Harmonic Exponential Families on Manifolds
This provides a flexible and efficient probabilistic modeling solution for fields like geosciences, robotics, and computer vision where data lies on manifolds, representing a novel method for a known bottleneck.
The paper tackles the lack of flexible, fast-to-train probabilistic models on manifolds by defining a class of harmonic exponential family distributions, showing that their log-likelihood gradient can be computed efficiently using a non-commutative generalization of the FFT. Experimental results demonstrate that these densities achieve significantly higher likelihood than the best competing method while being orders of magnitude faster to train.
In a range of fields including the geosciences, molecular biology, robotics and computer vision, one encounters problems that involve random variables on manifolds. Currently, there is a lack of flexible probabilistic models on manifolds that are fast and easy to train. We define an extremely flexible class of exponential family distributions on manifolds such as the torus, sphere, and rotation groups, and show that for these distributions the gradient of the log-likelihood can be computed efficiently using a non-commutative generalization of the Fast Fourier Transform (FFT). We discuss applications to Bayesian camera motion estimation (where harmonic exponential families serve as conjugate priors), and modelling of the spatial distribution of earthquakes on the surface of the earth. Our experimental results show that harmonic densities yield a significantly higher likelihood than the best competing method, while being orders of magnitude faster to train.