Distributed Differential Privacy in Multi-Armed Bandits

Authors: Sayak Ray Chowdhury, Xingyu Zhou

ICLR 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our theoretical findings are corroborated by numerical evaluations on both synthetic and real-world data. and 7 SIMULATION RESULTS We empirically evaluate the regret performance of our successive elimination scheme with Sec Agg protocol (Algorithm 1) under distributed trust model, which we abbreviate as Dist-DP-SE and Dist RDP-SE when the randomizer R is instantiated with P olya noise (for pure DP) and Skellam noise (for RDP), respectively. We compare them with the DP-SE algorithm of Sajed & Sheffet (2019) that achieves optimal regret under pure DP in the central model, but works only with continuous Laplace noise. We fix confidence level p = 0.1 and study comparative performances under varying privacy levels (ε < 1 for synthetic data, ε 1 for real data). We plot time-average regret Reg(T)/T in Figure 1 by averaging results over 20 randomly generated bandit instances.
Researcher Affiliation Collaboration Sayak Ray Chowdhury Microsoft Research Bengaluru, Karnataka, India t-sayakr@microsoft.com Xingyu Zhou Department of Electrical and Computer Engineering Wayne State University Detroit, USA xingyu.zhou@wayne.edu
Pseudocode Yes Algorithm 1 Private Batch-Based Successive Arm Elimination
Open Source Code Yes Code is available at https://github.com/sayakrc/Differentially-Private-Bandits.
Open Datasets Yes In the bottom panel, we generate bandit instances from Microsoft Learning to Rank dataset MSLR-WEB10K (Qin & Liu, 2013).
Dataset Splits No No explicit information regarding training, validation, and test dataset splits with percentages or sample counts was found. The paper mentions generating bandit instances and averaging results over 20 random instances, but not specific data partitioning for model training or validation.
Hardware Specification No No specific hardware details (such as GPU/CPU models, memory, or cloud computing instances) used for the experiments are mentioned in the paper.
Software Dependencies No No specific software dependencies with version numbers (e.g., programming language versions, library versions, or specific solver versions) are explicitly mentioned in the paper.
Experiment Setup Yes We fix confidence level p = 0.1 and study comparative performances under varying privacy levels (ε < 1 for synthetic data, ε 1 for real data). and Dist-RDP-SE(s=10) Dist-RDP-SE(s=100) in Figure 1 legends and text.