余剑峤
James Jianqiao Yu
首页 论文 服务 ENG

教授、博士生导师

计算机科学与技术学院

哈尔滨工业大学(深圳)

广东省深圳市南山区深圳大学城

jqyu(at)hit.edu.cn jqyu(at)ieee.org Google Scholar
Towards Crowdsourced Transportation Mode Identification: A Semi-supervised Federated Learning Approach

作者
Chenhan Zhang, Yuanshao Zhu, Christos Markos, Shui Yu, and James J.Q. Yu*

发表
IEEE Internet of Things Journal, Volume 9, Issue 14, July 2022, Pages 11868--11882

摘要
Privacy-preserving Transportation Mode Identification (TMI) is among the key challenges towards future intelligent transportation systems. With recent developments in federated learning (FL), crowdsourcing has emerged as a promising cost-effective data source for training powerful TMI classifiers without compromising users' data privacy. However, existing TMI approaches have relied heavily on the availability of transportation mode labels, which is often limited in real-world applications. While recent semi-supervised studies have partially addressed this issue by assigning pseudo-labels to unlabeled data, such practice often degrades classification performance as more unlabeled data are incorporated. In response to this issue, we present a semi-supervised FL scheme for TMI termed Mean Teacher Semi-Supervised Federated Learning (MTSSFL). MTSSFL trains a deep neural network ensemble under a novel semi-supervised FL framework, achieving highly accurate and privacy-protected crowdsourced TMI without depending on the availability of massive labeled data. MTSSFL introduces consistency-updating to insert the global model in the gradient updates of the local models that only have unlabeled data to improve their training. We also devise mean-teacher-averaging, a secure parameter aggregation mechanism that further boosts the global model's TMI performance without requiring additional training. Our extensive case studies on a real-world dataset demonstrate that MTSSFL's classification accuracy is merely 1.1% lower than the state-of-the-art semi-supervised TMI approach while being the only one to satisfy FL's privacy-preserving constraints. In addition, MTSSFL can achieve high accuracy with less training overhead due to the proposed semi-supervised learning design.