Learning with Structured Sparsity

Junzhou Huang, Tong Zhang, Dimitris N. Metaxas

ICML 2009 pp. 417-424

doi:10.1145/1553374.1553429 /icml/2009/huang2009icml-learning-a/

Abstract

This paper investigates a new learning formulation called structured sparsity, which is a natural extension of the standard sparsity concept in statistical learning and compressive sensing. By allowing arbitrary structures on the feature set, this concept generalizes the group sparsity idea. A general theory is developed for learning with structured sparsity, based on the notion of coding complexity associated with the structure. Moreover, a structured greedy algorithm is proposed to efficiently solve the structured sparsity problem. Experiments demonstrate the advantage of structured sparsity over standard sparsity.

PDF ICML Semantic Scholar

Cite

Text

Huang et al. "Learning with Structured Sparsity." International Conference on Machine Learning, 2009. doi:10.1145/1553374.1553429

Markdown

[Huang et al. "Learning with Structured Sparsity." International Conference on Machine Learning, 2009.](https://mlanthology.org/icml/2009/huang2009icml-learning-a/) doi:10.1145/1553374.1553429

BibTeX

@inproceedings{huang2009icml-learning-a,
  title     = {{Learning with Structured Sparsity}},
  author    = {Huang, Junzhou and Zhang, Tong and Metaxas, Dimitris N.},
  booktitle = {International Conference on Machine Learning},
  year      = {2009},
  pages     = {417-424},
  doi       = {10.1145/1553374.1553429},
  url       = {https://mlanthology.org/icml/2009/huang2009icml-learning-a/}
}