Computer Science graduate student at China West Normal University.
Competitive programmer, AI security researcher, and builder of practical XCPC infrastructure.

I turn research ideas and contest requirements into reproducible experiments, maintainable systems, and tools that can be used in real environments.

• 🎓 Computer Science graduate student at China West Normal University.
• 🏆 Competitive programmer with provincial gold and silver awards, as well as a national second prize in the Lanqiao Cup C++ competition.
• 🔐 Researching machine learning security, especially backdoor attacks and defenses.
• 🧠 Currently exploring backdoor security in graph neural networks and text-to-image diffusion models.
• 🛠️ Building contest systems, online judges, team-management tools, knowledge bases, and research code.
• 📚 Interested in number theory, trustworthy AI, open-source engineering, and technical education.

• 🥇 Gold Award — 重庆市大学生程序设计竞赛

  Chongqing Collegiate Programming Contest

• 🥈 Silver Award — 四川省大学生程序设计竞赛

  Sichuan Collegiate Programming Contest

• 🥈 National Second Prize — 蓝桥杯全国软件和信息技术专业人才大赛 C++ 组

  Lanqiao Cup National Software Competition — C++ Group

• XCPC Infrastructure — Online judges, contest management systems, team-training platforms, scoreboards, CCS APIs, live contest tooling, and reusable algorithm templates.
• AI Security Research — Reproducible experiments for GNN backdoor defense, diffusion-model backdoors, trigger detection, trigger inversion, and model purification.
• Knowledge and Education — Student-oriented knowledge bases, technical documentation, LaTeX templates, teaching materials, and open educational resources.

Graph Neural Network Security
├── Subgraph backdoor attacks
├── Edge influence analysis
├── Trigger detection and purification
└── Topological consistency

Diffusion Model Security
├── Text-triggered backdoors
├── Semantic and emotional triggers
├── Trigger inversion
└── Detection and mitigation

Competitive Programming
├── Algorithms and data structures
├── Online judge systems
├── Contest infrastructure
└── Training management

Programming Languages
AI and Research
Web Development
Data and Infrastructure
Engineering Tools

• Developing reliable defenses against graph and diffusion-model backdoors.
• Studying unified detection and inversion of textual, syntactic, and semantic triggers.
• Improving NovaJudge for XCPC training and onsite contest operations.
• Building long-term infrastructure and content governance for SCSWiki.
• Organizing reusable algorithms, tools, and documentation for competitive programmers.

Think clearly. Verify experimentally. Build systems that work.

_{Algorithms · Research · Engineering · Open Source}