conftrace_

2016 JMLR JMLR 2016

Subspace Learning with Partial Information

Abstract

The goal of subspace learning is to find a $k$-dimensional subspace of $\mathbb{R}^d$, such that the expected squared distance between instance vectors and the subspace is as small as possible. In this paper we study subspace learning in a partial information setting, in which the learner can only observe $r \le d$ attributes from each instance vector. We propose several efficient algorithms for this task, and analyze their sample complexity. [abs] [ pdf ][ bib ] © JMLR 2016. (edit, beta)

Authors

Alon Gonen, Dan Rosenbaum, Yonina C. Eldar, Shai Shalev-shwartz

Topics

Machine Learning > Core Methods > Representation Learning Machine Learning > Core Methods > Embedding Learning Machine Learning > Learning Types > Semi-Supervised Learning Machine Learning > Core Methods > Dimensionality Reduction Machine Learning > Core Methods > Feature Selection

Keywords

dimensionality reduction subspace learning feature selection sample complexity matrix approximation matrix completion partial information

More from JMLR 2016

All JMLR 2016 papers →

SPSD Matrix Approximation vis Column Selection: Theories, Algorithms, and Extensions

Combinatorial Multi-Armed Bandit and Its Extension to Probabilistically Triggered Arms

Differentially Private Data Releasing for Smooth Queries

Iterative Hessian Sketch: Fast and Accurate Solution Approximation for Constrained Least-Squares

Estimating Causal Structure Using Conditional DAG Models

Adaptive Lasso and group-Lasso for functional Poisson regression