Advancements and Challenges in Federated Learning : A Survey

Machine Learning (ML) has significantly impacted daily life by automating tasks and enhancing decision-making across diverse sectors such as healthcare, finance, and transportation. However, concerns regarding data privacy, particularly in sensitive domains like healthcare and finance, have impeded its widespread adoption. Federated Learning (FL), pioneered by Google in 2016, presents a promising solution by enabling devices to collaborate on model training without sharing raw data. In FL, each device retains its data locally, and only model updates are exchanged with a central server for aggregation. This decentralized approach preserves data privacy while still improving model performance through collaborative learning. FL has garnered interest across industries, with approximately 32% of companies intending to integrate it into their systems soon. Furthermore, investment in FL is projected to rise substantially from $107 million in 2020 to $538 million by 2025, as indicated by forecasts from KPMG. This paper provides an overview of FL, its applications across various sectors, current adoption trends, and future growth prospects, highlighting its significance in addressing data privacy concerns while advancing machine learning capabilities.