arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.12490cs.LGcs.AIcs.CV
收藏

迈向跨域持续学习

Towards Cross-Domain Continual Learning

Nanyang Technological University(南洋理工大学);University of South Australia(南澳大学);Singapore Institute of Manufacturing Technology (SIMTech)(新加坡制造技术研究院);Agency for Science, Technology and Research (A*STAR)(新加坡科技研究局)

Marcus de Carvalho, Mahardhika Pratama, Jie Zhang, Chua Haoyan, Edward Yapp

AI总结:

本文提出跨域持续学习方法CDCL,通过在紧凑卷积网络中结合任务间与任务内交叉注意力及伪标签机制,缓解灾难性遗忘并支持无监督跨域学习,在公开UDA数据集上验证了其有效性。

摘要:

持续学习是一个训练学习智能体在不重新访问过去数据的情况下,顺序掌握一系列任务或类别的过程。其挑战在于利用先前获得的知识高效学习新任务,同时避免灾难性遗忘。现有方法主要集中于单一领域,限制了它们对特定问题的适用性。在这项工作中,我们引入了一种名为跨域持续学习(Cross-Domain Continual Learning,CDCL)的新方法,以解决局限于单一监督领域的问题。我们的方法在一个紧凑的卷积网络中结合了任务间和任务内的交叉注意力机制。这种集成使模型能够保持与先前任务特征的对齐,从而延迟任务之间可能发生的数据漂移,同时在相关领域之间执行无监督跨域(UDA)。通过利用任务内特定的伪标签方法,我们确保有标签和无标签样本的准确输入配对,从而增强学习过程。为验证我们的方法,我们在公开的UDA数据集上进行了大量实验,展示了其在跨域持续学习挑战上的积极性能。此外,我们的工作引入了有助于该领域发展的增量思想。我们公开代码和模型,以鼓励进一步探索和复现我们的结果:\url{https://github.com/Ivsucram/CDCL}。

英文摘要:

Continual learning is a process that involves training learning agents to sequentially master a stream of tasks or classes without revisiting past data. The challenge lies in leveraging previously acquired knowledge to learn new tasks efficiently, while avoiding catastrophic forgetting. Existing methods primarily focus on single domains, restricting their applicability to specific problems. In this work, we introduce a novel approach called Cross-Domain Continual Learning (CDCL) that addresses the limitations of being limited to single supervised domains. Our method combines inter- and intra-task cross-attention mechanisms within a compact convolutional network. This integration enables the model to maintain alignment with features from previous tasks, thereby delaying the data drift that may occur between tasks, while performing unsupervised cross-domain (UDA) between related domains. By leveraging an intra-task-specific pseudo-labeling method, we ensure accurate input pairs for both labeled and unlabeled samples, enhancing the learning process. To validate our approach, we conduct extensive experiments on public UDA datasets, showcasing its positive performance on cross-domain continual learning challenges. Additionally, our work introduces incremental ideas that contribute to the advancement of this field. We make our code and models available to encourage further exploration and reproduction of our results: \url{https://github.com/Ivsucram/CDCL}

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。

↑