Towards Artificial General or Personalized Intelligence? A Survey on Foundation Models for Personalized Federated Intelligence

2025-05-11

AI Summary

To address core challenges in large language model (LLM) personalization—including privacy leakage, high edge-computing overhead, and inadequate user preference modeling—this paper proposes *Personalized Federated Intelligence* (PFI), a foundational enabling technology for Artificial Personalized Intelligence (API). Methodologically, PFI systematically integrates LLM zero-shot generalization with privacy-preserving federated learning mechanisms. We introduce three novel paradigms: (i) Efficient PFI, combining lightweight fine-tuning with on-device adaptation; (ii) Trustworthy PFI, featuring privacy-enhanced robust collaborative training; and (iii) RAG-Augmented PFI, leveraging local retrieval to empower personalized generation. Our work establishes the first comprehensive theoretical framework for PFI, identifies key enablers and bottlenecks in foundation-model–augmented federated learning, and delivers the first systematic technical blueprint and practical implementation paradigm for privacy-sensitive, edge-deployed personalized AI.

Technology Category

• Machine Learning: Privacy
• Natural Language Processing: (Large) Language Models
• Humans and AI: Intelligent User Interfaces

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

The rise of large language models (LLMs), such as ChatGPT, DeepSeek, and Grok-3, has reshaped the artificial intelligence landscape. As prominent examples of foundational models (FMs) built on LLMs, these models exhibit remarkable capabilities in generating human-like content, bringing us closer to achieving artificial general intelligence (AGI). However, their large-scale nature, sensitivity to privacy concerns, and substantial computational demands present significant challenges to personalized customization for end users. To bridge this gap, this paper presents the vision of artificial personalized intelligence (API), focusing on adapting these powerful models to meet the specific needs and preferences of users while maintaining privacy and efficiency. Specifically, this paper proposes personalized federated intelligence (PFI), which integrates the privacy-preserving advantages of federated learning (FL) with the zero-shot generalization capabilities of FMs, enabling personalized, efficient, and privacy-protective deployment at the edge. We first review recent advances in both FL and FMs, and discuss the potential of leveraging FMs to enhance federated systems. We then present the key motivations behind realizing PFI and explore promising opportunities in this space, including efficient PFI, trustworthy PFI, and PFI empowered by retrieval-augmented generation (RAG). Finally, we outline key challenges and future research directions for deploying FM-powered FL systems at the edge with improved personalization, computational efficiency, and privacy guarantees. Overall, this survey aims to lay the groundwork for the development of API as a complement to AGI, with a particular focus on PFI as a key enabling technique.

Problem

• Bridging AGI and personalized needs via foundation models
• Addressing privacy and efficiency in model customization
• Integrating federated learning with foundation models for edge deployment

Innovation

• Integrates federated learning with foundation models
• Enables privacy-protected personalized edge deployment
• Leverages retrieval-augmented generation for efficiency