PERSCEN: Learning Personalized Interaction Pattern and Scenario Preference for Multi-Scenario Matching

2025-06-23

AI Summary

Existing multi-scenario matching methods lack user-specific modeling, limiting their ability to capture personalized preferences. To address this, we propose a joint modeling framework comprising two core components: (1) a user-specific feature graph, enhanced by a lightweight graph neural network to capture high-order cross-scenario interactions; and (2) a vector quantization mechanism to extract fine-grained intra-scenario behavioral preferences, coupled with a progressive gated fusion module for low-latency, interpretable scenario-aware information integration. Our approach achieves significant improvements over state-of-the-art methods across multiple real-world industrial scenarios, striking an effective balance between recommendation accuracy and inference efficiency. It demonstrates strong scalability and practical deployment value, particularly in large-scale, latency-sensitive production environments.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

With the expansion of business scales and scopes on online platforms, multi-scenario matching has become a mainstream solution to reduce maintenance costs and alleviate data sparsity. The key to effective multi-scenario recommendation lies in capturing both user preferences shared across all scenarios and scenario-aware preferences specific to each scenario. However, existing methods often overlook user-specific modeling, limiting the generation of personalized user representations. To address this, we propose PERSCEN, an innovative approach that incorporates user-specific modeling into multi-scenario matching. PERSCEN constructs a user-specific feature graph based on user characteristics and employs a lightweight graph neural network to capture higher-order interaction patterns, enabling personalized extraction of preferences shared across scenarios. Additionally, we leverage vector quantization techniques to distil scenario-aware preferences from users' behavior sequence within individual scenarios, facilitating user-specific and scenario-aware preference modeling. To enhance efficient and flexible information transfer, we introduce a progressive scenario-aware gated linear unit that allows fine-grained, low-latency fusion. Extensive experiments demonstrate that PERSCEN outperforms existing methods. Further efficiency analysis confirms that PERSCEN effectively balances performance with computational cost, ensuring its practicality for real-world industrial systems.

Problem

• Capturing shared and scenario-specific user preferences in multi-scenario matching
• Addressing lack of personalized user modeling in existing methods
• Balancing recommendation performance with computational efficiency

Innovation

• User-specific feature graph with GNN
• Vector quantization for scenario-aware preferences
• Progressive scenario-aware gated linear unit