Spectrally refined unbiased Monte Carlo estimate of the Earth's global radiative cooling
PNAS 2024
CNRS researcher in Computer Graphics
megane.bati@irit.fr
I am a CNRS reasercher in Computer Graphics in the Storm team at IRIT.
My research is at the interface with Physics, focusing on the question of coupling different physics into a single path space in order to solve them using Monte Carlo methods.
For example, my previous work has dealt with the simulation of infrared images, involving conduction in solids, convection in fluids and radiation;
or the coupling of radiative transfer to spectroscopic models in the atmosphere.
The aim of these path-space formulations is to obtain accurate methods that scale up as the system becomes more complex (geometric, temporal complexity, etc.).
In particular, I am interested in the specific computating issues raised by coupled physics in terms of the operationality of these methods:
data structures, formulation of efficient and unbiased estimators, information storage for analysis and inversion.
Previously, I did a post-doc in the Grephe team at Laplace advised by Richard Fournier and Stéphane Blanco,
and in the Storm team at IRIT, collaborating with
Mathias Paulin and Nicolas Mellado.
I defended my Ph.D. thesis in December 2021,
about the inverse design of layered material appearance in the context of geometric optics,
supervised by Romain Pacanowski and Pascal Barla.
This work was conductedi as part of the Manao team, joint between LP2N and Inria Bordeaux Sud-Ouest.
Spectrally refined unbiased Monte Carlo estimate of the Earth's global radiative cooling
PNAS 2024
Coupling Conduction, Convection and Radiative Transfer in a Single Path-Space: Application to Infrared Rendering
ACM TOG 2023 ; presented at SIGGRAPH 2023