Haebin Seong

Senior Research Scientist, Maum.ai

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 6

1 paper

Friday, December 5, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

FedSVD: Adaptive Orthogonalization for Private Federated Learning with LoRA

#1209 · Seanie Lee, Sangwoo Park, Dong Bok Lee, Dominik Wagner, Haebin Seong, Tobias Bocklet, Juho Lee, Sung Ju Hwang

FedSVD mitigates noise amplification in differentially private federated fine-tuning with LoRA by periodically refactorizing LoRA matrices via SVD, improving stability and performance under DP-SGD.