Д. О. Медяков
1,2 · 2024
Read Д. О. Медяков by 1,2 (2024) 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-2412-14935.txt.
Description
Daniil Medyakov, Gleb Molodtsov, and Aleksandr Beznosikov present, in a 2024 paper, a method for solving variational inequalities when the data are split across many machines. Such problems cover tasks from training generative models to game theory. Sharing data between machines is costly, so squeezing what is sent, called compression, is a prized trick.
Their method adapts advanced tools first built for plain minimization to the inequality setting, where each machine sends compressed updates. They give a full proof of convergence for a class of well-behaved, strongly monotone problems. Tests back the theory and show strong practical speed. The work matters for federated learning, where phones or hospitals must train together without shipping raw data.
Catalog record
| CREATOR | 1,2 |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 47 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2412-14935