Adversary-Robust Learning from Fully Asynchronous Directional Derivative Estimates
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913107903774720 |
|---|---|
| author | Paul, Anik Kumar Roy, Nibedita Talagani, Nagesh Ganesh, Swetha Thoppe, Gugan Reiffers-Masson, Alexandre |
| author_facet | Paul, Anik Kumar Roy, Nibedita Talagani, Nagesh Ganesh, Swetha Thoppe, Gugan Reiffers-Masson, Alexandre |
| contents | We propose FAR-SIGN (Fully Asynchronous Robust optimization via SIGNed directional projections) for adversary-resilient learning in parameter-server--worker systems. FAR-SIGN achieves robustness through sign-based updates along carefully designed directions and mitigates the resulting bias via a two-timescale mechanism. It admits both first-order and zeroth-order implementations and enables fully asynchronous execution without requiring a private reference dataset at the server. We establish almost-sure convergence of FAR-SIGN to the set of stationary points for smooth, nonconvex objectives. Moreover, we prove the near-optimal rate of $O(n^{-1/4+ε})$ in the first-order setting and the standard $O(n^{-1/6+ε})$ in the zeroth-order setting, where $n$ is the iteration count and $ε>0$ can be chosen arbitrarily small. Experiments on MNIST show that FAR-SIGN outperforms robust aggregation-based methods in both accuracy and wall-clock time. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_09337 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Adversary-Robust Learning from Fully Asynchronous Directional Derivative Estimates Paul, Anik Kumar Roy, Nibedita Talagani, Nagesh Ganesh, Swetha Thoppe, Gugan Reiffers-Masson, Alexandre Machine Learning Optimization and Control We propose FAR-SIGN (Fully Asynchronous Robust optimization via SIGNed directional projections) for adversary-resilient learning in parameter-server--worker systems. FAR-SIGN achieves robustness through sign-based updates along carefully designed directions and mitigates the resulting bias via a two-timescale mechanism. It admits both first-order and zeroth-order implementations and enables fully asynchronous execution without requiring a private reference dataset at the server. We establish almost-sure convergence of FAR-SIGN to the set of stationary points for smooth, nonconvex objectives. Moreover, we prove the near-optimal rate of $O(n^{-1/4+ε})$ in the first-order setting and the standard $O(n^{-1/6+ε})$ in the zeroth-order setting, where $n$ is the iteration count and $ε>0$ can be chosen arbitrarily small. Experiments on MNIST show that FAR-SIGN outperforms robust aggregation-based methods in both accuracy and wall-clock time. |
| title | Adversary-Robust Learning from Fully Asynchronous Directional Derivative Estimates |
| topic | Machine Learning Optimization and Control |
| url | https://arxiv.org/abs/2605.09337 |