Headshot of Xin Li

Xin Li (李鑫)

Ph.D. Student

Nanyang Technological University

About Me

I am a Ph.D. student at Nanyang Technological University (NTU), advised by Prof. Chau Yuen. My work sits at the intersection of large language models and engineering automation, with three focus areas:

  • LLM Evaluation — Designing practical, trustworthy ways to measure what LLMs can and cannot do. Representative works: DafnyComp (Preprint; formal verification & compositional generalization), WirelessMathBench (Findings of ACL 2025; math-centric modeling for wireless), WritingPreferenceBench (Preprint; subjective writing preferences across cultures), and COIG-Writer (Preprint; Chinese creative writing with thought processes).
  • LLMs for Wireless Communication — Studying how LLMs can be integrated into wireless research (Project Maxwell). Representative works: LACP (AI4NextG @ NeurIPS 2025; agent communication protocol), WirelessMathBench (Findings of ACL 2025; domain benchmark), and WirelessMathLM (Preprint; 0.5B-7B models with verification-based RL on WirelessMathBench-XL).
  • LLMs for Robotics — Leveraging LLMs alongside classical perception/SLAM/navigation to build embedded AI systems.

Previously at MSRA, MEGVII, and Gausium Robotics, I led perception projects from prototype to deployment across SLAM, VIO, and radio/vision stacks.

I received my M.E. from Peking University and my B.E. from Northeastern University, China.

Open to Collaboration: If you are seeking any form of academic collaboration 🤝 on LLMs, wireless communications, or robotics, please feel free to email me at xin019@e.ntu.edu.sg.

I am actively looking for research internship opportunities in related areas. If you have any openings, I'd love to hear from you!

Recent News

Selected Publications

(* indicates equal contribution. For a full list, please see my Google Scholar ⤻.)

Local Success Does Not Compose: Benchmarking Large Language Models for Compositional Formal Verification

X. Xu*, Xin Li*, X. Qu, J. Fu, B. Yuan

arXiv, 2025

Homepage ⤻ arXiv ⤻

WirelessMathLM: Teaching Mathematical Reasoning for LLMs in Wireless Communications with Reinforcement Learning

Xin Li, M. Liu, Y. Zhu, W. Zhang, L. Wei, J. An, C. Yuen

arXiv, 2025

Homepage ⤻ arXiv ⤻

Beyond Correctness: Evaluating Subjective Writing Preferences Across Cultures

S. Ying, Y. Li, X. Qu, Xin Li, et al., G. Zhang, W. Huang, W. Che, C. Lin

arXiv, 2025

Homepage ⤻ arXiv ⤻

COIG-Writer: A High-Quality Dataset for Chinese Creative Writing with Thought Processes

Y. Li, S. Ying, X. Qu, Xin Li, et al., E. Zhang

arXiv, 2025

Homepage ⤻ arXiv ⤻

LACP: LLM Agent Communication Protocol Requires Urgent Standardization

Xin Li, M. Liu, C. Yuen

AI4NextG @ NeurIPS, 2025

Homepage ⤻ arXiv ⤻

Re:Form—Reducing Human Priors in Scalable Formal Software Verification with RL in LLMs

C. Yan*, F. Che*, X. Huang*, X. Xu*, Xin Li*, Y. Li*, X. Qu*, et al., J. Fu

Technical Report, 2025

arXiv ⤻ Code ⤻

WirelessMathBench: A Mathematical Modeling Benchmark for LLMs in Wireless Communications

Xin Li, M. Liu, L. Wei, J. An, M. Debbah, C. Yuen

Findings of ACL, 2025

Homepage ⤻ arXiv ⤻

Onboard Terrain Classification via Stacked Intelligent Metasurface-Diffractive Deep Neural Networks from SAR Level-0 Raw Data

M. Liu, Xin Li, J. An, C. Yuen

ML4RS @ ICLR, 2025

Homepage ⤻ arXiv ⤻

TransPathNet: A Two-Stage Framework for Indoor Radio Map Prediction

Xin Li, R. Liu, S. Xu, S. G. Razul, C. Yuen

ICASSP, 2025

Homepage ⤻ Code ⤻ arXiv ⤻

Co-planar Parametrization for Stereo-SLAM and Visual-Inertial Odometry

Xin Li*, Y. Li*, E. P. Örnek, J. Lin, F. Tombari

IEEE Robotics and Automation Letters (RA-L), 2020

arXiv ⤻ Code ⤻

Leveraging Planar Regularities for Point-Line Visual-Inertial Odometry

Xin Li*, Y. He*, J. Lin, X. Liu

IROS, 2020

arXiv ⤻ Code ⤻

Experience

  • Research Assistant @ NTU Singapore (Apr 2024 - Jan 2025) — Supervised by Prof. Chau Yuen, IEEE Fellow
  • SLAM Algorithm Engineer @ Gausium Robotics, Singapore (Mar 2022 - Feb 2024)
  • Research Intern @ Microsoft Research Asia, Beijing (Sep 2020 - Mar 2021) — Supervised by Dr. Yang Liu and Dr. Yizhong Zhang
  • Research Intern @ MEGVII, Beijing (Feb 2019 - Mar 2020) — Supervised by Dr. Yijia He

Education

Professional Activities & Awards

  • Conference Reviewer: NeurIPS (2024-25), ICLR (2025-2026), ICML (2025), SIGGRAPH (2025), AAAI (2026), AISTATS (2025-2026), ICRA (2022-25), IROS (2021-25)
  • Journal Reviewer: IEEE RA-L, ACM TOG, IEEE TNNLS
  • Workshop Organizer: AIR4D@IROS 2025
  • Grant: Cohere Labs Catalyst Grant
  • Grant: OpenAI Researcher Access Program
  • Award: NTU Research Scholarship