Journal
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
Volume 24, Issue 3, Pages 3528-3540Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TITS.2022.3225116
Keywords
Model watermark; cooperative intelligent transportation system (C-ITS); federated learning; intellectual property
Ask authors/readers for more resources
Federated learning is beneficial for building better cooperative intelligent transportation systems (C-ITS) with intellectual property protection. Existing research on watermark-based protection in centralized models is not effective in federated learning due to differences in watermark distribution and loss of global model accuracy. To address these issues, we propose a multi-party entangled watermark algorithm in federated learning. Our scheme includes a local watermark enhancement algorithm to solve local watermark failures and a global entanglement aggregation algorithm to mitigate the loss of global model accuracy. Experimental results show significant advantages of our proposal in model accuracy and watermark success rate compared to existing watermark schemes in federated learning.
Federated learning is good for building better cooperative intelligent transportation system (C-ITS). Intellectual property protection in C-ITS brings many benefits to all vehicles. Although the protection of model intellectual property by watermark has received much research attention, the existing works only deploy watermark in centralized models. Due to the difference of watermark distribution among vehicles, the global model accuracy of watermark in federated learning is significantly reduced or the local watermark is invalid. To solve these problems, we propose a multi-party entangled watermark algorithm in federated learning. Specifically, in the local training, we propose a watermark enhancement algorithm, which solves the problem of local watermark failure. Then, in the global aggregation, we propose an entanglement aggregation algorithm, which solves the problem of a great loss of global model accuracy. We conduct extensive experiments on public datasets to show the superiority of our proposal. The results show that our scheme can obtain more than 16 % and 31 % advantages in model accuracy and watermark success rate, respectively, compared with existing watermark schemes in federated learning.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available