AI Paper Atlas

Research map / Optimization / Distributed and compressed

Variance Reduced Median-of-Means Estimator for Byzantine-Robust Distributed Inference

Jiyuan Tu, Weidong Liu, Xiaojun Mao, Xi Chen

JMLR 2021 · Optimization › Distributed and compressed · approach: Byzantine-robust learning · 14 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
replace median with variance reduced median-of-means (VRMOM) aggregationmean estimator · Gaussian mean estimation p=30, no Byzantine machinesRMSE: 0.0211 -> 0.0175Improved
replace median with VRMOM aggregationmean estimator · Gaussian mean estimation p=1, no Byzantine machinesRMSE: 0.0030 -> 0.0025Improved
replace median with VRMOM aggregation under 15% Byzantine machinesmean estimator · Gaussian mean estimation p=30, Byzantine fraction 0.15RMSE: 0.0249 -> 0.0227Improved
increase number of quantile levels K from 10 to 100 (ablation)VRMOM · Gaussian mean estimation p=30, no Byzantine machinesRMSE: 0.0175 -> 0.0174No clear effect
use VRMOM as gradient aggregator in robust CSL instead of MOMRCSL · Linear regression p=30, no attackRMSE: 0.0319 -> 0.0231Improved
use VRMOM as gradient aggregator under Gaussian attack alpha=0.15RCSL · Linear regression p=30, Gaussian attackRMSE: 0.0398 -> 0.0351Improved
use VRMOM as gradient aggregator under omniscient attack alpha=0.15RCSL · Linear regression p=30, omniscient attackRMSE: 0.0449 -> 0.0396Improved
use VRMOM as gradient aggregator under bit-flip attack alpha=0.15RCSL · Linear regression p=30, bit-flip attackRMSE: 0.0343 -> 0.0250Improved
use VRMOM as gradient aggregator in balanced logistic regressionRCSL · Logistic regression p=30, mu_x=0, no ByzantineRMSE: 0.0699 -> 0.0504Improved
use VRMOM as gradient aggregator in imbalanced logistic regressionRCSL · Logistic regression p=30, mu_x=0.5, no ByzantineRMSE: 0.0830 -> 0.0583Improved
fix iteration count at T=5 instead of T=10 (ablation)RCSL · Linear regression p=30, no attackRMSE: 0.0231 -> 0.0231No clear effect
fix iteration count at T=5 instead of T=10 (ablation)RCSL · Logistic regression p=30, mu_x=0, alpha=0.15RMSE: 0.0702 -> 0.0701No clear effect

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