Harmonizing Semantic and Collaborative in LLMs: Reasoning-based Embedding Generator for Sequential Recommendation
2026-06-15 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors study ways to improve Sequential Recommender Systems (SRS), which suggest what users might like next based on their past interactions. They point out that existing methods using Large Language Models (LLMs) don't fully use the models' reasoning abilities and lack clear ways to align collaborative information. To fix this, the authors create ReaEmb, which uses a two-step process combining reasoning-focused contrastive learning and reinforcement learning with collaborative feedback. Their experiments show ReaEmb works better than previous methods for recommending items. They also provide their code to help others reproduce the results.
Sequential Recommender SystemsLarge Language Modelsembeddingcontrastive learningreinforcement learningcollaborative filteringlatent reasoningrecommendation algorithms
Authors
Qidong Liu, Mingyao Huang, Moranxin Wang, Wenxuan Yang, Haiping Zhu
Abstract
Sequential Recommender Systems (SRS) predict the next item of interest based on users' interaction histories and have been widely deployed, but hindered by long-tail problem. Large Language Models (LLMs), with strong semantic understanding and reasoning capabilities, offer a promising way to enrich item semantics and have recently been used as embedding generators. However, two fundamental gaps remain. First, current LLM-based embedding methods fail to exploit the model's inner reasoning capacity. Second, existing methods often inject collaborative signals implicitly via supervised fine-tuning, lacking explicit guidance for collaborative embedding alignment. In this paper, we introduce ReaEmb, a novel framework that resolves both issues via a Latent Reasoning-enhanced Contrastive Learning (LRCL) stage and a Collaborative Reward Reinforcement Learning (CRRL) stage. LRCL exploits the LLMs' inner reasoning capacity through a two-pass forward process with an additional attention module. CRRL subsequently explicitly injects collaborative signals into the LLM via a tailored reinforcement learning. Extensive experiments on three real-world datasets demonstrate superior effectiveness of ReaEmb across multiple SRS models. To ease reproducibility, we release the code online.