Overlapping Cover Local Regression Machines
Abstract
We present the Overlapping Domain Cover (ODC) notion for kernel machines, as a set of overlapping subsets of the data that covers the entire training set and optimized to be spatially cohesive as possible. We show how this notion benefit the speed of local kernel machines for regression in terms of both speed while achieving while minimizing the prediction error. We propose an efficient ODC framework, which is applicable to various regression models and in particular reduces the complexity of Twin Gaussian Processes (TGP) regression from cubic to quadratic. Our notion is also applicable to several kernel methods (e.g., Gaussian Process Regression(GPR) and IWTGP regression, as shown in our experiments). We also theoretically justified the idea behind our method to improve local prediction by the overlapping cover. We validated and analyzed our method on three benchmark human pose estimation datasets and interesting findings are discussed.
Cite
@article{arxiv.1701.01218,
title = {Overlapping Cover Local Regression Machines},
author = {Mohamed Elhoseiny and Ahmed Elgammal},
journal= {arXiv preprint arXiv:1701.01218},
year = {2017}
}
Comments
Long Article with more experiments and analysis of conference paper "Overlapping Domain Cover for Scalable and Accurate Regression Kernel Machines", presented orally 2015 at the British Machine Vision Conference 2015 (BMVC)