Jianhao Yuan

web_prof.jpg

OX1 3PJ, 23 Banbury Road

Oxford, UK

I am a DPhil (PhD) student of University of Oxford, supervised by Prof. Daniele De Martini, Prof. Paul Newman, and Prof. Philip Torr. I also closely work with Prof. Lars Kunze and Dr. Matt Gadd. Prior to this, I completed my undergraduate studies in Engineering Science also at University of Oxford. Previously, I worked as research intern at Meta FAIR, Amazon, and Oxa. My research focuses on developing physically plausible video world models and their applications.

Personal email: yuanjianhao2019@gmail.com

Work email: jianhaoyuan@robots.ox.ac.uk

News

Jan 28, 2026 LikePhys has been accepted to ICLR 2026!
Nov 1, 2025 Our solution WMReward won first place in the ICCV 2025 Physics-IQ Challenge!
Jun 16, 2025 Excited to start as a Research Scientist Intern at FAIR Meta!
Sep 2, 2024 Excited to start as a Applied Scientist Intern at Amazon!
Jun 17, 2024 Excited to start as a Research Engineer Intern at Oxa!

Selected Publications

  1. ICLR
    LikePhys: Evaluating Intuitive Physics Understanding in Video Diffusion Models via Likelihood Preference
    Jianhao Yuan, Fabio Pizzati, Francesco Pinto, and 5 more authors
    ICLR, 2026
    arXiv Code Website
  2. arXiv
    Inference-time Physics Alignment of Video Generative Models with Latent World Models
    Jianhao Yuan, Xiaofeng Zhang, Felix Friedrich, and 7 more authors
    arXiv preprint, 2026
    arXiv
  3. RSS
    RAG-Driver: Generalisable Driving Explanations with Retrieval-Augmented In-Context Learning in Multi-Modal Large Language Model
    Jianhao Yuan, Shuyang Sun, Daniel Omeiza, and 4 more authors
    RSS, 2024
    arXiv Code Website
  4. ICML
    Not Just Pretty Pictures: Toward Interventional Data Augmentation Using Text-to-Image Generators
    Jianhao Yuan, Francesco Pinto, Adam Davies, and 1 more author
    ICML, 2024
    arXiv PDF Code Website
  5. ICLR
    Real-Fake: Effective Training Data Synthesis Through Distribution Matching
    Jianhao Yuan, Jie Zhang, Shuyang Sun, and 2 more authors
    ICLR, 2024
    arXiv PDF Code Website