Differentially Private Two-Stage Empirical Risk Minimization and Applications to Individualized Treatment Rule
Joowon Lee, Guanhua Chen · 2026
Read Differentially Private Two-Stage Empirical Risk Minimization and Applications to Individualized Treatment Rule by Joowon Lee, Guanhua Chen (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-2602-12604.txt.
Description
Differential privacy limits how much any single point can sway a stats output, by adding tuned random noise. Joowon Lee led this work with one co-author, Guanhua Chen, studying private guessing in two-step procedures common in cause reading and personal care rule learning. Data-driven weights first force factor balance, then a value of focus follows from weighted sample fitting.
Their method, DP-2ERM, makes the final tool private directly through goal noise tuned to data-driven response of the full line. Analysis joins fixed weight-noise bounds for flipped tendency, spread balancing, and max gap weighting, with chance response bounds for the second-step answer. The final tuning is tighter than step-wise mixing, and tests plus a standard care-rule use show an improved privacy-use trade-off.
Catalog record
| CREATOR | Joowon Lee, Guanhua Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 93 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-12604