Henry Hong Xu
Associate Professor
Works on machine learning systems, data center networks and AI for IT operations.
Henry Hong Xu's lab at CUHK works in AI Systems & Hardware. 66 papers since 2023, 31 at top venues, mostly ASPLOS and DAC. 19 PhD students. 16 alumni. Applications close .
At a glance
- papers, 2023–2026
- 17/ year
- at T1 venues
- 47%
- PhD students
- 19
- MS students
- 0
- alumni listed
- 16
Papers by yearfrom the lab homepage, 2023–2026
- 15 papers2026 · T1 8 · collecting
- 19 papers2025 · T1 9
- 19 papers2024 · T1 9
- 13 papers2023 · T1 5
T1 venuesOtherVenues: ASPLOS, DAC, INFOCOM, ICML, SIGCOMM
Research focuspaper titles that use each topic's words
- networks4 papers
- data center0 papers
- operations0 papers
- learning systems0 papers
- center0 papers
Paper list
From the lab homepage.
202615 papers
- ICML (T1)AGZO: Activation-Guided Zeroth-Order Optimization for LLM Fine-Tuning (opens in a new tab)
- SIGCOMM (T1)Dynamic Compute and Network Orchestration for Disaggregated RL (opens in a new tab)
- ASPLOS (T1)Dynamic Sparsity in Large-Scale Video DiT Training (opens in a new tab)
- USENIX OSDIEfficient GPU-Centric Evolving Graph Processing at Scale (opens in a new tab)
- HiGen Workshop at CVPREvolutionary Eye-Hand Coordination with Large Language Models (opens in a new tab)
- ICML (T1)FOCUS: DLLMs Know How to Tame Their Compute Bound (opens in a new tab)
- ICAIFMemory × Self-Evolution: When Agents with Cognitive Frameworks Meet Market Paradigm Shifts in Investment (opens in a new tab)
- USENIX NSDIOffloading Cloud Network Services at Production Scale with SONiC DASH SmartSwitch (opens in a new tab)
- MLSys (T1)PRISM: Parametrically Refactoring Inference for Draft Models in Speculative Decoding (opens in a new tab)
- IWQoSReasonCache: Accelerating Large Reasoning Model Serving through KV Cache Sharing (opens in a new tab)
- NeurIPS (T1)RoutingBench: Can Agentic Routing Analysis Scale to Production Datacenter Networks? (opens in a new tab)
- APNetSaving Lost Intents in Network Configurations (opens in a new tab)
202519 papers
- ACM Transactions on Storage, accepted in DecC3: Finding Counting-related Crash Consistency Atomicity Bugs for Persistent Memory (opens in a new tab)
- FAISysCloud-Native Systems for Generative AI Applications: Current Trends and Open Challenges (opens in a new tab)
- DAC (T1)Easz: An Agile Transformer-based Image Compression Framework for Resource-constrained IoTs (opens in a new tab)
- PACMIFrontier: Simulating the Next Generation of LLM Inference Systems (opens in a new tab)
- IEEE Transactions on Cloud ComputingGHPFL: Advancing Personalized Edge-Based Learning Through Optimized Bandwidth Utilization (opens in a new tab)
- USENIX ATCInferring Likely Counting-related Atomicity Program Properties for Persistent Memory (opens in a new tab)
- IEEE Transactions on Network Science and EngineeringLearning for Accelerated Traffic Engineering with Differentiable Network Modeling (opens in a new tab)
- NeurIPS (T1)Learning Provably Improves the Convergence of Gradient Descent (opens in a new tab)
- EMNLP (T1)Logits-Based Finetuning (opens in a new tab)
- SOSP (T1)Mycroft: Tracing Dependencies in Collective Communication Towards Reliable LLM Training (opens in a new tab)
- LLM-NETNetSophon: Enabling Runtime Copilot for Programmable Dataplane for Cloud Operators (opens in a new tab)
- ASPLOS (T1)Performance Prediction of On-NIC Network Functions with Multi-Resource Contention and Traffic Awareness (opens in a new tab)
202419 papers
- INFOCOM (T1)A Learning-Only Method for Multi-Cell Multi-User MIMO Sum Rate Maximization (opens in a new tab)
- IEEE Transactions on NetworkingA Practical Congestion Control Algorithm for Low-Latency Interactive Video Streaming (opens in a new tab)
- IEEE/ACM IWQoSAdaptive Personalized Federated Learning for Non-IID Data with Continual Distribution Shift (opens in a new tab)
- ICCAn Online Auction Approach to UAV Scheduling and Trajectory Planning (opens in a new tab)
- ACM ICPPArlo: Serving Transformer-based Language Models with Dynamic Input Lengths (opens in a new tab)
- SIGCOMM (T1)Automating Network Configuration with Natural Language Intents (opens in a new tab)
- INFOCOM (T1)BREAK: A Holistic Approach for Efficient Container Deployment among Edge Clouds (opens in a new tab)
- DAC (T1)Disentangle, Align and Generalize: Learning A Timing Predictor from Different Technology Nodes (opens in a new tab)
- INFOCOM (T1)Dynamic Learning-based Link Restoration in Traffic Engineering with Archie (opens in a new tab)
- IEEE ICNCFed2Com: Towards Efficient Compression in Federated Learning (opens in a new tab)
- DAC (T1)Fracturing-aware Curvilinear ILT via Circular E-beam Mask Writer (opens in a new tab)
- ACM HotStorageIs Low Similarity Threshold A Bad Idea in Delta Compression? (opens in a new tab)
202313 papers
- IEEE CogMIABC: Automatic Bottom-up Construction of Configuration Knowledge Base for Multi-Vendor Networks (opens in a new tab)
- USENIX ATCAccelerating Distributed MoE Training and Inference with Lina (opens in a new tab)
- EMNLP (T1)Adaptive Gating in Mixture-of-Experts based Language Models (opens in a new tab)
- INFOCOM (T1)Cross-Camera Inference on the Constrained Edge (opens in a new tab)
- Elsevier Computer NetworkEfficient Real-time Video Conferencing with Adaptive Frame Delivery (opens in a new tab)
- IEEE Transactions on Mobile ComputingEfficient Time-Series Data Delivery in IoT with Xender (opens in a new tab)
- IEEE ICNPEnabling Load Balancing for Lossless Datacenters (opens in a new tab)
- Workshop on the Synergy of Scientific and Machine Learning Modeling, ICMLLearning to Optimize Non-Convex Sum-Rate Maximization Problems (opens in a new tab)
- DAC (T1)LRSDP: Low-Rank SDP for Triple Patterning Lithography Layout Decomposition (opens in a new tab)
- EuroSys (T1)Lyra: Elastic Scheduling for Deep Learning Clusters (opens in a new tab)
- SIGCOMM (T1)Meili: Towards SmartNIC as a Service (opens in a new tab)
- ACM MMMoby: Empowering 2D Models for Efficient Point Cloud Analytics on the Edge (opens in a new tab)
Where alumni went16 listed
Apply to CUHKFall 2027
Computer Science and Engineering MPhil/PhD
- Round
- HKPFS / main round
- Opens
- —
- GRE
- See program page
- English
- Interview in English
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- CNTsinghua
Shengen Yan
18 T1 - USStanford
Fredrik Kjolstad
- Compilers & programming languages
- CNZhejiang
De Ma
- Accelerators, chips & architecture
- CHEPFL
Mahsa Shoaran
- Accelerators, chips & architecture
- USUCSD
Arun Kumar
- Data management & big data
- USUCSD
Bill Lin
- Accelerators, chips & architecture
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
CUHKcloses