Timezone: »
Building on prior work (Xu et al., 2020), we pro-pose a newer version of the Rank-Search algo-rithm, which aims to solve the Threshold BanditProblem when both duels and pulls are possible.Our algorithm, which requires an additional as-sumption connecting the borda scores of armswith their mean rewards, is able to perform in set-tings in which previous Rank-Search algorithmscannot. We conduct experiments comparing theperformance of the new algorithm with that ofolder versions of Rank-Search in various settings.Finally we prove upper bounds for the total num-ber of duels and pulls required by our proposedalgorithm, which we call Rank Search with Link(RS-L).
Author Information
Keshav Narayan (Carnegie Mellon University)
I'm currently a Master's Student in Machine Learning at Carnegie Mellon University, and previously did my undergraduate degree in Computer Science at CMU as well. I'm joining DRW Cumberland as a Quantitative Trader in July, where I will be applying Machine Learning techniques to create algorithms for trading cryptocurrencies.
Aarti Singh (Carnegie Mellon University)
Related Events (a corresponding poster, oral, or spotlight)
-
2022 : Threshold Bandit Problem with Link Assumption between Pulls and Duels »
Sat. Jul 23rd 06:20 -- 06:40 PM Room
More from the Same Authors
-
2020 : Contributed Talk: Catch Me if I Can: Detecting Strategic Behaviour in Peer Assessment »
Ivan Stelmakh · Nihar Shah · Aarti Singh -
2023 Poster: Weighted Tallying Bandits: Overcoming Intractability via Repeated Exposure Optimality »
Dhruv Malik · Conor Igoe · Yuanzhi Li · Aarti Singh -
2023 Poster: The Virtues of Laziness in Model-based RL: A Unified Objective and Algorithms »
Anirudh Vemula · Yuda Song · Aarti Singh · J. Bagnell · Sanjiban Choudhury -
2018 Poster: Nonparametric Regression with Comparisons: Escaping the Curse of Dimensionality with Ordinal Information »
Yichong Xu · Hariank Muthakana · Sivaraman Balakrishnan · Aarti Singh · Artur Dubrawski -
2018 Oral: Nonparametric Regression with Comparisons: Escaping the Curse of Dimensionality with Ordinal Information »
Yichong Xu · Hariank Muthakana · Sivaraman Balakrishnan · Aarti Singh · Artur Dubrawski -
2017 Poster: Near-Optimal Design of Experiments via Regret Minimization »
Zeyuan Allen-Zhu · Yuanzhi Li · Aarti Singh · Yining Wang -
2017 Talk: Near-Optimal Design of Experiments via Regret Minimization »
Zeyuan Allen-Zhu · Yuanzhi Li · Aarti Singh · Yining Wang -
2017 Poster: Uncorrelation and Evenness: a New Diversity-Promoting Regularizer »
Pengtao Xie · Aarti Singh · Eric Xing -
2017 Talk: Uncorrelation and Evenness: a New Diversity-Promoting Regularizer »
Pengtao Xie · Aarti Singh · Eric Xing