Embedded Files
Haoyu Wang (王 昊宇)
Email : haoyu.wang.2024@phdcs.smu.edu.sg
Github : https://github.com/haoyuwang99
LinkedIn : https://www.linkedin.com/in/haoyu-wang-898003239/
Wechat : StanWang_142
I am currently a PhD candidate at the Singapore Management University co-advised by Prof. C.M. Poskitt and Prof. Jun Sun, where I am part of the Center for Research on Intelligent Software Engineering (RISE). My current research interest mainly focus on enforcing the safety and security for AI software. Besides, I am also interested in fuzzing fundamental software systems including compilers, smart contracts, etc.
EDUCATION
2024.8 to now > PhD in Computer Science, Singapore Management University
2021.9 to 2024.1 > Master in Software Engineering, Tianjin University
2017.9 to 2021.7 > Bachelor in Software Engineering, Tianjin University
WORK EXPERIENCES
2024.4 to present > Research Engineer, Singapore Management University
2020.5 to 2020.10 > Backend Developer Intern, ByteDance
PUBLICATIONS
AI safety/security:
LLM-enabled Applications Require System-Level Threat Monitoring (arxiv)
Yedi Zhang, Haoyu Wang, Xianglin Yang, Jinsong Dong, and Jun Sun.
Pro2Guard: Proactive Runtime Enforcement of LLM Agent Safety via Probabilistic Model Checking (arxiv)
Haoyu Wang, C.M. Poskitt, Jun Sun, and Jiali Wei.
AgentSpec: Customizable Runtime Enforcement for Safe and Reliable LLM Agents (ICSE '26)
Haoyu Wang, C.M. Poskitt and Jun Sun.
Evaluating and Mitigating Linguistic Discrimination in Large Language Models: Perspectives on Safety Equity and Knowledge Equity (IJCAI '25)
Guoliang Dong, Haoyu Wang, Jun Sun and Xinyu Wang.
Traditional Software Testing:
Natural Adversaries: Fuzzing Autonomous Vehicles with Realistic Roadside Object Placements (ICST '26)
Yang Sun, Haoyu Wang, C.M. Poskitt, and Jun Sun.
Exploring JVM Garbage Collector Testing with Event-Coverage (TOSEM '24)
Kai Zheng, Zan Wang, Yingquan Zhao, Junjie Chen, Hanmo You, Haoyu Wang, Yiheng Du, Tianchang Gao
MLIRSmith: Random Program Generation for Fuzzing MLIR Compiler Infrastructure (ASE '23)
Haoyu Wang, Junjie Chen, Chuyue Xie, Shuang Liu, Zan Wang, Qingchao Shen and Yingquan Zhao.
The compiler fuzzing tool MLIRSmith detected 100+ previously unknown bugs in MLIR .
sFuzz2.0: Storage-Access Pattern Guided Smart Contract Fuzzing (JSEP '23)
Haoyu Wang, Zan Wang, Shuang Liu, Jun Sun, Yingquan Zhao, Yan Wan and Tai D. Nguyen.
Natural Test Generation for Precise Testing of Question Answering Software (ASE '22)
Qingchao Shen, Junjie Chen, Jie M Zhang, Haoyu Wang, Shuang Liu and Menghan Tian.
SERVICES
TA of CS441 Network Security, CS427 AI Safety.
Artifact Evaluation for FormaliSE 2026.
INVITED TALKS
[Jan. 2026] I gave a live code demonstration to MSD inc. about using AgentSpec and Pro2Guard to enforce the safety of LLM agents. [Virtual]
[April. 2025] I gave a talk to AI security group of MSD inc. about enforcing safe and reliable agent via AgentSpec. [Virtual]
[Sep. 2023] I gave a talk at ASE about the fuzzing the MLIR infrastructure with MLIRSmith. [Luxemburg]
[Dec. 2022] I gave a talk at ChinaSoft about the smart contract fuzzing. [Virtual]
PROMOTING LINK
Peixin Zhang: Research Scientist at Singapore Management University
Yedi Zhang: Post-doc at National University of Singapore
Yige Li: Research Scientist at Singapore Management University
Ningke Li: PhD at National University of Singapore
Yingquan Zhao: Associate Professor at Tianjin University
Qingchao Shen : PhD at Tianjin University
Bo Wang: Associate Professor at Beijing Jiaotong University
Page updated
Google Sites
Report abuse