AI Paper Atlas

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

FLoRG: Federated Fine-tuning with Low-rank Gram Matrices and Procrustes Alignment

Chuiyang Meng, Ming Tang, Vincent W.S. Wong

ICLR 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated prompt and LoRA tuning · 2 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
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · MRPCaccuracy: 83.14 -> 86.54Improved
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · QQPaccuracy: 85.85 -> 88.71Improved
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · MNLIaccuracy: 80.93 -> 87.20Improved
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · QNLIaccuracy: 86.72 -> 89.69Improved
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · WNLIaccuracy: 59.81 -> 65.41Improved
add Procrustes alignment after decomposition (ablation)FLoRG OPT-125M · RTEaccuracy: 64.32 -> 68.77Improved
add Procrustes alignment after decomposition (ablation)FLoRG RoBERTa-large · MRPCaccuracy: 86.50 -> 89.87Improved
add Procrustes alignment after decomposition (ablation)FLoRG RoBERTa-large · MNLIaccuracy: 88.93 -> 91.39Improved
add Procrustes alignment after decomposition (ablation)FLoRG RoBERTa-large · QNLIaccuracy: 88.62 -> 92.48Improved
use semi-orthogonal instead of Kaiming initialization (ablation)FLoRG OPT-125M · MNLIaccuracy: 85.23 -> 87.20Improved
use semi-orthogonal instead of SVD initialization (ablation)FLoRG OPT-125M · MNLIaccuracy: 83.19 -> 87.20Improved
use semi-orthogonal instead of Kaiming initialization (ablation)FLoRG OPT-125M · RTEaccuracy: 69.31 -> 68.77Hurt

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