No-regret Algorithms for Multi-task Bayesian Optimization

Sayak Ray Chowdhury, Aditya Gopalan
Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, PMLR 130:1873-1881, 2021.

Abstract

We consider multi-objective optimization (MOO) of an unknown vector-valued function in the non-parametric Bayesian optimization (BO) setting. Our aim is to maximize the expected cumulative utility of all objectives, as expressed by a given prior over a set of scalarization functions. Most existing BO algorithms do not model the fact that the multiple objectives, or equivalently, tasks can share similarities, and even the few that do lack rigorous, finite-time regret guarantees that capture explicitly inter-task structure. In this work, we address this problem by modelling inter-task dependencies using a multi-task kernel and develop two novel BO algorithms based on random scalarization of the objectives. Our algorithms employ vector-valued kernel regression as a stepping stone and belong to the upper confidence bound class of algorithms. Under a smoothness assumption that the unknown vector-valued function is an element of the reproducing kernel Hilbert space associated with the multi-task kernel, we derive worst-case regret bounds for our algorithms that explicitly capture the similarities between tasks. We numerically benchmark our algorithms on both synthetic and real-life MOO problems, and show the advantages offered by learning with multi-task kernels.

Cite this Paper


BibTeX
@InProceedings{pmlr-v130-chowdhury21c, title = { No-regret Algorithms for Multi-task Bayesian Optimization }, author = {Chowdhury, Sayak Ray and Gopalan, Aditya}, booktitle = {Proceedings of The 24th International Conference on Artificial Intelligence and Statistics}, pages = {1873--1881}, year = {2021}, editor = {Banerjee, Arindam and Fukumizu, Kenji}, volume = {130}, series = {Proceedings of Machine Learning Research}, month = {13--15 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v130/chowdhury21c/chowdhury21c.pdf}, url = {https://proceedings.mlr.press/v130/chowdhury21c.html}, abstract = { We consider multi-objective optimization (MOO) of an unknown vector-valued function in the non-parametric Bayesian optimization (BO) setting. Our aim is to maximize the expected cumulative utility of all objectives, as expressed by a given prior over a set of scalarization functions. Most existing BO algorithms do not model the fact that the multiple objectives, or equivalently, tasks can share similarities, and even the few that do lack rigorous, finite-time regret guarantees that capture explicitly inter-task structure. In this work, we address this problem by modelling inter-task dependencies using a multi-task kernel and develop two novel BO algorithms based on random scalarization of the objectives. Our algorithms employ vector-valued kernel regression as a stepping stone and belong to the upper confidence bound class of algorithms. Under a smoothness assumption that the unknown vector-valued function is an element of the reproducing kernel Hilbert space associated with the multi-task kernel, we derive worst-case regret bounds for our algorithms that explicitly capture the similarities between tasks. We numerically benchmark our algorithms on both synthetic and real-life MOO problems, and show the advantages offered by learning with multi-task kernels. } }
Endnote
%0 Conference Paper %T No-regret Algorithms for Multi-task Bayesian Optimization %A Sayak Ray Chowdhury %A Aditya Gopalan %B Proceedings of The 24th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2021 %E Arindam Banerjee %E Kenji Fukumizu %F pmlr-v130-chowdhury21c %I PMLR %P 1873--1881 %U https://proceedings.mlr.press/v130/chowdhury21c.html %V 130 %X We consider multi-objective optimization (MOO) of an unknown vector-valued function in the non-parametric Bayesian optimization (BO) setting. Our aim is to maximize the expected cumulative utility of all objectives, as expressed by a given prior over a set of scalarization functions. Most existing BO algorithms do not model the fact that the multiple objectives, or equivalently, tasks can share similarities, and even the few that do lack rigorous, finite-time regret guarantees that capture explicitly inter-task structure. In this work, we address this problem by modelling inter-task dependencies using a multi-task kernel and develop two novel BO algorithms based on random scalarization of the objectives. Our algorithms employ vector-valued kernel regression as a stepping stone and belong to the upper confidence bound class of algorithms. Under a smoothness assumption that the unknown vector-valued function is an element of the reproducing kernel Hilbert space associated with the multi-task kernel, we derive worst-case regret bounds for our algorithms that explicitly capture the similarities between tasks. We numerically benchmark our algorithms on both synthetic and real-life MOO problems, and show the advantages offered by learning with multi-task kernels.
APA
Chowdhury, S.R. & Gopalan, A.. (2021). No-regret Algorithms for Multi-task Bayesian Optimization . Proceedings of The 24th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 130:1873-1881 Available from https://proceedings.mlr.press/v130/chowdhury21c.html.

Related Material