughout the entire stream. Ablation experiments on drift thresholds and reset strengths further show that drift-aware resetting is essential for preventing collapse and achieving reliable long-horizon CTTA.
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。
