Align$^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912725937946624 |
|---|---|
| author | Ye, Wencai Sun, Mingjie Chen, Shuhang Wu, Wenjin Jiang, Peng |
| author_facet | Ye, Wencai Sun, Mingjie Chen, Shuhang Wu, Wenjin Jiang, Peng |
| contents | Large Language Models (LLMs) demonstrate significant advantages in leveraging structured world knowledge and multi-step reasoning capabilities. However, fundamental challenges arise when transforming LLMs into real-world recommender systems due to semantic and behavioral misalignment. To bridge this gap, we propose Align$^3$GR, a novel framework that unifies token-level, behavior modeling-level, and preference-level alignment. Our approach introduces: Dual tokenization fusing user-item semantic and collaborative signals. Enhanced behavior modeling with bidirectional semantic alignment. Progressive DPO strategy combining self-play (SP-DPO) and real-world feedback (RF-DPO) for dynamic preference adaptation. Experiments show Align$^3$GR outperforms the SOTA baseline by +17.8% in Recall@10 and +20.2% in NDCG@10 on the public dataset, with significant gains in online A/B tests and full-scale deployment on an industrial large-scale recommendation platform. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_11255 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Align$^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation Ye, Wencai Sun, Mingjie Chen, Shuhang Wu, Wenjin Jiang, Peng Information Retrieval Large Language Models (LLMs) demonstrate significant advantages in leveraging structured world knowledge and multi-step reasoning capabilities. However, fundamental challenges arise when transforming LLMs into real-world recommender systems due to semantic and behavioral misalignment. To bridge this gap, we propose Align$^3$GR, a novel framework that unifies token-level, behavior modeling-level, and preference-level alignment. Our approach introduces: Dual tokenization fusing user-item semantic and collaborative signals. Enhanced behavior modeling with bidirectional semantic alignment. Progressive DPO strategy combining self-play (SP-DPO) and real-world feedback (RF-DPO) for dynamic preference adaptation. Experiments show Align$^3$GR outperforms the SOTA baseline by +17.8% in Recall@10 and +20.2% in NDCG@10 on the public dataset, with significant gains in online A/B tests and full-scale deployment on an industrial large-scale recommendation platform. |
| title | Align$^3$GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation |
| topic | Information Retrieval |
| url | https://arxiv.org/abs/2511.11255 |