An Experimental Design Approach for Regret Minimization in Logistic Bandits

Authors: Blake Mason, Kwang-Sung Jun, Lalit Jain7736-7743

AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical Evaluation To verify the performance of WAR numerically, we have drawn 20 arms from the a three-dimensional unit sphere. The unknown θ was drawn the same way but scaled to have the norm S {2,4,8}. We have run the naive warmup (5), WAR (2), and the oracle warmup that solves g = minλ X maxx X x 2 Hλ(θ ) 1. We then computed the total number of samples required to satisfy the warmup condition (2) from each method, ignoring the integer effect for simplicity. We repeat this process 5 times and report the result in Table 1 where WAR is significantly better than the naive warmup and not far from the oracle warmup.
Researcher Affiliation Academia Blake Mason1, Kwang-Sung Jun2, and Lalit Jain3 1Rice University 2University of Arizona 3University of Washington
Pseudocode Yes Algorithm 1: HOMER: H Optimal MEthod for Regret
Open Source Code No The paper does not provide an explicit statement or a link to open-source code for the methodology described.
Open Datasets No The paper describes a synthetic data generation process for its numerical evaluation ('we have drawn 20 arms from the a three-dimensional unit sphere') but does not refer to or provide access information for a publicly available or open dataset.
Dataset Splits No The paper describes a numerical evaluation using synthetically generated data and does not specify training, validation, or test dataset splits in terms of percentages or sample counts.
Hardware Specification No The paper does not provide specific details about the hardware (e.g., GPU/CPU models, memory) used to run the numerical evaluations.
Software Dependencies No The paper does not specify any software dependencies with version numbers (e.g., programming languages, libraries, frameworks, or solvers) used for implementation or experiments.
Experiment Setup Yes Numerical Evaluation To verify the performance of WAR numerically, we have drawn 20 arms from the a three-dimensional unit sphere. The unknown θ was drawn the same way but scaled to have the norm S {2,4,8}. We have run the naive warmup (5), WAR (2), and the oracle warmup that solves g = minλ X maxx X x 2 Hλ(θ ) 1. We then computed the total number of samples required to satisfy the warmup condition (2) from each method, ignoring the integer effect for simplicity. We repeat this process 5 times and report the result in Table 1 where WAR is significantly better than the naive warmup and not far from the oracle warmup.