Guanqiao Qu

Guanqiao Qu

Auto-generated profile, based on publicly available information

Guanqiao Qu

Department of Electrical and Computer Engineering, The University of Hong Kong, Hong Kong, SAR, China

Research Focus & Metrics

Academic Career Length

6Years (2021-Present)

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

Representative Publications

View All Publications
25 June 2025 in IEEE Transactions on Networking
Verified
Journal Article
Closed access
ZLZheng Lin
Zheng Lin
GQGuanqiao Qu
Guanqiao Qu
WWWei Wei
Wei Wei
Xianhao ChenXianhao Chen
Xianhao Chen
KLKin K. Leung
Kin K. Leung
The increasing complexity of deep neural networks poses significant barriers to democratizing AI to resource-limited edge devices. To address this challenge, split federated learning (SFL) has emerged as a promising solution that enables device-server co-training through model splitting. However, although system optimization substantially influences the performance of SFL, the problem remains largely uncharted. In this paper, we first provide a unified convergence analysis of SFL, which quantifies the impact of model splitting (MS) and client-side model aggregation (MA) on its learning performance, laying a theoretical foundation for this field. Based on this convergence bound, we introduce AdaptSFL, an adaptive SFL framework to accelerate SFL under resource-constrained edge computing systems. Specifically, AdaptSFL adaptively controls MS and client-side MA to balance communication-computing latency and training convergence. Extensive simulations across various datasets validate that our proposed AdaptSFL framework takes considerably less time to achieve target accuracy than existing benchmarks.

16 July 2024 in TechRxiv
Verified
Preprint
Open access
GQGuanqiao Qu
Guanqiao Qu
QCQiyuan Chen
Qiyuan Chen
WWWei Wei
Wei Wei
ZLZheng Lin
Zheng Lin
XCXianhao Chen
Xianhao Chen
KHKaibin Huang
Kaibin Huang
On-device large language models (LLMs), referring to running LLMs on edge devices, have raised considerable interest owing to their superior privacy, reduced latency, and bandwidth saving. Nonetheless, the capabilities of on-device LLMs are intrinsically constrained by the limited capacity of edge devices compared to the much more powerful cloud centers. To bridge the gap between cloud-based and on-device AI, mobile edge intelligence (MEI) presents a viable solution to this problem by provisioning AI capabilities within the edge of mobile networks with improved privacy and latency relative to cloud computing. MEI sits between on-device AI and cloud-based AI, featuring wireless communications and more powerful computing resources than end devices. This article provides a contemporary survey on harnessing MEI for LLMs. We first cover the preliminaries of LLMs, starting with LLMs and MEI, followed by resource-efficient LLM techniques. We then illustrate several killer applications to demonstrate the need for deploying LLMs at the network edge and present an architectural overview of MEI for LLMs (MEI4LLM). Subsequently, we delve into various aspects of MEI4LLM, extensively covering edge LLM caching and delivery, edge LLM training, and edge LLM inference. Finally, we identify future research opportunities. We aim to inspire researchers in the field to leverage mobile edge computing to facilitate LLM deployment in close proximity to users, thereby unleashing the potential of LLMs across various privacy-and delay-sensitive applications.

1 January 2021 in Wireless Communications and Mobile Computing
Verified
Journal Article
Open access
PDF file
CHChenguang He
Chenguang He
GQGuanqiao Qu
Guanqiao Qu
LYLiang Ye
Liang Ye
SWShouming Wei
Shouming Wei
Recently, the research on the vehicular ad hoc network (VANET) has been paid more attention by researchers with the quick development of the autonomous driving technology. In the VANET, vehicles can communicate with everything through the route established by routing algorithms. However, the topology of the VANET changes fast because the vehicles move fast. Also, as the number of vehicles increases, the probability of data collision and the transmission latency will also increase when communicating. Therefore, the VANET needs a stable, low‐latency, and efficient route for vehicles to communicate with each other. However, the existing routing algorithms are either unable to aggregate data or are not suitable for the large‐size VANET. In this paper, we consider the vehicle attribute information comprehensively and cluster the vehicles on the road by using the cluster algorithm we propose. We dynamically select the cluster heads at each moment according to their attribute information. We consider all kinds of nodes in the network and the vehicle nodes will communicate with each other through the cluster heads under the two‐level communicating algorithm we propose. Compared with the existing cluster routing algorithm, the algorithm we propose is much more suitable for the large‐size VANET because the cluster heads do not need a gateway to help them communicate. In the simulation part, we set some real street scenes in Simulation of Urban Mobility (SUMO) and the vehicles can move by the traffic rules like in the real world, which is more suitable for the VANET. After analysing the communication performance in Network Simulator version 2 (NS2), we can get a conclusion that the algorithm proposed is superior to the traditional routing algorithm. The route established by the algorithm we propose is much more stable and efficient. And the latency is also lower than the former.

10 May 2024 in IEEE Network
Verified
Journal Article
Closed access
SLSuning Liu
Suning Liu
LYLiang Ye
Liang Ye
HJHanbo Jia
Hanbo Jia
GQGuanqiao Qu
Guanqiao Qu
XLXiaomeng Luo
Xiaomeng Luo
WMWeixiao Meng
Weixiao Meng
Owing to its high mobility and flexible topology, unmanned aerial vehicles (UAV) are envisioned to play an important role in both civil and military fields to enable agile and ubiquitous connectivity. With the development of the B5G/6G communication technologies, the demand for wireless communication services is experiencing a significant increase. Consequently, the spectrum resources are becoming increasingly scarce. However, almost existing full-duplex UAV networking schemes still use the centralized architecture, which is fragile in the circumstances where communication conditions are unpleasant. Thus motivated, we propose a novel decentralized full-duplex UAV ad-hoc network architecture to improve its invulnerability and spectrum efficiency. Then we design the efficient clustering and multiple access protocols respectively under the full-duplex pattern to improve the packet transmission rate and reduce the decentralized network access time effectively. We also conduct a preliminary simulation study to validate the feasibility of the adopted architecture. Finally, we explore the primary challenges and potential avenues for the decentralized full-duplex UAV adhoc networks in conjunction with other cutting-edge technologies.

30 April 2026 in IEEE Transactions on Networking
Verified
Journal Article
Closed access
GQGuanqiao Qu
Guanqiao Qu
ZLZheng Lin
Zheng Lin
QCQian Chen
Qian Chen
JLJian Li
Jian Li
FLFangming Liu
Fangming Liu
KHKaibin Huang
Kaibin Huang
Next-generation mobile networks are expected to facilitate fast AI model downloading to end users. By caching models on edge servers, mobile networks can deliver models to end users with low latency, resulting in a paradigm of >edge model caching. In this paper, we develop a novel model placement framework, called parameter-sharing model caching (TrimCaching). TrimCaching exploits the key observation that a wide range of AI models, such as convolutional neural networks or large language models, can share a significant proportion of parameter blocks containing reusable knowledge, thereby improving storage efficiency. To this end, we formulate a parameter-sharing model placement problem to maximize the cache hit ratio in multi-edge wireless networks by balancing the fundamental tradeoff between storage efficiency and service latency. We show that the formulated problem is a submodular maximization problem with submodular constraints, for which no polynomial-time approximation algorithm exists. To tackle this challenge, we study an important special case, where a small fixed number of parameter blocks are shared across models, which often holds in practice. In such a case, a polynomial-time algorithm with a (1 − ϵ) /2-approximation guarantee is developed. Subsequently, we address the original problem for the general case by developing a greedy algorithm. Simulation results demonstrate that the proposed TrimCaching framework significantly improves the cache hit ratio compared with state-of-the-art content caching without exploiting shared parameters in AI models.

31 March 2026 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
GQGuanqiao Qu
Guanqiao Qu
TLTao Li
Tao Li
QCQian Chen
Qian Chen
Xianhao ChenXianhao Chen
Xianhao Chen
SZSheng Zhou
Sheng Zhou
To support on-device inference, the next-generation mobile networks are expected to support real-time model downloading services to mobile users. However, powerful AI models typically have large model sizes, resulting in excessive end-to-end (E2E) downloading-and-inference (DAI) latency. To address this issue, we propose a simultaneous model downloading and inference (SLIDE) framework, which allows users to perform inference with downloaded layers while simultaneously receiving the remaining layers of the model. To this end, we formulate a task throughput maximization problem by jointly optimizing model provisioning, spectrum bandwidth allocation, and computing resource allocation for multi-user downlink systems. Unlike traditional DAI frameworks, SLIDE introduces recursive dependencies across layers, where inference latency depends recursively on the downloading bandwidth and computing resource allocation for each of the preceding layers. To solve this challenging problem, we design an efficient algorithm that acquires the optimal solution with polynomial-time complexity. Simulation results demonstrate that the proposed SLIDE framework significantly improves task throughput under latency and communication resource constraints compared with the conventional model downloading schemes.