Paper Overview
Field: Machine Learning Authors: Ruizhong Qiu, Yinglong Xia, Dongqi Fu Published: 2025-06-23 arXiv: 2506.18494
Summary
Generative recommendation is an emerging paradigm that has shown promise in industrial recommendation systems, aiming to predict users' next interactions from their historical behaviors. At the core of generative recommendation lies item tokenization, which bridges item semantics and recommendation models.
However, existing methods often struggle to effectively organize and inject complex user-behavioral and item-semantic contexts into recommendation models simultaneously:
- Graph-based integration methods (e.g., graph serialization, graph neural networks) either suffer from scalability issues or exploit only local graph information.
- Semantic tokenization methods typically rely on heuristics and lack explicit supervision signals, which may lead to inaccurate or suboptimal semantic representations.
The Proposed Approach
To address these limitations in user interest context modeling, the authors propose G2Rec, a scalable framework that unifies holistic graph-based user co-engagement modeling with semantic tokenization for industrial-scale generative recommendation.
Overall, G2Rec enables recommendation models to capture holistic and semantically grounded user interest prototypes without requiring ground-truth user interest labels, thereby providing more comprehensive and accurate user behavioral context modeling. Online deployment and extensive experiments demonstrate G2Rec's superiority over existing methods.
--- *Source: arXiv:2506.18494*