A Non-Generative Framework and Convex Relaxations for Unsupervised Learning
Abstract
We give a novel formal theoretical framework for unsupervised learning with two distinctive characteristics. First, it does not assume any generative model and based on a worst-case performance metric. Second, it is comparative, namely performance is measured with respect to a given hypothesis class. This allows to avoid known computational hardness results and improper algorithms based on convex relaxations. We show how several families of unsupervised learning models, which were previously only analyzed under probabilistic assumptions and are otherwise provably intractable, can be efficiently learned in our framework by convex optimization.
Cite
Text
Hazan and Ma. "A Non-Generative Framework and Convex Relaxations for Unsupervised Learning." Neural Information Processing Systems, 2016.Markdown
[Hazan and Ma. "A Non-Generative Framework and Convex Relaxations for Unsupervised Learning." Neural Information Processing Systems, 2016.](https://mlanthology.org/neurips/2016/hazan2016neurips-nongenerative/)BibTeX
@inproceedings{hazan2016neurips-nongenerative,
title = {{A Non-Generative Framework and Convex Relaxations for Unsupervised Learning}},
author = {Hazan, Elad and Ma, Tengyu},
booktitle = {Neural Information Processing Systems},
year = {2016},
pages = {3306-3314},
url = {https://mlanthology.org/neurips/2016/hazan2016neurips-nongenerative/}
}