Fine-grained Semantics Integration for Large Language Model-based Recommendation

2026-02-26

AI Summary

This work addresses the limitations in semantic identifier (SID) space initialization for large language model (LLM)-based recommender systems, where initial embeddings often lack semantic richness and alignment is performed at a coarse item-level granularity. To overcome these issues, the authors propose a semantic-aware SID embedding initialization method coupled with an item-cluster-based token-level semantic alignment mechanism. Specifically, a teacher model extracts salient keywords from item descriptions, which are then combined with mean pooling to initialize SID embeddings; subsequent supervised fine-tuning enables fine-grained, token-level alignment between user queries and item representations. This approach transcends the conventional item-level optimization paradigm. Extensive experiments on two real-world datasets demonstrate that the proposed method significantly outperforms both generative and traditional recommendation baselines, confirming its effectiveness and state-of-the-art performance.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Recommender Systems
• Natural Language Processing: Sentence-level Semantics, Textual Inference, etc.

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

Recent advances in Large Language Models (LLMs) have shifted in recommendation systems from the discriminative paradigm to the LLM-based generative paradigm, where the recommender autoregressively generates sequences of semantic identifiers (SIDs) for target items conditioned on historical interaction. While prevalent LLM-based recommenders have demonstrated performance gains by aligning pretrained LLMs between the language space and the SID space, modeling the SID space still faces two fundamental challenges: (1) Semantically Meaningless Initialization: SID tokens are randomly initialized, severing the semantic linkage between the SID space and the pretrained language space at start point, and (2) Coarse-grained Alignment: existing SFT-based alignment tasks primarily focus on item-level optimization, while overlooking the semantics of individual tokens within SID sequences.To address these challenges, we propose TS-Rec, which can integrate Token-level Semantics into LLM-based Recommenders. Specifically, TS-Rec comprises two key components: (1) Semantic-Aware embedding Initialization (SA-Init), which initializes SID token embeddings by applying mean pooling to the pretrained embeddings of keywords extracted by a teacher model; and (2) Token-level Semantic Alignment (TS-Align), which aligns individual tokens within the SID sequence with the shared semantics of the corresponding item clusters. Extensive experiments on two real-world benchmarks demonstrate that TS-Rec consistently outperforms traditional and generative baselines across all standard metrics. The results demonstrate that integrating fine-grained semantic information significantly enhances the performance of LLM-based generative recommenders.

Problem

• Semantic Alignment
• Token-level Semantics
• Large Language Models
• Recommendation Systems
• Semantic Initialization

Innovation

• Token-level Semantic Alignment
• Semantic-Aware Embedding Initialization
• LLM-based Recommendation
• Fine-grained Semantics
• Semantic Identifiers (SIDs)