Accelerated Methods with Complexity Separation Under Data Similarity for Federated Learning Problems
This work addresses communication efficiency in federated learning, but it appears incremental as it builds on existing methods under data similarity assumptions.
The paper tackled the challenge of data heterogeneity in federated learning by formalizing it as an optimization problem and proposing an optimal algorithm for convex cases, with validation through experiments.
Heterogeneity within data distribution poses a challenge in many modern federated learning tasks. We formalize it as an optimization problem involving a computationally heavy composite under data similarity. By employing different sets of assumptions, we present several approaches to develop communication-efficient methods. An optimal algorithm is proposed for the convex case. The constructed theory is validated through a series of experiments across various problems.