Concept-aware clustering for decentralized deep learning under temporal shift

TL;DR

This paper introduces a novel decentralized deep learning algorithm that automatically detects and adapts to evolving concepts in non-iid, temporally shifting data across clients, outperforming existing methods.

Contribution

It presents the first approach to handle decentralized learning with non-iid and dynamic data without prior knowledge of concept changes.

Findings

Outperforms previous decentralized learning methods

Effectively detects and adapts to concept shifts

Works well on standard benchmark datasets

Abstract

Decentralized deep learning requires dealing with non-iid data across clients, which may also change over time due to temporal shifts. While non-iid data has been extensively studied in distributed settings, temporal shifts have received no attention. To the best of our knowledge, we are first with tackling the novel and challenging problem of decentralized learning with non-iid and dynamic data. We propose a novel algorithm that can automatically discover and adapt to the evolving concepts in the network, without any prior knowledge or estimation of the number of concepts. We evaluate our algorithm on standard benchmark datasets and demonstrate that it outperforms previous methods for decentralized learning.