An end-to-end mobile system for proactive stress monitoring from wearable signals supported by an extensible split federated learning system that enables on-device LoRA fine-tuning of large language models.
Qi Guo, Qi-Yue Xu, Jia-Xiang Geng et al.· Proceedings of the Twenty-se...· 0 citations
Movable agents, such as autonomous vehicles and robots, should be able to autonomously adapt their pre-trained object recognition models to new environments without human annotations. While existing test-time adaptation (TTA) methods address label-free model adaptation, they typically treat test data as independent sna...
Ya-Nan Ma, Yi-Hang Tao, Zheng-Ru Fang et al.· 0 citations
HO-FL is introduced, a hybrid-order FL framework that trains a model's bottom segment with ZO optimization and its top segment with FO optimization, and shows that hybrid-order local training can retain much of the full-FO performance with substantially lower client memory requirements.
Qi-Yuan Chen, Xian Wu, Ya-Nan Ma et al.· 0 citations
CFLoRA is presented, a federated LoRA scheme that partitions latent LoRA channels into two complementary sets in every communication round, and eliminates bilinear terms in matrix multiplications, making federated aggregation exact.
A polynomial-time algorithm is developed that transforms the original problem into two sub-problems and obtains a sub-optimal solution with a constant approximation guarantee and demonstrates that BALANCE consistently outperforms conventional AD and SD and significantly improves task throughput.
Guan-Qiao Qu, Shuo Chen, Qian Chen et al.· 0 citations
A polynomial-time algorithm is developed that transforms the original problem into two sub-problems and obtains a sub-optimal solution with a constant approximation guarantee and demonstrates that BALANCE consistently outperforms conventional AD and SD and significantly improves task throughput.
LDE, Learning from Distributed ``Eyes", a novel framework that transforms collaborative perception (CP) into a source of high-quality supervision for model adaptation, is proposed, which consistently outperforms both the pre-trained models and state-of-the-art unsupervised adaptation methods.
Fluid antenna systems (FAS) have emerged as a promising paradigm for wireless communications, enabling channel reconfigurability that offers a novel spatial degree of freedom. Nevertheless, efficiently acquiring accurate and high-resolution channel state information (CSI) in FAS remains challenging, primarily due to it...
Xue-Feng Wang, Yu-Hang Li, Yang Lu et al.· IEEE Transactions on Wireles...· 0 citations
L-shaped SFT is presented, a split fine-tuning framework that removes the need for continuous client participation and introduces one-shot SFT, in which clients upload activations once and then go offline while the server continues optimization over cached representations.
FlexP-SFT is proposed, a novel aggregation-free framework for personalized split federated fine-tuning, which fundamentally eliminates the client-side aggregation process and introduces a layer-flexible alignment strategy to balance personalization and generalization capabilities.
Jiaxiang Geng, Tianjun Yuan, Pengchao Han et al.· 3 citations
SplitLite is proposed, a communication-efficient split federated LoRA fine-tuning method that exploits the low effective rank structure of consecutive-epoch activation and gradient residuals, thereby significantly reducing both activation uplink and gradient downlink traffic.
This work establishes an efficient optimization framework for SFL under resource-constrained networks that jointly optimizes model splitting and resource allocation to minimize training cost, which is defined as the weighted sum of latency and energy costs.
Wei Wei, Xian-Hao Chen· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.