Federated Learning without Full Labels: A Survey

TL;DR

This survey reviews federated learning approaches that operate effectively without fully labeled data, focusing on semi-supervised, self-supervised, and transfer learning techniques to address privacy and labeling challenges.

Contribution

It provides a comprehensive overview of methods combining federated learning with unlabeled data techniques and summarizes evaluation datasets and future research directions.

Findings

Various semi-supervised, self-supervised, and transfer learning methods are adapted for federated learning.

Evaluation datasets for FL without full labels are summarized.

Future research directions include improving model accuracy and privacy in unlabeled FL settings.

Abstract

Data privacy has become an increasingly important concern in real-world big data applications such as machine learning. To address the problem, federated learning (FL) has been a promising solution to building effective machine learning models from decentralized and private data. Existing federated learning algorithms mainly tackle the supervised learning problem, where data are assumed to be fully labeled. However, in practice, fully labeled data is often hard to obtain, as the participants may not have sufficient domain expertise, or they lack the motivation and tools to label data. Therefore, the problem of federated learning without full labels is important in real-world FL applications. In this paper, we discuss how the problem can be solved with machine learning techniques that leverage unlabeled data. We present a survey of methods that combine FL with semi-supervised learning, self-supervised learning, and transfer learning methods. We also summarize the datasets used to evaluate FL methods without full labels. Finally, we highlight future directions in the context of FL without full labels.