AI Paper Atlas

Research map / Federated learning and privacy / Differential privacy

GlucoSynth: Generating Differentially-Private Synthetic Glucose Traces

Josephine Lamp, Mark Derdzinski, Christopher Hannemann, Joost Van der Linden, Lu Feng, Tianhao Wang, David Evans

NeurIPS 2023 · Federated learning and privacy › Differential privacy · approach: Differentially private synthetic data · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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
GlucoSynth vs TimeGAN (nonprivate)TimeGAN · glucose traces, epsilon=infforecasting RMSE: 0.061 -> 7e{-3}Improved
GlucoSynth vs FF (nonprivate)FF · glucose traces, epsilon=infvalidation motif fraction VM: 0.642 -> 0.987Improved
GlucoSynth vs NVP (nonprivate)NVP · glucose traces, epsilon=infmotif distribution MSE: 1.9 -> 1.6Mixed
GlucoSynth vs dpGAN at epsilon=1dpGAN · glucose traces, epsilon=1forecasting RMSE: 0.030 -> 0.030No clear effect
GlucoSynth vs RGAN at epsilon=1RGAN · glucose traces, epsilon=1forecasting RMSE: 0.651 -> 0.030Improved
decrease privacy budget from infinity to 0.01GlucoSynth · glucose tracesforecasting RMSE: 7e{-3} -> 0.038Hurt
decrease privacy budget from infinity to 0.01GlucoSynth · glucose tracesmotif distribution MSE: 1.6 -> 99.0Hurt
decrease privacy budget from infinity to 0.01GlucoSynth · glucose tracescoverage: 0.534 -> 0.010Hurt
decrease epsilon from 10 to 0.1GlucoSynth · glucose tracesVM: 1.000 -> 1.000No clear effect
use PATE instead of DP-SGD for motif causality blockGlucoSynth · glucose tracesprivacy-utility trade-offImproved
GlucoSynth vs dpGAN coverage at low epsilondpGAN · glucose traces, epsilon=0.01coverage: 0.054 -> 0.010Hurt
GlucoSynth vs dpGAN motif MSE at epsilon=1dpGAN · glucose traces, epsilon=1motif distribution MSE: 23.2 -> 6.7Improved

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