Nonmyopic Multifidelity Acitve Search
Quan Nguyen, Arghavan Modiri, Roman Garnett
Abstract
Active search is a learning paradigm where we seek to identify as many members of a rare, valuable class as possible given a labeling budget. Previous work on active search has assumed access to a faithful (and expensive) oracle reporting experimental results. However, some settings offer access to cheaper surrogates such as computational simulation that may aid in the search. We propose a model of multifidelity active search, as well as a novel, computationally efficient policy for this setting that is motivated by state-of-the-art classical policies. Our policy is nonmyopic and budget aware, allowing for a dynamic tradeoff between exploration and exploitation. We evaluate the performance of our solution on real-world datasets and demonstrate significantly better performance than natural benchmarks.
BibTeX
@InProceedings{pmlr-v139-nguyen21f,
title = {Nonmyopic Multifidelity Acitve Search},
author = {Nguyen, Quan and Modiri, Arghavan and Garnett, Roman},
booktitle = {Proceedings of the 38th International Conference on Machine Learning},
pages = {8109--8118},
year = {2021},
editor = {Meila, Marina and Zhang, Tong},
volume = {139},
series = {Proceedings of Machine Learning Research},
month = {18--24 Jul},
publisher = {PMLR},
pdf = {http://proceedings.mlr.press/v139/nguyen21f/nguyen21f.pdf},
url = {https://proceedings.mlr.press/v139/nguyen21f.html},
abstract = {Active search is a learning paradigm where we seek to identify as many members of a rare, valuable class as possible given a labeling budget. Previous work on active search has assumed access to a faithful (and expensive) oracle reporting experimental results. However, some settings offer access to cheaper surrogates such as computational simulation that may aid in the search. We propose a model of multifidelity active search, as well as a novel, computationally efficient policy for this setting that is motivated by state-of-the-art classical policies. Our policy is nonmyopic and budget aware, allowing for a dynamic tradeoff between exploration and exploitation. We evaluate the performance of our solution on real-world datasets and demonstrate significantly better performance than natural benchmarks.}
}