Tree Structured Sparse Coding on Cubes
Abstract
Several recent works have discussed tree structured sparse coding [8, 10, 7, 3], where N data points in R written as thed × N matrixX are approximately decomposed into the product of matrices WZ. HereW is ad×K dictionary matrix, andZ is aK×N matrix of coefficients. In tree structured sparse coding, the rows of Z correspond to nodes on a tree, and the columns of Z are encouraged to be nonzero on only a few branches of the tree; or alternativel y, the columns are constrained to lie on at most a specified number of branches of the tree.
Cite
Text
Szlam. "Tree Structured Sparse Coding on Cubes." International Conference on Learning Representations, 2013.Markdown
[Szlam. "Tree Structured Sparse Coding on Cubes." International Conference on Learning Representations, 2013.](https://mlanthology.org/iclr/2013/szlam2013iclr-tree/)BibTeX
@inproceedings{szlam2013iclr-tree,
title = {{Tree Structured Sparse Coding on Cubes}},
author = {Szlam, Arthur},
booktitle = {International Conference on Learning Representations},
year = {2013},
url = {https://mlanthology.org/iclr/2013/szlam2013iclr-tree/}
}