arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.09002cs.LG
收藏

基于表示编码的联邦元学习用于少样本故障诊断

Federated Meta-Learning for Few-Shot Fault Diagnosis with Representation Encoding

Nanjing University of Science and Technology(南京理工大学);The Hong Kong Polytechnic University(香港理工大学);China Academy of Information and Communications Technology(中国信息通信研究院)

Jixuan Cui, Jun Li, Zhen Mei, Kang Wei, Sha Wei, Ming Ding, Wen Chen, Song Guo

AI总结:

针对联邦学习中域差异和数据稀缺导致故障诊断性能下降的问题,提出基于表示编码的联邦元学习框架REFML,利用客户端异质性实现少样本下的分布外泛化,在未见工况和设备上准确率显著提升。

摘要:

基于深度学习的故障诊断(FD)方法需要大量的训练数据,而这些数据难以获取,因为它们分布在不同实体中。联邦学习(FL)使得多个客户端能够在保证数据隐私的情况下协作训练共享模型。然而,客户端之间的域差异和数据稀缺问题会降低全局FL模型的性能。为了解决这些问题,我们提出了一种名为基于表示编码的联邦元学习(REFML)的新框架,用于少样本FD。首先,开发了一种基于表示编码和元学习的新型训练策略。它利用训练客户端之间固有的异质性,有效地将其转化为在未见工作条件或设备类型上的分布外泛化优势。此外,提出了一种自适应插值方法,计算局部模型和全局模型的最优组合作为局部训练的初始化。这有助于进一步利用局部信息来减轻域差异的负面影响。因此,在训练数据有限的情况下,可以在未见工作条件或设备类型上实现高诊断精度。与最先进的方法(如FedProx)相比,所提出的REFML框架在同一设备类型的未见工作条件下测试时,准确率提高了2.17%-6.50%,在完全未见设备类型上测试时,准确率提高了13.44%-18.33%。

英文摘要:

Deep learning-based fault diagnosis (FD) approaches require a large amount of training data, which are difficult to obtain since they are located across different entities. Federated learning (FL) enables multiple clients to collaboratively train a shared model with data privacy guaranteed. However, the domain discrepancy and data scarcity problems among clients deteriorate the performance of the global FL model. To tackle these issues, we propose a novel framework called representation encoding-based federated meta-learning (REFML) for few-shot FD. First, a novel training strategy based on representation encoding and meta-learning is developed. It harnesses the inherent heterogeneity among training clients, effectively transforming it into an advantage for out-of-distribution generalization on unseen working conditions or equipment types. Additionally, an adaptive interpolation method that calculates the optimal combination of local and global models as the initialization of local training is proposed. This helps to further utilize local information to mitigate the negative effects of domain discrepancy. As a result, high diagnostic accuracy can be achieved on unseen working conditions or equipment types with limited training data. Compared with the state-of-the-art methods, such as FedProx, the proposed REFML framework achieves an increase in accuracy by 2.17%-6.50% when tested on unseen working conditions of the same equipment type and 13.44%-18.33% when tested on totally unseen equipment types, respectively.

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。

↑