Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2605.19813 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918511623798784 |
|---|---|
| author | Li, Yicheng |
| author_facet | Li, Yicheng |
| contents | We prove a general lower bound for differentially private federated learning protocols with arbitrary public-transcript interactions. The protocol may use any number of adaptive rounds, and each client's local samples may be reused across rounds. For parameter estimation under squared \(\ell_2\) loss, we establish a federated van Trees lower bound for every estimator satisfying a total clientwise sample-level zero-concentrated differential privacy (zCDP) constraint. The main technical ingredient is a privacy-information contraction inequality for complete public transcripts. We illustrate the bound through applications to mean estimation, linear regression, and nonparametric regression. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_19813 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | General Lower Bounds for Differentially Private Federated Learning with Arbitrary Public-Transcript Interactions Li, Yicheng Machine Learning Statistics Theory We prove a general lower bound for differentially private federated learning protocols with arbitrary public-transcript interactions. The protocol may use any number of adaptive rounds, and each client's local samples may be reused across rounds. For parameter estimation under squared \(\ell_2\) loss, we establish a federated van Trees lower bound for every estimator satisfying a total clientwise sample-level zero-concentrated differential privacy (zCDP) constraint. The main technical ingredient is a privacy-information contraction inequality for complete public transcripts. We illustrate the bound through applications to mean estimation, linear regression, and nonparametric regression. |
| title | General Lower Bounds for Differentially Private Federated Learning with Arbitrary Public-Transcript Interactions |
| topic | Machine Learning Statistics Theory |
| url | https://arxiv.org/abs/2605.19813 |