ICML 2018oral66 citations

Multi-Fidelity Black-Box Optimization with Hierarchical Partitions

Rajat Sen, Kirthevasan Kandasamy, Sanjay Shakkottai

Abstract

Motivated by settings such as hyper-parameter tuning and physical simulations, we consider the problem of black-box optimization of a function. Multi-fidelity techniques have become popular for applications where exact function evaluations are expensive, but coarse (biased) approximations are available at much lower cost. A canonical example is that of hyper-parameter selection in a learning algorithm. The learning algorithm can be trained for fewer iterations – this results in a lower cost, but its validation error is only coarsely indicative of the same if the algorithm had been trained till completion. We incorporate the multi-fidelity setup into the powerful framework of black-box optimization through hierarchical partitioning. We develop tree-search based multi-fidelity algorithms with theoretical guarantees on simple regret. We finally demonstrate the performance gains of our algorithms on both real and synthetic datasets.

BibTeX
@InProceedings{pmlr-v80-sen18a,
  title = 	 {Multi-Fidelity Black-Box Optimization with Hierarchical Partitions},
  author =       {Sen, Rajat and Kandasamy, Kirthevasan and Shakkottai, Sanjay},
  booktitle = 	 {Proceedings of the 35th International Conference on Machine Learning},
  pages = 	 {4538--4547},
  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/sen18a/sen18a.pdf},
  url = 	 {https://proceedings.mlr.press/v80/sen18a.html},
  abstract = 	 {Motivated by settings such as hyper-parameter tuning and physical simulations, we consider the problem of black-box optimization of a function. Multi-fidelity techniques have become popular for applications where exact function evaluations are expensive, but coarse (biased) approximations are available at much lower cost. A canonical example is that of hyper-parameter selection in a learning algorithm. The learning algorithm can be trained for fewer iterations – this results in a lower cost, but its validation error is only coarsely indicative of the same if the algorithm had been trained till completion. We incorporate the multi-fidelity setup into the powerful framework of black-box optimization through hierarchical partitioning. We develop tree-search based multi-fidelity algorithms with theoretical guarantees on simple regret. We finally demonstrate the performance gains of our algorithms on both real and synthetic datasets.}
}