Scenario-Adaptive Fine-Grained Personalization Network: Tailoring User Behavior Representation to the Scenario Context

• DOI: 10.48550/arxiv.2404.09709
• Publication Date: 2024-04-15
• Authors: Moyu Zhang; Yongxiang Tang; Jinxin Hu; Yanwen Zhang
• Keywords: Representation

Existing methods often adjust representations adaptively only after aggregating user behavior sequences. This coarse-grained approach to re-weighting the entire sequence hampers model's ability accurately model interest migration across different scenarios. To enhance capacity capture interests from historical sequences in each scenario, we develop a ranking framework named Scenario-Adaptive Fine-Grained Personalization Network (SFPNet), which designs kind of fine-grained method for multi-scenario personalized recommendations. Specifically, SFPNet comprises series blocks as Scenario-Tailoring Block, stacked sequentially. Each block initially deploys parameter personalization unit integrate scenario information at level by redefining fundamental features. Subsequently, consolidate scenario-adaptively adjusted feature serve context information. By employing residual connection, incorporate this into representation behavior, allowing context-aware customization scenario-level, turn supports scenario-aware modeling.