Estimating Variable Structure and Dependence in Multitask Learning via Gradients
Abstract
We consider the problem of hierarchical or multitask modeling where we simultaneously learn the regression function and the underlying geometry and dependence between variables. We demonstrate how the gradients of the multiple related regression functions over the tasks allow for dimension reduction and inference of dependencies across tasks jointly and for each task individually. We provide Tikhonov regularization algorithms for both classification and regression that are efficient and robust for high-dimensional data, and a mechanism for incorporating a priori knowledge of task (dis)similarity into this framework. The utility of this method is illustrated on simulated and real data.
Cite
Text
Guinney et al. "Estimating Variable Structure and Dependence in Multitask Learning via Gradients." Machine Learning, 2011. doi:10.1007/S10994-010-5217-4Markdown
[Guinney et al. "Estimating Variable Structure and Dependence in Multitask Learning via Gradients." Machine Learning, 2011.](https://mlanthology.org/mlj/2011/guinney2011mlj-estimating/) doi:10.1007/S10994-010-5217-4BibTeX
@article{guinney2011mlj-estimating,
title = {{Estimating Variable Structure and Dependence in Multitask Learning via Gradients}},
author = {Guinney, Justin and Wu, Qiang and Mukherjee, Sayan},
journal = {Machine Learning},
year = {2011},
pages = {265-287},
doi = {10.1007/S10994-010-5217-4},
volume = {83},
url = {https://mlanthology.org/mlj/2011/guinney2011mlj-estimating/}
}