Symmetric Linear Bandits with Hidden Symmetry

Authors: Phuong Nam Tran, The Anh Ta, Debmalya Mandal, Long Tran-Thanh

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental To illustrate the performance of our algorithm, we conduct simulations where the entries of θ satisfy three cases: sparsity, non-crossing partitions and non-nesting partitions.
Researcher Affiliation Academia Nam Phuong Tran Department of Computer Science University of Warwick Coventry, United Kingdom nam.p.tran@warwick.ac.uk The Anh Ta CSIRO s Data61 Marsfield, NSW, Australia theanh.ta@csiro.au Debmalya Mandal Department of Computer Science University of Warwick Coventry, United Kingdom debmalya.mandal@warwick.ac.uk Long Tran-Thanh Department of Computer Science University of Warwick Coventry, United Kingdom long.tran-thanh@warwick.ac.uk
Pseudocode Yes Algorithm 1 Explore Models then Commit
Open Source Code Yes Code is available at: https://github.com/Nam Tran Kek L/Symmetric-Linear-Bandit-with-Hidden-Symmetry.git.
Open Datasets No The paper uses synthetic data generated for its simulations and does not provide access information for a pre-existing public dataset.
Dataset Splits No The paper describes an exploration phase followed by a commitment phase, but does not explicitly mention separate training/validation/test dataset splits or cross-validation.
Hardware Specification No The paper describes its simulations and discusses computational complexity, but does not provide specific details about the hardware (e.g., CPU/GPU models, memory) used to run these experiments.
Software Dependencies No The paper refers to algorithms and techniques (e.g., Lasso regression, OFUL algorithm) but does not list specific software packages or libraries with version numbers required to replicate the experiments.
Experiment Setup Yes The set of arms X is d Sd 1, σ = 0.1, and (d, d0) {(40, 4), (80, 10), (100, 15)}. We let exploratory distribution ν be the uniform distribution on the unit sphere. The ground-truth partition πG and θ are randomized before each simulation.