AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

A Unified Solution for Privacy and Communication Efficiency in Vertical Federated Learning

Ganyu Wang, Bin Gu, Qingsong Zhang, Xiang Li, Boyu Wang, Charles Ling

NeurIPS 2023 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Vertical federated learning · 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
replace first-order backward messages with ZOO (AvgRand-100) (ablation)VFL-CZOFO · dist-MNISTtest accuracy: 97.36 -> 95.30Hurt
replace first-order backward messages with ZOO (AvgRand-100) (ablation)VFL-CZOFO · dist-MNISTbackward cost: 3073 MB -> 75 MBImproved
decrease ZOO sampling count q from 100 to 10 (ablation)AvgRand ZOO · dist-MNISTtest accuracy: 95.30 -> 93.33Hurt
decrease ZOO sampling count q from 10 to 1 (ablation)AvgRand ZOO · dist-MNISTtest accuracy: 93.33 -> 79.66Hurt
increase sampling count q (ablation)VFL-CZOFO · dist-MNISTconvergence speedImproved
compress forward message to 8 bits (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 95.02No clear effect
compress forward message to 4 bits (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 93.35Hurt
compress forward message to 2 bits (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 79.10Hurt
compress forward message to 1 bit (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 30.20Hurt
compress backward message to 8 bits (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 94.58No clear effect
compress backward message to 2 bits (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 94.16No clear effect
compress backward message to 1 bit (ablation)AvgRand-100 · dist-MNISTtest accuracy: 95.30 -> 90.67Hurt

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