ICML 2018oral22 citations
Approximation Algorithms for Cascading Prediction Models
Abstract
We present an approximation algorithm that takes a pool of pre-trained models as input and produces from it a cascaded model with similar accuracy but lower average-case cost. Applied to state-of-the-art ImageNet classification models, this yields up to a 2x reduction in floating point multiplications, and up to a 6x reduction in average-case memory I/O. The auto-generated cascades exhibit intuitive properties, such as using lower-resolution input for easier images and requiring higher prediction confidence when using a computationally cheaper model.
BibTeX
@InProceedings{pmlr-v80-streeter18a,
title = {Approximation Algorithms for Cascading Prediction Models},
author = {Streeter, Matthew},
booktitle = {Proceedings of the 35th International Conference on Machine Learning},
pages = {4752--4760},
year = {2018},
editor = {Dy, Jennifer and Krause, Andreas},
volume = {80},
series = {Proceedings of Machine Learning Research},
month = {10--15 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v80/streeter18a/streeter18a.pdf},
url = {https://proceedings.mlr.press/v80/streeter18a.html},
abstract = {We present an approximation algorithm that takes a pool of pre-trained models as input and produces from it a cascaded model with similar accuracy but lower average-case cost. Applied to state-of-the-art ImageNet classification models, this yields up to a 2x reduction in floating point multiplications, and up to a 6x reduction in average-case memory I/O. The auto-generated cascades exhibit intuitive properties, such as using lower-resolution input for easier images and requiring higher prediction confidence when using a computationally cheaper model.}
}