Kaiwei Mo

Kaiwei Mo

Auto-generated profile, based on publicly available information

Kaiwei Mo

City University of Hong Kong, Hong Kong, SAR, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
17 July 2025 in Computer Networks
Verified
Journal Article
Closed access
KMKaiwei Mo
Kaiwei Mo
XLXianglong Li
Xianglong Li
ZLZongpeng Li
Zongpeng Li
HXHong Xu
Hong Xu

11 December 2024 in IEEE Internet of Things Journal
Verified
Journal Article
Closed access
XLXianglong Li
Xianglong Li
KMKaiwei Mo
Kaiwei Mo
YHYeqiao Hou
Yeqiao Hou
ZLZongpeng Li
Zongpeng Li
HXHong Xu
Hong Xu
NGNan Guan
Nan Guan
We study resource allocation and task scheduling for mobile Artificial Intelligence Generated Content (AIGC) in a three-layer cloud-edge-device network. Escalating industry demand for computational resources presents significant challenges in resource allocation and optimization, particularly for edge-side AIGC, which faces high computational costs and requires advanced techniques for efficient model deployment on mobile devices. Optimal resource allocation in mobile AIGC networks is naturally formulated into a 0-1 ILP, which is proven NP-hard. We reformulate the problem into both its Comp-Exp and dual forms. Then we design an online auction framework OATS to optimize decisions on instances and time schedules, maximizing social welfare for the AIGC ecosystem. Our analysis demonstrates that OATS achieves high social welfare through appropriate bid acceptance and resource allocation. Simulation results corroborate the theoretical analysis, showcasing the efficacy of our online algorithms.

17 April 2024 in Computer Networks
Verified
Journal Article
Closed access
XLXianglong Li
Xianglong Li
KMKaiwei Mo
Kaiwei Mo
YHYeqiao Hou
Yeqiao Hou
ZLZongpeng Li
Zongpeng Li
HXHong Xu
Hong Xu
CXChun Jason Xue
Chun Jason Xue

11 February 2025 in IEEE Transactions on Cloud Computing
Verified
Journal Article
Closed access
KMKaiwei Mo
Kaiwei Mo
WLWei Lin
Wei Lin
JLJiaxun Lu
Jiaxun Lu
CXChun Jason Xue
Chun Jason Xue
YSYunfeng Shao
Yunfeng Shao
HXHong Xu
Hong Xu
Federated learning (FL) is increasingly adopted to combine knowledge from clients in training without revealing their private data. In order to improve the performance of different participants, personalized FL has recently been proposed. However, considering the non-independent and identically distributed (non-IID) data and limited bandwidth at clients, the model performance could be compromised. In reality, clients near each other often tend to have similar data distributions. In this work, we train the personalized edge-based model in the client-edge-server FL. While considering the differences in data distribution, we fully utilize the limited bandwidth resources. To make training efficient and accurate at the same time, An intuitive idea is to learn as much useful knowledge as possible from other edges and reduce the accuracy loss incurred by non-IID data. Therefore, we devise Grouping Hierarchical Personalized Federated Learning (GHPFL). In this framework, each edge establishes physical connections with multiple clients, while the server physically connects with edges. It clusters edges into groups and establishes client-edge logical connections for synchronization. This is based on data similarities that the nodes actively identify, as well as the underlying physical topology. We perform a large-scale evaluation to demonstrate GHPFL's benefits over other schemes.

6 February 2026 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
XLXianglong Li
Xianglong Li
KMKaiwei Mo
Kaiwei Mo
GFGuang Fang
Guang Fang
ZLZongpeng Li
Zongpeng Li
In maritime navigation, vessels rely on Internet access for communication and entertainment, typically provided by terrestrial-based stations via relay links. For routes that are difficult to cover from the shore, long-endurance drones are deployed to accompany ships and provide Internet connectivity. However, existing systems consider only Internet Service Provider (ISP) competition and do not incorporate the resource selection process that depends on users’ Internet service demand. Furthermore, the flexibility of user utility is limited by ISPs’ pre-defined time-slot allocation methods. To address the unique challenges of maritime environments, including variable connectivity, complex vessel mobility, UAV energy constraints, potential communication disruptions, and fluctuating user demand, we propose an enhanced double auction mechanism that explicitly accounts for real-world constraints that are not typically encountered in terrestrial networks. Our proposed Online Maritime Double Auction Mechanism (OMDAM) aims to maximize the social welfare of the maritime network. We introduce an online algorithm, $A_{online}$ , to solve the social welfare maximization problem, with an inner algorithm, $A_{core}$ , handling the selection of Internet access devices and task allocation between ships and ISPs. Theoretical analysis demonstrates that OMDAM ensures budget balance, individual rationality, and economic efficiency. Simulation results show a performance improvement of up to 17% in social welfare compared with prior art, underscoring the advantages of a maritime-specific design under realistic operational settings.

22 May 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
XLXianglong Li
Xianglong Li
KMKaiwei Mo
Kaiwei Mo
HCHaifeng Chen
Haifeng Chen
XWXinyao Wang
Xinyao Wang
ZLZongpeng Li
Zongpeng Li
UAV network slicing enables flexible, task-oriented resource allocation. However, most current deployments lack true slicing support and face scalability and coordination challenges. Existing methods fail to jointly handle multi-type tasks, UAV deadlines, and hierarchical bandwidth quotas, leading to suboptimal welfare in dynamic environments. We study QoS-aware resource allocation in multi-UAV networks, with the objective of maximizing social welfare. We present a holistic mechanism called Advanced Dynamic Slicing and Task Allocation that supports per-task cross-slice orchestration and per-UAV multi-slice service under per-UAV capacity, coverage, and deadline constraints; we design an online primal–dual algorithm with a dual subroutine that prices multi-slice demands and applies a unit-cost heuristic (travel cost amortized by multi-slice bandwidth requirement) to select feasible UAVs; and we reformulate the integer program to capture travel constraints, bidding prices, and slice-level quotas, using complementary-slackness guidance for real-time task acceptance and resource allocation. In contrast to existing online primal–dual or auction-based UAV schedulers, the proposed mechanism explicitly combines (i) per-UAV concurrent multi-slice service, (ii) per-task cross-slice composition, and (iii) travel-cost-aware deadline constraints in a single optimization and algorithmic framework. Our duality analysis establishes a one-to-one correspondence between dual solutions and primal allocations, ensuring feasible online decisions. Empirical studies verify consistent performance gains, achieving about 10\% higher social welfare than other allocation baselines. We also discuss modeling assumptions, economic interpretation, and limitations, providing a clearer positioning of the proposed mechanism relative to prior work and outlining directions for further empirical validation.