Who You Are Matters: Bridging Topics and Social Roles via LLM-Enhanced Logical Recommendation

TL;DR

This paper tackles the limitation of conventional recommender systems that focus on item topics by introducing user roles and behavioral logic as explicit modeling targets. TagCF combines multi-modal LLM-based tag extraction with LLM-driven logic graphs to connect user roles to item topics, updating tag sets and logic relations daily and distilling scalable models. The framework integrates tag semantics through a tag-based encoder, learning augmentation with contrastive objectives, and an explicit tag-logic inference extension to boost accuracy and diversity. Empirical results from online A/B tests and offline datasets show that user-role modeling improves recommendation quality and diversity, and that the extracted tag-logic graphs are transferable across tasks, with practical deployment considerations and evidence of long-term user engagement benefits.

Abstract

Recommender systems filter contents/items valuable to users by inferring preferences from user features and historical behaviors. Mainstream approaches follow the learning-to-rank paradigm, which focus on discovering and modeling item topics (e.g., categories), and capturing user preferences on these topics based on historical interactions. However, this paradigm often neglects the modeling of user characteristics and their social roles, which are logical confounders influencing the correlated interest and user preference transition. To bridge this gap, we introduce the user role identification task and the behavioral logic modeling task that aim to explicitly model user roles and learn the logical relations between item topics and user social roles. We show that it is possible to explicitly solve these tasks through an efficient integration framework of Large Language Model (LLM) and recommendation systems, for which we propose TagCF. On the one hand, TagCF exploits the (Multi-modal) LLM's world knowledge and logic inference ability to extract realistic tag-based virtual logic graphs that reveal dynamic and expressive knowledge of users, refining our understanding of user behaviors. On the other hand, TagCF presents empirically effective integration modules that take advantage of the extracted tag-logic information, augmenting the recommendation performance. We conduct both online experiments and offline experiments with industrial and public datasets as verification of TagCF's effectiveness, and we empirically show that the user role modeling strategy is potentially a better choice than the modeling of item topics. Additionally, we provide evidence that the extracted logic graphs are empirically a general and transferable knowledge that can benefit a wide range of recommendation tasks. Our code is available in https://github.com/Code2Q/TagCF.

Figures (13)

• Figure 1: The toy example of the progress from traditional methods to tag-logic modeling.
• Figure 2: The main framework of the proposed TagCF.
• Figure 3: The deployment of TagCF in the online recommender system.
• Figure 4: The ablation results of the three key methods of the tag-logic integration module.
• Figure 5: The model performance with different $\beta_0$ and $\beta_1$.