Jianhao Yuan

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 interest lies in multimodal generative modeling.

Beyond Visual Quality: A Study of Test-Time Planning with World Action Models

World action models generate actions together with visual predictions of their consequences. Sampling several pairs creates a choice. We study when these predictions provide useful evidence for making that choice.

LikePhys: Evaluating Intuitive Physics Understanding in Video Diffusion Models via Likelihood Preference

LikePhys evaluates whether video diffusion models distinguish physically plausible events from impossible ones, without additional training. It compares model likelihoods for paired videos across twelve scenarios spanning rigid bodies, fluids, deformable materials, and optics, providing a measure that aligns with human judgements of physical plausibility.