Federated $\mathcal{X}$-armed Bandit with Flexible Personalisation

Authors: Ali Arabzadeh, James A. Grant, David S. Leslie

TMLR 2024 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Theoretical analysis and empirical evaluations demonstrate the effectiveness of our approach compared to existing methods. Potential applications of this work span various domains, including healthcare, smart home devices, and e-commerce, where balancing personalisation with global insights is crucial. ... In this section we evaluate the empirical performance of PF-XAB through a series of experiments employing synthetic objective functions and real-world dataset.
Researcher Affiliation Academia Ali Arabzadeh EMAIL School of Mathematical Sciences Lancaster University James A. Grant EMAIL School of Mathematical Sciences Lancaster University David S. Leslie EMAIL School of Mathematical Sciences Lancaster University
Pseudocode Yes Algorithm 1 Fed-XAB: server ... Algorithm 2 PF-XAB: m-th client
Open Source Code No No explicit statement or link to source code release for the methodology described in this paper was found.
Open Datasets Yes In addition to our initial experiments with synthetic objective functions, we extended our experiments to a real-world-inspired scenario using the landmine dataset (Liu et al., 2007).
Dataset Splits No The paper does not explicitly provide details about training/test/validation dataset splits. It mentions using the 'landmine dataset' and 'synthetic objective functions' but no specific split percentages or methodologies.
Hardware Specification No No specific hardware details (e.g., GPU models, CPU models, memory amounts) used for running the experiments are mentioned in the paper.
Software Dependencies No The paper does not provide specific software dependencies with version numbers. It mentions using a 'Support Vector Machine (SVM) classifier equipped with RBF kernels' but no software versions.
Experiment Setup Yes Throughout this section, all experiments has been conducted on a federation with M = 10 clients. ... with M = 10 machines over T = 2 106 Time Steps. ... Our experiments explore d = 2 kernel parameters, γ ranging in [0.01, 10], and the regularisation parameter C, selected from the range [10 4, 104], forms the domain space of our experiments.