Undergraduate researcher in Computer Science and Technology (Hongyi Honor Class) at Wuhan University.
My research interests are representation learning, explainable AI, responsible AI, and multimodal learning. I work with Prof. Mang Ye at Wuhan University. In summer 2026, I completed a research internship at Microsoft Research Asia, mentored by Fangzhao Wu. I am also a co-founder and technical lead of Synesis AI.
- Cross-Species Animal Re-Identification with Semantic Consistency Learning — ECCV 2026, equal contribution. Paper · Code
- See Before You Code: Learning Visual Priors for Spatially Aware Educational Animation Generation — arXiv 2026. Paper · Code
- FedDiG: Federated Continual Graph Learning with Disentangled Generative Replay — under review.
| Project | What it does |
|---|---|
| Synesis AI | AI-assisted learning products and multimodal teaching tools; co-founder and technical lead. |
DeepGraph  |
Research agent for literature discovery, proposal generation, experiment planning, and code execution. |
| ClawOSS | Agent for open-source repository maintenance, from issue selection through pull request follow-up. |
| Codeshift | Lightweight coding agent with repository maps and dependency-aware Python type-hint migration. |
Other public work: LatentCompress and CPU2026spring. I also contribute to OpenClaw.
Email: liyuejia@whu.edu.cn · Academic homepage · Google Scholar