计算机科学
适应(眼睛)
水准点(测量)
背景(考古学)
领域(数学分析)
考试(生物学)
域适应
质量(理念)
培训(气象学)
软件部署
时域
人工智能
分割
机器学习
软件工程
心理学
计算机视觉
数学
分类器(UML)
气象学
认识论
神经科学
哲学
物理
古生物学
地理
生物
数学分析
大地测量学
作者
Robert A. Marsden,Mario Döbler,Bin Yang
出处
期刊:Cornell University - arXiv
日期:2022-08-16
被引量:1
标识
DOI:10.48550/arxiv.2208.07736
摘要
Experiencing domain shifts during test-time is nearly inevitable in practice and likely results in a severe performance degradation. To overcome this issue, test-time adaptation continues to update the initial source model during deployment. A promising direction are methods based on self-training which have been shown to be well suited for gradual domain adaptation, since reliable pseudo-labels can be provided. In this work, we address two problems that exist when applying self-training in the setting of test-time adaptation. First, adapting a model to long test sequences that contain multiple domains can lead to error accumulation. Second, naturally, not all shifts are gradual in practice. To tackle these challenges, we introduce GTTA. By creating artificial intermediate domains that divide the current domain shift into a more gradual one, effective self-training through high quality pseudo-labels can be performed. To create the intermediate domains, we propose two independent variations: mixup and light-weight style transfer. We demonstrate the effectiveness of our approach on the continual and gradual corruption benchmarks, as well as ImageNet-R. To further investigate gradual shifts in the context of urban scene segmentation, we publish a new benchmark: CarlaTTA. It enables the exploration of several non-stationary domain shifts.
科研通智能强力驱动
Strongly Powered by AbleSci AI