arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.08464cs.LG
收藏

通过动态潜在表示增强演化域泛化

Enhancing Evolving Domain Generalization through Dynamic Latent Representations

The Chinese University of Hong Kong(香港中文大学);Hong Kong Baptist University(香港浸会大学)

Binghui Xie, Yongqiang Chen, Jiaqi Wang, Kaiwen Zhou, Bo Han, Wei Meng, James Cheng

AI总结:

本文提出MISTS框架,通过信息论约束解耦动态与不变特征,并结合域自适应分类器,解决非平稳环境中演化新域的泛化问题。

摘要:

域泛化是机器学习系统面临的一项关键挑战。先前的域泛化方法专注于在若干静态域中提取域不变特征,以便能够泛化到新域。然而,在非平稳任务中,新域会沿着诸如时间等潜在的连续结构不断演化,仅仅提取不变特征不足以泛化到不断演化的新域。尽管如此,由于演化特征与不变特征之间存在冲突,在单一模型中同时学习二者并非易事。为弥合这一差距,我们构建因果模型来刻画与这两种模式相关的分布偏移,并提出通过一个名为基于互信息的序列自编码器(MISTS)的新框架来同时学习动态特征和不变特征。MISTS 将信息论约束施加到序列自编码器上,以解耦动态特征和不变特征,并利用域自适应分类器基于演化信息和不变信息进行预测。我们在合成数据集和真实世界数据集上的实验结果表明,MISTS 能够成功捕获演化信息和不变信息,并在演化域泛化任务中展现出有前景的结果。

英文摘要:

Domain generalization is a critical challenge for machine learning systems. Prior domain generalization methods focus on extracting domain-invariant features across several stationary domains to enable generalization to new domains. However, in non-stationary tasks where new domains evolve in an underlying continuous structure, such as time, merely extracting the invariant features is insufficient for generalization to the evolving new domains. Nevertheless, it is non-trivial to learn both evolving and invariant features within a single model due to their conflicts. To bridge this gap, we build causal models to characterize the distribution shifts concerning the two patterns, and propose to learn both dynamic and invariant features via a new framework called Mutual Information-Based Sequential Autoencoders (MISTS). MISTS adopts information theoretic constraints onto sequential autoencoders to disentangle the dynamic and invariant features, and leverage a domain adaptive classifier to make predictions based on both evolving and invariant information. Our experimental results on both synthetic and real-world datasets demonstrate that MISTS succeeds in capturing both evolving and invariant information, and present promising results in evolving domain generalization tasks.

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。

↑