On Interruptible Pure Exploration in Multi-Armed Bandits

Authors: Alexander Shleyfman, Antonín Komenda, Carmel Domshlak

AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our experimental evaluation demonstrates that the corresponding instances of DB favorably compete both with the currently popular strategies UCB1 and ε-Greedy, as well as with the conservative uniform sampling.
Researcher Affiliation Academia Alexander Shleyfman Technion Israel Institute of Technology, Haifa, Israel Anton ın Komenda Dept. of Computer Science, Faculty of Electrical Engineering, Czech Technical University, Prague, Czech Republic Carmel Domshlak Technion Israel Institute of Technology, Haifa, Israel
Pseudocode Yes Figure 1: The Discriminative Bucketing sampling scheme
Open Source Code No The paper does not provide any links to open-source code or state that the code is available.
Open Datasets No The paper describes how experimental instances were generated (e.g., 'For i [K], the action ai is set to a Bernoulli random variable Ber(pi) with parameter pi U[0, 1]') rather than using pre-existing public datasets with access information.
Dataset Splits No The paper mentions repeating experiments but does not specify training, validation, or test dataset splits.
Hardware Specification No The paper does not specify any hardware used for the experiments.
Software Dependencies No The paper mentions 'ε-Greedy with ε = 1/2' as an experimental setting, but does not list any specific software dependencies or versions.
Experiment Setup Yes Following Tolpin and Shimony (2012), we used ε-Greedy with ε = 1/2.