AI Paper Atlas

Research map / Federated learning and privacy / Communication-efficient FL

Federated Gaussian Process: Convergence, Automatic Personalization and Multi-Fidelity Modeling

Xubo Yue, Raed Al Kontar

TPAMI 2024 · Federated learning and privacy › Communication-efficient FL · approach: Local drift correction · 25 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
use FGPR (federated global prior) instead of separate HF-only GPFGPR · CURRINRMSE vs true HF model: 0.301 -> 0.148Improved
use FGPR instead of centralized Deep multi-fidelity modelFGPR · PARKRMSE vs true HF model: 0.013 -> 0.012Improved
use FGPR instead of centralized Deep multi-fidelity modelFGPR · BRANINRMSE vs true HF model: 0.213 -> 0.260Hurt
use FGPR instead of AR multi-fidelity modelFGPR · BoreholeRMSE vs true HF model: 0.615 -> 0.604Improved
increase HF client weight p1 from 0.17 toward 1 (decrease LF weight) (ablation)FGPR · CURRINRMSE: 0.17 to 1Mixed
use FGPR instead of federated polynomial regressionFGPR · NASA C-MAPSS Sensor 2averaged RMSE x10: 6.79 -> 5.45Improved
use FGPR instead of federated neural networkFGPR · NASA C-MAPSS Sensor 7averaged RMSE x10: 6.71 -> 5.76Improved
use FGPR instead of Ditto-personalized neural networkFGPR · SARCOS Output 5averaged RMSE x10: 2.89 -> 2.20Improved
use FGPR instead of Ditto-personalized neural networkFGPR · SARCOS Output 1std of RMSE x10: 1.70 -> 1.84Hurt
start FGPR from a bad initialization (flat-line fit, theta=(1,10,1)) (ablation)FGPR · sin(x) toy, two clientsaveraged RMSENo clear effect
use few data points per device (K=100, N=800) (ablation)FGPR · Homogeneous GP simulationconvergence: N=800, M_k=8Mixed

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