Sequential Experimental Design for Transductive Linear Bandits

Authors: Tanner Fiez, Lalit Jain, Kevin G. Jamieson, Lillian Ratliff

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we present simulations for the linear bandit pure exploration problem and the general transductive bandit problem. We compare our proposed algorithm with both adaptive and nonadaptive strategies.
Researcher Affiliation Academia Tanner Fiez Electrical & Computer Engineering University of Washington ... Lalit Jain Allen School of Computer Science & Engineering University of Washington ... Kevin Jamieson Allen School of Computer Science & Engineering University of Washington ... Lillian Ratliff Electrical & Computer Engineering University of Washington
Pseudocode Yes Algorithm 1: RAGE(X, Z, , r( ), δ): Randomized Adaptive Gap Elimination
Open Source Code No The paper does not provide an explicit statement about releasing the source code or a link to a code repository for the described methodology.
Open Datasets Yes To conduct an experiment based on real data, we build a problem using the Yahoo! Webscope Dataset R6A.5 ... 5https://webscope.sandbox.yahoo.com/
Dataset Splits No The paper describes problem setups and simulations but does not provide specific details on train/validation/test dataset splits for reproducibility.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments or simulations.
Software Dependencies No The paper does not provide specific software dependency names with version numbers required to replicate the experiment.
Experiment Setup Yes We run each algorithm at a confidence level of δ = 0.05. To compute the samples for RAGE, we first used the Frank-Wolfe algorithm (with a precise stopping condition in the supplementary) to find λt, and then the rounding procedure from [27] with = 1/10. ... We also include known parameters 1 = 1 and 2 = 0.5 ... The weights of the parameter vector are drawn from a discrete uniform distribution with a range of [ 0.3, 0.3] and a granularity of 0.01. ... We then fit a regularized least squares estimate using a regularization parameter of 0.01 to obtain .