Learning Single-Index Models in Gaussian Space

Rishabh Dudeja, Daniel Hsu
Proceedings of the 31st Conference On Learning Theory, PMLR 75:1887-1930, 2018.

Abstract

We consider regression problems where the response is a smooth but non-linear function of a $k$-dimensional projection of $p$ normally-distributed covariates, contaminated with additive Gaussian noise. The goal is to recover the range of the $k$-dimensional projection, i.e., the index space. This model is called the multi-index model, and the $k=1$ case is called the single-index model. For the single-index model, we characterize the population landscape of a natural semi-parametric maximum likelihood objective in terms of the link function and prove that it has no spurious local minima. We also propose and analyze an efficient iterative procedure that recovers the index space up to error $\epsilon$ using a sample size $\tilde{O}(p^{O(R^2/\mu)} + p/\epsilon^2)$, where $R$ and $\mu$, respectively, parameterize the smoothness of the link function and the signal strength. When a multi-index model is incorrectly specified as a single-index model, we prove that essentially the same procedure, with sample size $\tilde{O}(p^{O(kR^2/\mu)} + p/\epsilon^2)$, returns a vector that is $\epsilon$-close to being completely in the index space.

Cite this Paper

BibTeX
@InProceedings{pmlr-v75-dudeja18a, title = {Learning Single-Index Models in Gaussian Space}, author = {Dudeja, Rishabh and Hsu, Daniel}, booktitle = {Proceedings of the 31st Conference On Learning Theory}, pages = {1887--1930}, year = {2018}, editor = {Bubeck, Sébastien and Perchet, Vianney and Rigollet, Philippe}, volume = {75}, series = {Proceedings of Machine Learning Research}, month = {06--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v75/dudeja18a/dudeja18a.pdf}, url = {https://proceedings.mlr.press/v75/dudeja18a.html}, abstract = {We consider regression problems where the response is a smooth but non-linear function of a $k$-dimensional projection of $p$ normally-distributed covariates, contaminated with additive Gaussian noise. The goal is to recover the range of the $k$-dimensional projection, i.e., the index space. This model is called the multi-index model, and the $k=1$ case is called the single-index model. For the single-index model, we characterize the population landscape of a natural semi-parametric maximum likelihood objective in terms of the link function and prove that it has no spurious local minima. We also propose and analyze an efficient iterative procedure that recovers the index space up to error $\epsilon$ using a sample size $\tilde{O}(p^{O(R^2/\mu)} + p/\epsilon^2)$, where $R$ and $\mu$, respectively, parameterize the smoothness of the link function and the signal strength. When a multi-index model is incorrectly specified as a single-index model, we prove that essentially the same procedure, with sample size $\tilde{O}(p^{O(kR^2/\mu)} + p/\epsilon^2)$, returns a vector that is $\epsilon$-close to being completely in the index space.} }
Endnote
%0 Conference Paper %T Learning Single-Index Models in Gaussian Space %A Rishabh Dudeja %A Daniel Hsu %B Proceedings of the 31st Conference On Learning Theory %C Proceedings of Machine Learning Research %D 2018 %E Sébastien Bubeck %E Vianney Perchet %E Philippe Rigollet %F pmlr-v75-dudeja18a %I PMLR %P 1887--1930 %U https://proceedings.mlr.press/v75/dudeja18a.html %V 75 %X We consider regression problems where the response is a smooth but non-linear function of a $k$-dimensional projection of $p$ normally-distributed covariates, contaminated with additive Gaussian noise. The goal is to recover the range of the $k$-dimensional projection, i.e., the index space. This model is called the multi-index model, and the $k=1$ case is called the single-index model. For the single-index model, we characterize the population landscape of a natural semi-parametric maximum likelihood objective in terms of the link function and prove that it has no spurious local minima. We also propose and analyze an efficient iterative procedure that recovers the index space up to error $\epsilon$ using a sample size $\tilde{O}(p^{O(R^2/\mu)} + p/\epsilon^2)$, where $R$ and $\mu$, respectively, parameterize the smoothness of the link function and the signal strength. When a multi-index model is incorrectly specified as a single-index model, we prove that essentially the same procedure, with sample size $\tilde{O}(p^{O(kR^2/\mu)} + p/\epsilon^2)$, returns a vector that is $\epsilon$-close to being completely in the index space.
APA
Dudeja, R. & Hsu, D.. (2018). Learning Single-Index Models in Gaussian Space. Proceedings of the 31st Conference On Learning Theory, in Proceedings of Machine Learning Research 75:1887-1930 Available from https://proceedings.mlr.press/v75/dudeja18a.html.

Related Material