Unleashing the Native Recommendation Potential: LLM-Based Generative Recommendation via Structured Term Identifiers

2026-01-11

arXiv.org

AI Summary

This work addresses the challenges faced by large language model (LLM)-based generative recommender systems, which often suffer from hallucination or semantic gaps due to the excessively large output space of item identifiers, thereby compromising both accuracy and generalization. To mitigate these issues, the authors propose Structured Term-based Identifiers (TIDs) as a novel representation for items and introduce the GRLM framework, which integrates context-aware term generation, unified instruction tuning, and flexible identifier alignment to effectively bridge LLMs with structured item metadata. This approach circumvents the hallucination pitfalls of conventional textual IDs and the high alignment cost of semantic IDs. Extensive experiments on multiple real-world datasets demonstrate that the proposed method significantly outperforms existing baselines, achieving superior recommendation performance and strong cross-scenario generalization capabilities.

Technology Category

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

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Leveraging the vast open-world knowledge and understanding capabilities of Large Language Models (LLMs) to develop general-purpose, semantically-aware recommender systems has emerged as a pivotal research direction in generative recommendation. However, existing methods face bottlenecks in constructing item identifiers. Text-based methods introduce LLMs'vast output space, leading to hallucination, while methods based on Semantic IDs (SIDs) encounter a semantic gap between SIDs and LLMs'native vocabulary, requiring costly vocabulary expansion and alignment training. To address this, this paper introduces Term IDs (TIDs), defined as a set of semantically rich and standardized textual keywords, to serve as robust item identifiers. We propose GRLM, a novel framework centered on TIDs, employs Context-aware Term Generation to convert item's metadata into standardized TIDs and utilizes Integrative Instruction Fine-tuning to collaboratively optimize term internalization and sequential recommendation. Additionally, Elastic Identifier Grounding is designed for robust item mapping. Extensive experiments on real-world datasets demonstrate that GRLM significantly outperforms baselines across multiple scenarios, pointing a promising direction for generalizable and high-performance generative recommendation systems.

Problem

• generative recommendation
• Large Language Models
• item identifiers
• Semantic IDs
• hallucination

Innovation

• Term IDs
• Generative Recommendation
• Large Language Models
• Structured Identifiers
• Instruction Fine-tuning