A Comprehensive Review on Harnessing Large Language Models to Overcome Recommender System Challenges

2025-07-17

arXiv.org

AI Summary

To address inherent limitations in recommender systems—including data sparsity, cold-start challenges, insufficient personalization, and weak semantic understanding—this paper proposes a novel LLM-based recommendation paradigm, positioning large language models as the foundational architecture rather than merely auxiliary components. Methodologically, it integrates prompt-driven retrieval, language-native ranking, retrieval-augmented generation (RAG), and conversational interaction to enable zero-shot and few-shot cross-task generalization; multi-stage candidate generation coupled with external knowledge injection further enhances semantic alignment and interpretability. The key contribution is the first structured, LLM-enhanced recommendation framework, which significantly improves personalization accuracy and semantic comprehension—particularly under cold-start and long-tail conditions. Additionally, the work systematically investigates synergistic optimization pathways balancing accuracy, scalability, and real-time responsiveness.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Recommender Systems
• Natural Language Processing: (Large) Language Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Recommender systems have traditionally followed modular architectures comprising candidate generation, multi-stage ranking, and re-ranking, each trained separately with supervised objectives and hand-engineered features. While effective in many domains, such systems face persistent challenges including sparse and noisy interaction data, cold-start problems, limited personalization depth, and inadequate semantic understanding of user and item content. The recent emergence of Large Language Models (LLMs) offers a new paradigm for addressing these limitations through unified, language-native mechanisms that can generalize across tasks, domains, and modalities. In this paper, we present a comprehensive technical survey of how LLMs can be leveraged to tackle key challenges in modern recommender systems. We examine the use of LLMs for prompt-driven candidate retrieval, language-native ranking, retrieval-augmented generation (RAG), and conversational recommendation, illustrating how these approaches enhance personalization, semantic alignment, and interpretability without requiring extensive task-specific supervision. LLMs further enable zero- and few-shot reasoning, allowing systems to operate effectively in cold-start and long-tail scenarios by leveraging external knowledge and contextual cues. We categorize these emerging LLM-driven architectures and analyze their effectiveness in mitigating core bottlenecks of conventional pipelines. In doing so, we provide a structured framework for understanding the design space of LLM-enhanced recommenders, and outline the trade-offs between accuracy, scalability, and real-time performance. Our goal is to demonstrate that LLMs are not merely auxiliary components but foundational enablers for building more adaptive, semantically rich, and user-centric recommender systems

Problem

• Overcoming sparse data and cold-start issues in recommender systems
• Enhancing personalization and semantic understanding through LLMs
• Enabling zero-shot reasoning for long-tail recommendation scenarios

Innovation

• LLMs unify recommendation tasks via language-native mechanisms
• LLMs enable prompt-driven retrieval and conversational recommendation
• LLMs support zero-shot reasoning for cold-start scenarios