Decentralized Exploration in Multi-Armed Bandits

Authors: Raphael Feraud, Reda Alami, Romain Laroche

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

Reproducibility Variable Result LLM Response
Research Type Experimental To illustrate and complete the analysis of DECENTRALIZED ELIMINATION, we run three synthetic experiments: (...) The algorithms are compared with respect to two key performance indicators: the sample complexity and the communication cost. For all the experiments, ϵ is set to 0.25, and δ is set to 0.05. The privacy level η is set to 0.9. All the curves and the measures are averaged over 20 trials.
Researcher Affiliation Industry 1Orange Labs 2Microsoft Research.
Pseudocode Yes Algorithm 1 DECENTRALIZED EXPLORATION PROBLEM (...) Algorithm 2 DECENTRALIZED ELIMINATION
Open Source Code No The paper does not provide any concrete access information (e.g., repository link, explicit statement of code release) for the source code of the described methodology.
Open Datasets No The paper uses synthetic experiments ('Problem 1: Uniform distribution of players', 'Problem 2: 50% of players generates 80% of events', 'Problem 3: non-stationary rewards') rather than external public datasets, and provides no access information or citations for any dataset.
Dataset Splits No The paper mentions running experiments over '20 trials' but does not specify any training, validation, or test dataset splits or a methodology for data partitioning.
Hardware Specification No The paper does not provide any specific details regarding the hardware specifications (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper mentions using algorithms like UGAPEC and SER3, but does not provide specific version numbers for any software dependencies or libraries.
Experiment Setup Yes For all the experiments, ϵ is set to 0.25, and δ is set to 0.05. The privacy level η is set to 0.9.