Randomized Federated Learning Methods for Nonsmooth, Nonconvex, and Hierarchical Optimization (Final Technical Report)

Yousefian, Farzad, Shanbhag, Uday V. · 2025

Read Randomized Federated Learning Methods for Nonsmooth, Nonconvex, and Hierarchical Optimization (Final Technical Report) by Yousefian, Farzad, Shanbhag, Uday V. (2025) 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/doe-osti-3011291-randomized-federated-learning-methods-fo.txt.

Description

Federated learning methods for nonsmooth, nonconvex, and hierarchical problems are developed in the final report for award DE-SC0023303. Farzad Yousefian of Rutgers led the work, with Uday Shanbhag as co-investigator. The project ran from September 2022 to August 2025, counting a no-cost extension year, with total costs of 400,000 dollars.

Three thrusts produced smoothing-enabled zeroth-order methods with provable bounds on communication, iteration, and sample counts. Inexact implicit variants cover the hierarchical cases. The work also gave early non-asymptotic rate analyses for stochastic programs with equilibrium constraints. Doctoral students at three universities trained on the project.

Catalog record

CREATORYousefian, Farzad, Shanbhag, Uday V.
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
LICENSEPublic Domain
RIGHTSPUBLIC DOMAIN
SOURCEDOE OSTI

Source: https://www.osti.gov/biblio/3011291

Browse the full library · About the Tower of Babel