Binglun Wang

I'm a PhD student at University College London (UCL), advised by Prof. He Wang and Prof. Niloy Mitra. My PhD (2024–2028) is supported by the UCL Research Studentship. Previously, I worked at Adobe Research London and UCL with Dr. Duygu Ceylan, Prof. Niloy Mitra, and Prof. Ray Dolan.

My research interests span Computer Graphics, Machine Learning, and Computational Physics, which together help me study motion, learning, and physical systems. I aim to explore how intelligent agents can learn to interact with complex physical environments. In my PhD, I focus on physics-based character-fluid motion.

Outside of research, I enjoy music, art, and sports .

Publication

SWIM: Compact Environment Representation for Reinforcement Learning Human Swimming
Binglun Wang, Doğa Yılmaz, Niloy Mitra, Edmond S. L. Ho, He Wang
Preprint, 2026
Project Page / Preprint

A framework for reinforcement learning of human swimming control. Our main contributions are: (1) the first work to learn stable, controllable, and natural full-body swimming policies, from only a single reference motion; and (2) a novel body–water environment representation that enables this learning. SWIM achieves improved stability, goal satisfaction, and physical realism, with a 63.9% success rate on held-out generalization, compared with 36.8% for the strongest baseline.

ProteusNeRF: Fast Lightweight NeRF Editing using 3D-Aware Image Context
Binglun Wang, Niladri Shekhar Dutt, Niloy Mitra
Proceedings of the ACM on Computer Graphics and Interactive Techniques
I3D 2024 / Journal
Project Page / Paper / Presentation / Code / arXiv

A fast 3D editing framework powered by generative AI. It addresses two key challenges: cross-view inconsistency and slow optimization. Its key novelty is a simple yet efficient idea that enables generative models to perform 3D editing without additional training, reducing editing time from hours to minutes.