Compute Less to Get More: Using ORC to Improve Sparse Filtering
Abstract
Sparse Filtering is a popular feature learning algorithm for image classification pipelines. In this paper, we connect the performance of Sparse Filtering with spectral properties of the corresponding feature matrices. This connection provides new insights into Sparse Filtering; in particular, it suggests early stopping of Sparse Filtering. We therefore introduce the Optimal Roundness Criterion (ORC), a novel stopping criterion for Sparse Filtering. We show that this stopping criterion is related with pre-processing procedures such as Statistical Whitening and demonstrate that it can make image classification with Sparse Filtering considerably faster and more accurate.
Cite
Text
Lederer and Guadarrama. "Compute Less to Get More: Using ORC to Improve Sparse Filtering." AAAI Conference on Artificial Intelligence, 2015. doi:10.1609/AAAI.V29I1.9784Markdown
[Lederer and Guadarrama. "Compute Less to Get More: Using ORC to Improve Sparse Filtering." AAAI Conference on Artificial Intelligence, 2015.](https://mlanthology.org/aaai/2015/lederer2015aaai-compute/) doi:10.1609/AAAI.V29I1.9784BibTeX
@inproceedings{lederer2015aaai-compute,
title = {{Compute Less to Get More: Using ORC to Improve Sparse Filtering}},
author = {Lederer, Johannes and Guadarrama, Sergio},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {2015},
pages = {3797-3803},
doi = {10.1609/AAAI.V29I1.9784},
url = {https://mlanthology.org/aaai/2015/lederer2015aaai-compute/}
}