AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

Differentially Private Federated Bayesian Optimization with Distributed Exploration

Zhongxiang Dai, Bryan Kian Hsiang Low, Patrick Jaillet

NeurIPS 2021 · Federated learning and privacy › Differential privacy · approach: Differentially private federated learning · 4 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add distributed exploration to DP-FTSDP-FTS · Synthetic and real-world experimentsconvergenceImproved
add distributed explorationFTS · Synthetic and real-world experimentsconvergenceImproved
use federated information through Thompson samplingstandard TS · Real-world experimentsconvergenceImproved
add local sub-region exploration at initialization (ablation)distributed exploration · Synthetic experimentsperformanceImproved
weight agents more heavily when assigned to a particular sub-region (ablation)distributed exploration · Synthetic experimentsperformanceImproved
increase the number of sub-regions P (ablation)DP-FTS-DE · Synthetic experimentsperformanceImproved
use an overly small clipping threshold S (ablation)DP-FTS-DE · Synthetic experimentsperformanceHurt
use an excessively large clipping threshold S (ablation)DP-FTS-DE · Synthetic experimentsperformanceHurt
increase the subsampling probability q (ablation)DP-FTS-DE · Synthetic and real-world experimentsutilityImproved
increase Gaussian noise multiplier z (ablation)DP-FTS-DE · Synthetic and real-world experimentsutilityHurt
use a smaller privacy loss by increasing z or decreasing q (ablation)DP-FTS-DE · Real-world experimentsutilityHurt
make the influence of other agents decay faster (ablation)FTS-DE · Synthetic experiments with heterogeneous agentsperformanceImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy