A Hybrid Neural-Microfacet BRDF Model for Real-Time Rendering
A compact neural correction augments a GGX microfacet model to better fit measured materials while retaining editable analytical parameters and practical importance sampling.
Rendering · simulation · computational physics
Research for practical graphics and the physics behind it.
I am a Research Scientist at Ubisoft La Forge in Montreal. My core work focuses on real-time rendering, light transport, Monte Carlo methods, and fluid simulation. I also explore numerical relativity, spectral rendering, and non-Euclidean geometry.
Selected work
Recent research
Hierarchical Differentiable Fluid Simulation
A two-step algorithm for differentiable fluid simulation that significantly reduces memory usage by first optimizing bulk forces at reduced resolution, then refining local details over sub-domains while maintaining differentiability.
A Monte Carlo Method for Fluid Simulation
A Monte Carlo-based fluid simulation approach capable of pointwise and stochastic estimation of fluid motion with an in-depth discussion of its distinct advantages, evaluations, and future work.