Literature Survey of Federated Learning Approaches for Privacy Preserving Data Analytics
Abstract
Two signs of the massive developments in AI are the rise of large language models (LLMs) and foundation models. However, the training of these models usually depends on centralized aggregation of massive, usually sensitive datasets, creating major risks to privacy, challenges to regulatory compliance, like, GDPR, HIPAA, and vulnerability to data breaches or inference attacks. Federated learning provides an effective decentralized method that allows collaborative model training without raw data being shared, however, yet its application to large-scale foundation models and LLMs is explored only in a limited manner in terms of systematic synthesis of approaches, performance trade-offs, and practical feasibility. This study addresses the need for a comprehensive overview by conducting a systematic literature survey to critically analyze federated learning approaches for privacy-preserving analytics in foundation models and LLMs. The significance of the study lies in bridging the gap between privacy requirements and the performance demands of modern AI, allowing secure, collaborative development of powerful models in privacy-sensitive domains like healthcare, finance, and edge computing. A qualitative research design was employed, using purposive sampling to select 32 high-impact since-2020 studies from major repositories, like, Google Scholar. Thematic analysis was applied to synthesize methods, strengths, limitations, and trends from the data obtained from the existing…
Read the paper
Related studies
- Active Learning during Federated Learning for Object Detection
- Fairness-aware federated learning
- Decentralized Federated Learning
- Building a Personally Identifiable Information Recognizer in a Privacy Preserved Manner Using Automated Annotation and Federated Learning
- Preservation of Patient Level Privacy: Federated Classification and Calibration Models
- Efficient explainable federated learning
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.