Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Epsilon Best Arm Identification in Spectral Bandits

Authors: Tomáš Kocák, Aurélien Garivier

IJCAI 2021 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental For the experiments, we used bandit problem µ = (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.4, 0.3, 0.2, 0.1, 0) with K = 11 arms... The following plot demonstrates the effect of smoothness parameter R on both theoretical and empirical stopping times. The green curve represents the average stopping time of 10 runs of Spectral Ta S while the red curve represents the characteristic time.
Researcher Affiliation Academia Tom aˇs Koc ak and Aur elien Garivier Unit e de Math ematiques Pures et Appliqu ees et Laboratoire de l Informatique du Parall elisme Ecole Normale Sup erieure de Lyon, Universit e de Lyon EMAIL, EMAIL
Pseudocode Yes Algorithm 1 Spectral Ta S
Open Source Code No No statement about releasing code or links to a code repository are provided.
Open Datasets No The paper defines a synthetic bandit problem µ = (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.4, 0.3, 0.2, 0.1, 0) with K=11 arms for its experiments, which is generated for the paper rather than being a pre-existing public dataset with specific access information.
Dataset Splits No The paper describes a simulation setup for a bandit problem and averages results over "10 runs" but does not specify explicit training, validation, or test dataset splits.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper does not specify any software names with version numbers.
Experiment Setup Yes For the experiments, we used bandit problem µ = (0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.4, 0.3, 0.2, 0.1, 0) with K = 11 arms... ε = 0.05, and different values of R. Algorithm 1 Spectral Ta S Input and initialization: L : graph Laplacian ε, δ : tolerance and confidence parameters R : upper bound on the smoothness of µ