A solvable model using disordered systems theory shows interactions among students in unsupervised federated learning enhance pattern recovery, with optimal Bayesian conditions derived as functions of sample complexity, noise, and interaction strength.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.dis-nn 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
A solvable model for unsupervised federated learning
A solvable model using disordered systems theory shows interactions among students in unsupervised federated learning enhance pattern recovery, with optimal Bayesian conditions derived as functions of sample complexity, noise, and interaction strength.