I am an Associate Professor at the School of Computer Science and Technology, National University of Defense Technology (NUDT), working at the intersection of AI and high-performance computing (HPC) systems.
I received my Ph.D. from the University of New South Wales (UNSW), Sydney, in August 2019. Prior to that, I obtained my B.E. and M.S. degrees from NUDT.
My current research spans two complementary directions: robust & reliable machine learning (adversarial robustness, interpretability, graph neural networks) and AI for HPC systems (using ML to optimize parallel file systems, I/O performance, and burst buffers for large-scale HPC workloads).
I have published in venues such as AAAI, IJCAI, WWW, NeurIPS, ACM MM, ECML/PKDD, ECAI, ACM TWEB, IEEE TPDS, MSST, ACM ICS, and CLUSTER.
What's new
- Mar 2025 One paper on collaborative burst buffer for HPC systems was accepted by ACM ICS 2025.
- Jul 2024 Two papers on data clustering were accepted by ACM MM 2024.
- Jul 2024 One paper on large-scale data dispatching for HPC was accepted by CLUSTER 2024.
- Jul 2024 One paper on generalized adversarial defense for GNNs was accepted by ECAI 2024.
- Jul 2023 One paper on HPC I/O configuration optimizations was accepted by CLUSTER 2023.
- Apr 2023 One paper on heterophilic graph learning was accepted by ECML/PKDD 2023.
Latest publications
See the full publication list (31 papers, 2017–2026) with author lists, venues, DBLP / DOI links, and ✉ corresponding-author markers.
- From Islands to Archipelago: Towards Collaborative and Adaptive Burst Buffer for HPC Systems — ACM ICS, 2025 ✉
- DeloopSGNN: Revisiting Spectral GNNs Through the Lens of Spatial Aggregation — AAAI, 2026 ✦✉
- A Survey on Machine Learning-Based HPC I/O Analysis and Optimization — IEEE TPDS, 2026
- Fully Decentralized Data Distribution for Large-Scale HPC Systems — IEEE TPDS, 2026
- PallasGNN: Curriculum-Based Pattern Mining for Robust GNNs — PAKDD (3), 2026 ✉
- MIST: Towards MPI Instant Startup and Termination on Tianhe HPC Systems — IEEE TPDS, 2025
- MergeFS: Optimizing Node-Local Burst Buffers for Complex HPC Workflows — ICPADS, 2025 ✉
- Talos: A More Effective and Efficient Adversarial Defense for GNN Models — ECAI, 2024 ✉
- Adversarial Examples for Graph Data: Deep Insights into Attack and Defense — IJCAI, 2019 — selected as one of the most influential papers at IJCAI 2019
Research interests
AI for HPC
Applying machine learning to optimize parallel file systems, I/O stacks, and burst buffers in exascale HPC environments.
Storage Systems
Deduplication, caching, and data placement in primary and backup storage systems for cloud and HPC workloads.
Robust Machine Learning
Adversarial robustness, interpretability, and defense mechanisms for deep neural networks and graph neural networks.
Graph Neural Networks
Representation learning on graphs, heterophilic graph learning, and GNN applications to real-world systems.
Education
-
2015 – 2019
Ph.D. in Computer Science
University of New South Wales (UNSW), Sydney, Australia
-
M.S. in Computer Science
National University of Defense Technology (NUDT), Changsha, China
-
B.E. in Computer Science
National University of Defense Technology (NUDT), Changsha, China
Contact
Email: wuhuijun[at]nudt[dot]edu[dot]cn
Office: School of Computer Science and Technology, NUDT, Changsha, Hunan, China
GitHub: sktzwhj
Google Scholar: Huijun Wu