A Lie-algebraic kernel reparameterizes 3D rotationally anisotropic Gaussian processes with explicit principal length-scales and SO(3) orientations, matching full SPD flexibility but improving interpretability over axis-aligned ARD.
Mardia and Peter E
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Interpretable Machine Learning for Spatial Science: A Lie-Algebraic Kernel for Rotationally Anisotropic Gaussian Processes
A Lie-algebraic kernel reparameterizes 3D rotationally anisotropic Gaussian processes with explicit principal length-scales and SO(3) orientations, matching full SPD flexibility but improving interpretability over axis-aligned ARD.