Biography
Welcome! I am an Assistant Professor at The University of Hong Kong, directing the Networked AI Systems and Security Lab (NAISS Lab) in the Department of Data and Systems Engineering. I earned my Ph.D. and M.S. in Computer Science at The University of Chicago, advised by Prof. Nick Feamster. I also worked closely with Prof. Vyas Sekar. I received my B.Eng. (Honors) from the Yingcai Honors College of UESTC. I am a recipient of the 2025 ML and Systems Rising Stars recognition, the Carnegie Bosch Fellowship, the ACTION AI Fellowship, and the Daniels Fellowship.
My research lies in computer networking and security. I focus on developing accessible, reliable, and performant machine learning systems for network data analysis, and using network data analysis to address critical issues in security and privacy. My work spans LLM agentic systems, enterprise networks, the Internet of Things, and cyber-physical systems. It has been recognized and published at venues including USENIX Security, USENIX NSDI, ACM SIGMETRICS, ACM CoNEXT, NDSS, and UbiComp.
I served as head of the NSF ACTION AI Institute student advisory council. I serve SIGCOMM, CCS, USENIX Security, IMC, PETS/PoPETS, and USENIX NSDI through technical program committees and pre-review work. I also review for NeurIPS, UbiComp/IMWUT, USENIX ATC, IEEE INFOCOM, IEEE TDSC, IEEE TIFS, and IEEE IoTJ. My research has been featured in Forbes, The Wall Street Journal, and ACM TechNews.
Prospective Students & Researchers
I am recruiting Ph.D. students, MPhil students, RAs, and postdoctoral researchers for Spring / Fall 2027. If you are interested in our research, please get in touch. English flyer · 中文介绍
I am receiving a high volume of applications and may take longer to respond. Thank you for your patience.
Research
Select an area to explore its research and related publications.
Applications
Operational ML Frameworks
Inference
Monitoring & Adaptation
Training → Inference ⇄ Monitoring & Adaptation
Network Management
Supporting network operations with traffic measurement, synthetic data generation, efficient traffic inference, and adaptation to changing network conditions. These systems support downstream tasks such as traffic classification, performance prediction, and intrusion detection.
Related Publications
Network Security & Privacy
Studying the security and privacy of networked and cyber-physical systems, including IoT traffic, GPS spoofing, synthetic data, and shared computing infrastructure.
Related Publications
Network Data Analysis
My research connects operational machine learning frameworks with applications in network management, security, and privacy.
Research Impact
- Deployment. LEAF was deployed in a billion-dollar-scale cellular network system at Verizon to address concept drift.
- Patent Application. NetDiffusion led to Diffusion-Based Network Traffic Generation (US 20260089070 A1).
- Recognition. Our distributed multi-antenna GPS spoofing work received the Best Full Paper Award at ACM WiSec 2025.
- Media Coverage. Our GPS spoofing research was featured in Forbes, The Wall Street Journal, and ACM TechNews. NetDiffusion was also featured by UChicago Computer Science.
Related Publications
Break Network Data Silos
Generating high-fidelity network traces to support learning across scarce or fragmented data, while preserving protocol structure and multi-flow behavior. Related work evaluates synthetic tabular data for analytical queries.
Related Publications
Merge Multimodal Information
Combining network traffic with video and sensor data to understand connected environments, while examining what data is necessary and what privacy it exposes.
Related Publications
Handle Large Traffic Volumes
Making ML traffic analysis practical through efficient serving, compact representations, fast–slow model architectures, and end-to-end pipeline optimization. Related work extends efficient inference to long-context LLMs at the edge.
Related Publications
Deal with Network Evolution
Detecting and adapting to changing traffic patterns, concept drift, and class imbalance so learning systems can continue to operate as networks evolve.
Related Publications
Selected Publications
Underlined authors: NAISS Lab members. * Equal contribution.
2026
FlowWise: Stateful Fast-Slow Model Serving for Streaming Traffic Intelligence
ACM Symposium on Cloud Computing (SoCC), 2026.
GhostAccess: Attacking the GPU on the Multi-tenant Cloud via CPU LLC under Unified Memory
IEEE/ACM MICRO, 2026.
WiFinger: Fingerprinting Noisy IoT Event Traffic Using Packet-level Sequence Matching
NDSS, 2026. (Acceptance rate: 17.89%; reviewed submissions)
Towards On-Device Evidence Gathering for Intimate Partner Infiltration: A Feasibility Study for Joint Identity–Action Detection
ACM IMWUT/UbiComp, 2026.
NetSSM: Multi-Flow and State-Aware Network Trace Generation using State-Space Models
ACM CoNEXT, 2026.
Quantifying the Privacy Implications of High-Fidelity Synthetic Network Traffic
ACM IMC, 2026. (Acceptance rate: 18%; cycle 2)
Generative Active Adaptation for Drifting Imbalanced Network Intrusion Detection
ACM CoNEXT, 2026.
2025
Algorithmic Data Minimization for Machine Learning over Internet-of-Things Data Streams
Proceedings of the VLDB Endowment (PVLDB), 18(13):5652–5661, 2025.
JITI: Dynamic Model Serving for Just-in-Time Traffic Inference
ACM CoNEXT, 2025. (Acceptance rate: 18.35%; PACMNET)
Distributed Multi-Antenna GPS Spoofing Attack Using Off-the-Shelf Devices
ACM WiSec, 2025.
CATO: End-to-end Optimization of ML Traffic Analysis Pipelines
USENIX NSDI, 2025. (Acceptance rate: 12.5%)
2024
2023
AMIR: Active Multimodal Interaction Recognition from Video and Network Traffic in Connected Environments
UbiComp / IMWUT, 2023.
LEAF: Navigating Concept Drift in Cellular Networks
ACM CoNEXT, 2023. (Acceptance rate: 18.5%; long papers)
Generative, High-Fidelity Network Traces
ACM Workshop on Hot Topics in Networks (HotNets), 2023.
2021
News & Updates
Talk
I gave an invited talk at Rockfish Data on Building Practical Systems for High-Fidelity Synthetic Data. Thanks to Prof. Vyas Sekar for the invitation!
Paper
Jul 2026 · Our paper GhostAccess has been accepted at IEEE/ACM MICRO 2026!
Paper
Jun 2026 · Our work on on-device evidence gathering has been accepted at ACM IMWUT/UbiComp 2026!
Paper
Jun 2026 · Our paper on generative active adaptation has been accepted at ACM CoNEXT 2026!
Service
Jun 2026 · I am serving on the USENIX Security 2027 program committee.
More Updates
Award
Jun 2026 · New Smart Traffic Fund project on road works and temporary traffic arrangements approved; serving as Co-I.
Award
Jun 2026 · I was awarded RMB 500,000 in GPU research compute from Miracle Plus.
Paper
Apr 2026 · FlowWise accepted at ACM SoCC 2026. Details
Service
Apr 2026 · I am serving on the PETS 2027 and FMSys 2026 program committees, and the ML and Systems Rising Stars 2026 selection committee.
Award
Feb 2026 · I was awarded HKD 400,000 from the Faculty Interdisciplinary Fund as PI.
Talk
Jan 2026 · I was invited to speak at the Harvard AI and Robotics Seminar. Details
Service
Jan 2026 · I am serving on the SIGCOMM, CCS, IMC, and PETS 2026 program committees.
Paper
Dec 2025 · WiFinger accepted at NDSS 2026. Details
Milestone
Dec 2025 · NAISS Lab attended CoNEXT 2025 and MobiCom 2025 in Hong Kong.
Paper
Sept 2025 · JITI accepted at ACM CoNEXT 2025. Details
Paper
Aug 2025 · Our work on data minimization for IoT streams accepted at VLDB 2026. Details
Milestone
Aug 2025 · I joined The University of Hong Kong as an Assistant Professor and launched NAISS Lab. Details
Award
Jul 2025 · Our GPS spoofing paper received the Best Full Paper Award at ACM WiSec 2025. Details
Award
May 2025 · I was recognized as a 2025 ML and Systems Rising Star. Details