Coresets via Bilevel Optimization for Continual Learning and Streaming
Abstract
Coresets are small data summaries that are sufficient for model training. They can be maintained online, enabling efficient handling of large data streams under resource constraints. However, existing constructions are limited to simple models such as k-means and logistic regression. In this work, we propose a novel coreset construction via cardinality-constrained bilevel optimization. We show how our framework can efficiently generate coresets for deep neural networks, and demonstrate its empirical benefits in continual learning and in streaming settings.
Cite
Text
Borsos et al. "Coresets via Bilevel Optimization for Continual Learning and Streaming." Neural Information Processing Systems, 2020.Markdown
[Borsos et al. "Coresets via Bilevel Optimization for Continual Learning and Streaming." Neural Information Processing Systems, 2020.](https://mlanthology.org/neurips/2020/borsos2020neurips-coresets/)BibTeX
@inproceedings{borsos2020neurips-coresets,
title = {{Coresets via Bilevel Optimization for Continual Learning and Streaming}},
author = {Borsos, Zalán and Mutny, Mojmir and Krause, Andreas},
booktitle = {Neural Information Processing Systems},
year = {2020},
url = {https://mlanthology.org/neurips/2020/borsos2020neurips-coresets/}
}