Balancing Fine-tuning and RAG: A Hybrid Strategy for Dynamic LLM Recommendation Updates

TL;DR

This paper proposes a hybrid update strategy combining fine-tuning and RAG for LLM-based recommendation systems, improving adaptability, cost-efficiency, and user satisfaction in dynamic environments.

Contribution

It introduces a novel hybrid update method that balances fine-tuning and RAG, validated through large-scale experiments for dynamic recommendation systems.

Findings

Hybrid approach outperforms pure fine-tuning or RAG in user satisfaction.

Hybrid method reduces update costs while maintaining knowledge relevance.

Live experiments show significant improvements in recommendation quality.

Abstract

Large Language Models (LLMs) empower recommendation systems through their advanced reasoning and planning capabilities. However, the dynamic nature of user interests and content poses a significant challenge: While initial fine-tuning aligns LLMs with domain knowledge and user preferences, it fails to capture such real-time changes, necessitating robust update mechanisms. This paper investigates strategies for updating LLM-powered recommenders, focusing on the trade-offs between ongoing fine-tuning and Retrieval-Augmented Generation (RAG). Using an LLM-powered user interest exploration system as a case study, we perform a comparative analysis of these methods across dimensions like cost, agility, and knowledge incorporation. We propose a hybrid update strategy that leverages the long-term knowledge adaptation of periodic fine-tuning with the agility of low-cost RAG. We demonstrate through live A/B experiments on a billion-user platform that this hybrid approach yields statistically significant improvements in user satisfaction, offering a practical and cost-effective framework for maintaining high-quality LLM-powered recommender systems.