Chowdhury, SR and Gopalan, A (2017) On kernelized multi-armed bandits. In: 34th International Conference on Machine Learning, ICML 2017, 6 - 11 August 2017, Sydney, pp. 1397-1422.
PDF
ICML_2017.pdf - Published Version Restricted to Registered users only Download (570kB) | Request a copy |
Abstract
We consider the stochastic bandit problem with a continuous set of arms, with the expected reward function over the arms assumed to be fixed but unknown. We provide two new Gaussian process-based algorithms for continuous bandit optimization-Improved GP-UCB (IGP-UCB) and GP-Thomson sampling (GP-TS), and derive corresponding regret bounds. Specifically, the bounds hold when the expected reward function belongs to the reproducing kernel Hilbert space (RKHS) that naturally corresponds to a Gaussian process kernel used as input by the algorithms. Along the way, we derive a new self-normalized concentration inequality for vector-valued martingales of arbitrary, possibly infinite, dimension. Finally, experimental evaluation and comparisons to existing algorithms on synthetic and real-world environments are carried out that highlight the favorable gains of the proposed strategies in many cases.
Item Type: | Conference Paper |
---|---|
Publication: | 34th International Conference on Machine Learning, ICML 2017 |
Publisher: | International Machine Learning Society (IMLS) |
Additional Information: | The copyright for this article belongs to the International Machine Learning Society (IMLS). |
Keywords: | Artificial intelligence; Gaussian distribution; Gaussian noise (electronic); Learning algorithms; Stochastic systems, Bandit problems; Concentration inequality; Experimental evaluation; Gaussian Processes; Multi armed bandit; Real world environments; Reproducing Kernel Hilbert spaces; Reward function, Learning systems |
Department/Centre: | Division of Electrical Sciences > Electrical Communication Engineering |
Date Deposited: | 04 Aug 2022 09:55 |
Last Modified: | 04 Aug 2022 09:55 |
URI: | https://eprints.iisc.ac.in/id/eprint/74680 |
Actions (login required)
View Item |