I am an undergraduate student in Information Security at the School of Computer Science and Technology, Tongji University. My research focuses on AI security, LLM, and graph representation learning. I previously worked with Prof. Chengβs Financial Big Data Lab, where I took my first steps into research and contributed as a second author on a paper submitted to TKDE (a CCF-A journal), co-developing a fraud detection framework with generative graph representations.
π₯ News
- 2025.05: Β ππ Proposed a VAE-based IPv6 active address prediction framework, enabling efficient large-scale network scanning.
- 2024.11: Β ππ Developed an adversarial attack pipeline for facial recognition systems with masked momentum PGD, achieving controllable perturbation generation.
π Publications
TKDE 2025
Defending Attacks on Anti-fraud Model with Generative Graph Representations
Jiasheng Wu, Xincheng Wang, Dawei Cheng et al.
Project
- Proposed GSRGNN framework with generative feature reconstruction and structural enhancement modules
- Achieved dual defense against local adversarial perturbations through adaptive fusion mechanism
- Demonstrated SOTA robustness on Web Chat Pay and commercial datasets
- [TKDE (Under Review, CCF-A)] Defending Attacks on Anti-fraud Model with Generative Graph Representations, Jiasheng Wu, Xincheng Wang, Dawei Cheng et al.
π Honors and Awards
- 2024.12 National Endeavor Scholarship
- 2024.12 Tongji University Outstanding Student Scholarship (Second Class)
- 2023.12 National Endeavor Scholarship
- 2024.11 The 8th in the finals, AI Security Track - 7th StrongNet Mimic Defense International Challenge
- 2024.03 Provincial 2nd Prize, C++ Group - 16th Lanqiao Cup Programming Contest
π Educations
- 2021.09 - Present B.Eng. in Information Security, School of Computer Science and Technology, Tongji University.
π» Internships
- 2024.09 - Present Research Assistant, Financial Big Data Lab, Tongji University
Advisor: Prof. Dawei Cheng |Focus: Graph-based Anti-fraud Systems