Selective Sampling for Online Best-arm Identification

Authors: Romain Camilleri, Zhihan Xiong, Maryam Fazel, Lalit Jain, Kevin G. Jamieson

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section we present a benchmark experiment validating the fundamental trade-offs that are theoretically characterized in Theorem 1 and Theorem 2.
Researcher Affiliation Academia Romain Camilleri , Zhihan Xiong , Maryam Fazel, Lalit Jain, Kevin Jamieson University of Washington, Seattle, WA {camilr,zhihanx,mfazel,lalitj,jamieson}@uw.edu
Pseudocode Yes Algorithm 1 Selective Sampling for Best-arm Identification
Open Source Code No The paper does not include an unambiguous statement of code release or a link to a code repository for the described methodology.
Open Datasets No The paper describes a synthetic data generation process for its experiments but does not use or provide access information for a publicly available or open dataset.
Dataset Splits No The paper describes an online streaming setting where data is generated IID, rather than using fixed dataset splits (training, validation, test) for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details, such as library names with version numbers, needed to replicate the experiment.
Experiment Setup Yes In this problem, the angle ω is small enough that the item (cos(ω), sin(ω)) is hard to discriminate from the best item e1. ... We swept over the values of τ and plotted on the y-axis the amount of labeled data needed before termination, as shown in Figure 1. ... Our Algorithm uses the solution to (5) for b Pℓ, where we take µb = 2 10 5.