FiBeR : A Differentially Private Optimizer with Filter-Aware Innovation Bias Correction
Duc Dm, Thao Do, Minh Son Hoang, Tran Le Duc Anh, Daeyoung Kim †, Huy Nguyen † · 2026
Read FiBeR : A Differentially Private Optimizer with Filter-Aware Innovation Bias Correction by Duc Dm, Thao Do, Minh Son Hoang, Tran Le Duc Anh, Daeyoung Kim †, Huy Nguyen † (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-03425.txt.
Description
Duc Dm, Thao Do, Minh Son Hoang, Anh Le Duc Tran, Daeyoung Kim, and Huy Nguyen present FiBeR, a differentially private optimizer built for temporally filtered gradients. Filtering privatized gradients cuts variance. But it also changes the noise statistics that adaptive optimizers assume. Standard bias corrections then become miscalibrated.
FiBeR filters in innovation space. It decouples observation geometry from filter gain, for independent tuning. It subtracts a filter-aware noise term, computed in closed form. Across vision and language benchmarks, under equal privacy budgets, it consistently beat state-of-the-art private optimizers. The gains show that optimizer design must match the filtering actually applied.
Catalog record
| CREATOR | Duc Dm, Thao Do, Minh Son Hoang, Tran Le Duc Anh, Daeyoung Kim †, Huy Nguyen † |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 100 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.03425