About Me

Tianqi Shen


My name is Tianqi Shen (沈天琦).


I am a third-year Ph.D. student in the MOFA Lab in the Department of Computer Science at the City University of Hong Kong (CityUHK), advised by Ziye Ma. I am supported by a Postgraduate Studentship (Presidential Assistant Professors Scheme).


I welcome academic discussions and potential collaborations. Feel free to contact me at tianqshen5-c [at] my.cityu.edu.hk.


Education and Employment

  • 2024–Present: Ph.D. Student; Department of Computer Science; City University of Hong Kong (CityUHK).
  • 2023–2024: Researcher, Institute of Mining Artificial Intelligence, China Coal Research Institute (Chinese Academy of Coal Science).
  • 2020–2023: M.Sc. (Research degree); Department of Image Processing; School of Astronautics; Beihang University.
  • 2018–2020: Second B.Sc. degree; Department of Applied Mathematics; School of Mathematical Sciences; Beihang University.
  • 2016–2020: B.Sc. degree; Department of Planning, Navigation and Control; School of Astronautics; Beihang University.

Research Interests

  • Non-convex optimization and landscape analysis
  • Optimizer design for modern large language models (LLMs)
  • World model and scene reconstruction
  • Visual planning, navigation and tracking

News and Updates

  • Jul 2026: One patent was granted; link to be added soon.
  • Jun 2026: Our manuscript on a river-valley perspective for landscape analysis of the Muon optimizer is now available on arXiv. (link).
  • Feb 2026: Our manuscript on simulated tensor over-parameterization for non-convex matrix sensing is now available on arXiv. (link).
  • Mar 2026: Our paper on efficient aerial tracker was accepted by IEEE Transactions on Cognitive and Developmental Systems. (link)
  • Oct 2025: I received CityUHK’s Outstanding Academic Performance Award for Research Degree Students (non-local UGC-funded students).
  • Sep 2025: I passed the Qualifying Examination (QE) and advanced to Ph.D. candidacy.
  • Dec 2024: Our paper on topology-aware 3D Gaussian Splatting was accepted to AAAI 2025 and selected for an oral presentation. (link)
  • Jun 2024: One patent was granted. (link)