I am currently pursuing a Diploma of Advanced Studies in Data Science with specialization in Machine Learning at ETH Zurich. My topics of interest are Gaussian processes, reinforcement learning, and deep learning. I am also currently collaborating on two research projects in pure machine learning.
Daniel Robert-Nicoud, Andreas Krause and Viacheslav Borovitskiy
AISTATS (2024), under review.
Various applications ranging from robotics to climate science require modeling signals on non-Euclidean domains, such as the sphere. Gaussian process models on manifolds have recently been proposed for such tasks, in particular when uncertainty quantification is needed. In the manifold setting, vector-valued signals can behave very differently from scalar-valued ones, with much of the progress so far focused on modeling the latter. The former, however, are crucial for many applications, such as modeling wind speeds or force fields of unknown dynamical systems. In this paper, we propose novel Gaussian process models for vector-valued signals on manifolds that are intrinsically defined and account for the geometry of the space in consideration. We provide computational primitives needed to deploy the resulting Hodge-Matérn Gaussian vector fields on the two-dimensional sphere and the hypertori. Further, we highlight two generalization directions: discrete two-dimensional meshes and "ideal" manifolds like hyperspheres, Lie groups, and homogeneous spaces. Finally, we show that our Gaussian vector fields constitute considerably more refined inductive biases than the extrinsic fields proposed before.
Copyright © 2023 Daniel Robert-Nicoud - All Rights Reserved.