I am a second-year undergraduate student expecting to complete my BSc degree in Applied Mathematics at the University of Liverpool in 2028. Concurrently, I serve as a Research Assistant at PremiLab, where I am supervised by Prof. Qiufeng Wang. I am currently seeking research internships or visiting student opportunities. If you are interested in my research, please feel free to contact me.
My research interests span AI for math reasoning, autoformalization, multimodal large language models and Reinforcement Learning. In my spare time, I also engage in research on optimization theory and mathematical physics. I am also fascinated by representation theory and ontology.
π§ Future
- In the coming year, I plan to launch a large language model for solving Lean4 analytic geometry problems, a dual-feedback prover-model with Parse2Reason step, and a large Olympic physics competition special benchmark. If you are interested in my research, please feel free to contact me.
π₯ News
- 2026.02οΌΒ ππ My first paper βHilbert-Geo: Solving Solid Geometric Problems by Neural-Symbolic Reasoningβ was accepted by CVPR2026.
- 2026.02: Β ππ We have constructed a large benchmark of multimodal geometry problems, Omni-Geo and live-Omni-Geo, which is expected to be submitted to ECCV2026.
- 2026.01: Β ππ This is my first attempt in the field of optimization, thanks for the Mathematical analysis course this semester. I submitted a paper on the optimization behavior of non-smooth perception to ICML2026.
- 2025.11: Β ππ I have completed a manuscript on solving multimodal mathematical problems and submitted it to CVPR2026.
- 2025.06: Β ππ I was supported by the Summer Undergraduate Research Fellowship(SURF) at XJTLU to study large language models for math reasoning.
- 2025.03: Β ππ I will intern at Shanghai AI Lab for several months.
- 2025.03: Β ππ I will serve as the Head of the Academic Department for the Math Club and Physics Club at XJTLU.
- 2025.02: Β ππ I joined PremiLab as a research assistant.
- 2024.12: Β ππ I translated a textbook: Method in contemporary mathematical physics.
π Publications
Hilbert-Geo: Solving Solid Geometric Problems by Neural-Symbolic Reasoning
Ruoran Xu, Haoyu Cheng, Bin Dong, Qiufeng Wang
- By integrating ontology and topological structures, we developed a Multimodal Formalization Parser to enable cross-modal formalization of geometry. Leveraging this foundational integration, we further designed a Reasoning Engine equipped with formal verification mechanisms.
Ruoran Xu, Borong She, Qiufeng Wang, Xiaobo Jin
- Proposed a geometry-aware theoretical paradigm for non-smooth optimization, quantifying local instability via Clarke subdifferential analysis. Proposed the LGI metric, establishing a rigorous and efficient proxy for subdifferential diameter based on randomized directional derivatives. Proved the convergence of S-Adam, which converges to Clarke stationary points at a $O(1/\sqrt{T})$ rate, providing fundamental theoretical support for singularity-aware optimization.
Omni-Geo: Full-Domain Geometry Benchmark with Multimodal Diagram Generation
Ruoran Xu, Wending Gao, Haoyu Cheng, Chengrui Zhang, Maizhen Ning, Qiufeng Wang
- Comingβ¦
π Educations
- 2024.09 - 2028.06 (Expected), University of Liverpool, BSc. Applied Mathematics β’ GPA: 3.81/4.00
π Honors and Awards
- 2025 Honorable Mention, Mathematical Contest in Modeling (MCM)
- 2024 Bronze Award, University Physics Competition in Modeling (UPC)
- 2022 First Prize, Chinese Physics Olympiad (CPHO)
- 2020 Second Prize, National Olympiad in Informatics in Provinces (NOIP)
π» Internships
- 2025.03 - 2025.06, Shanghai AI Lab, Shanghai.
