Sparse Dueling Bandits

Kevin Jamieson, Sumeet Katariya, Atul Deshpande, Robert Nowak
Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, PMLR 38:416-424, 2015.

Abstract

The dueling bandit problem is a variation of the classical multi-armed bandit in which the allowable actions are noisy comparisons between pairs of arms. This paper focuses on a new approach for finding the best arm according to the Borda criterion using noisy comparisons. We prove that in the absence of structural assumptions, the sample complexity of this problem is proportional to the sum of the inverse gaps squared of the Borda scores of each arm. We explore this dependence further and consider structural constraints on the pairwise comparison matrix (a particular form of sparsity natural to this problem) that can significantly reduce the sample complexity. This motivates a new algorithm called Successive Elimination with Comparison Sparsity (SECS) that exploits sparsity to find the Borda winner using fewer samples than standard algorithms. We also evaluate the new algorithm experimentally with synthetic and real data. The results show that the sparsity model and the new algorithm can provide significant improvements over standard approaches.

Cite this Paper


BibTeX
@InProceedings{pmlr-v38-jamieson15, title = {{Sparse Dueling Bandits}}, author = {Jamieson, Kevin and Katariya, Sumeet and Deshpande, Atul and Nowak, Robert}, booktitle = {Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics}, pages = {416--424}, year = {2015}, editor = {Lebanon, Guy and Vishwanathan, S. V. N.}, volume = {38}, series = {Proceedings of Machine Learning Research}, address = {San Diego, California, USA}, month = {09--12 May}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v38/jamieson15.pdf}, url = {https://proceedings.mlr.press/v38/jamieson15.html}, abstract = {The dueling bandit problem is a variation of the classical multi-armed bandit in which the allowable actions are noisy comparisons between pairs of arms. This paper focuses on a new approach for finding the best arm according to the Borda criterion using noisy comparisons. We prove that in the absence of structural assumptions, the sample complexity of this problem is proportional to the sum of the inverse gaps squared of the Borda scores of each arm. We explore this dependence further and consider structural constraints on the pairwise comparison matrix (a particular form of sparsity natural to this problem) that can significantly reduce the sample complexity. This motivates a new algorithm called Successive Elimination with Comparison Sparsity (SECS) that exploits sparsity to find the Borda winner using fewer samples than standard algorithms. We also evaluate the new algorithm experimentally with synthetic and real data. The results show that the sparsity model and the new algorithm can provide significant improvements over standard approaches.} }
Endnote
%0 Conference Paper %T Sparse Dueling Bandits %A Kevin Jamieson %A Sumeet Katariya %A Atul Deshpande %A Robert Nowak %B Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2015 %E Guy Lebanon %E S. V. N. Vishwanathan %F pmlr-v38-jamieson15 %I PMLR %P 416--424 %U https://proceedings.mlr.press/v38/jamieson15.html %V 38 %X The dueling bandit problem is a variation of the classical multi-armed bandit in which the allowable actions are noisy comparisons between pairs of arms. This paper focuses on a new approach for finding the best arm according to the Borda criterion using noisy comparisons. We prove that in the absence of structural assumptions, the sample complexity of this problem is proportional to the sum of the inverse gaps squared of the Borda scores of each arm. We explore this dependence further and consider structural constraints on the pairwise comparison matrix (a particular form of sparsity natural to this problem) that can significantly reduce the sample complexity. This motivates a new algorithm called Successive Elimination with Comparison Sparsity (SECS) that exploits sparsity to find the Borda winner using fewer samples than standard algorithms. We also evaluate the new algorithm experimentally with synthetic and real data. The results show that the sparsity model and the new algorithm can provide significant improvements over standard approaches.
RIS
TY - CPAPER TI - Sparse Dueling Bandits AU - Kevin Jamieson AU - Sumeet Katariya AU - Atul Deshpande AU - Robert Nowak BT - Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics DA - 2015/02/21 ED - Guy Lebanon ED - S. V. N. Vishwanathan ID - pmlr-v38-jamieson15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 38 SP - 416 EP - 424 L1 - http://proceedings.mlr.press/v38/jamieson15.pdf UR - https://proceedings.mlr.press/v38/jamieson15.html AB - The dueling bandit problem is a variation of the classical multi-armed bandit in which the allowable actions are noisy comparisons between pairs of arms. This paper focuses on a new approach for finding the best arm according to the Borda criterion using noisy comparisons. We prove that in the absence of structural assumptions, the sample complexity of this problem is proportional to the sum of the inverse gaps squared of the Borda scores of each arm. We explore this dependence further and consider structural constraints on the pairwise comparison matrix (a particular form of sparsity natural to this problem) that can significantly reduce the sample complexity. This motivates a new algorithm called Successive Elimination with Comparison Sparsity (SECS) that exploits sparsity to find the Borda winner using fewer samples than standard algorithms. We also evaluate the new algorithm experimentally with synthetic and real data. The results show that the sparsity model and the new algorithm can provide significant improvements over standard approaches. ER -
APA
Jamieson, K., Katariya, S., Deshpande, A. & Nowak, R.. (2015). Sparse Dueling Bandits. Proceedings of the Eighteenth International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 38:416-424 Available from https://proceedings.mlr.press/v38/jamieson15.html.

Related Material