Using LLMs to Capture Users' Temporal Context for Recommendation

2025-08-11

AI Summary

Traditional user profiling methods struggle to model fine-grained, dynamically evolving contextual preferences—such as short-term interests versus long-term tastes. This paper proposes a time-aware, LLM-driven user profiling framework that models user interaction history as dynamic context, explicitly disentangling and adaptively fusing short- and long-term behavioral features to generate semantically rich, temporally grounded user embeddings. Its key contribution is the first systematic empirical validation of LLMs’ capability in temporal preference disentanglement and fusion, alongside the construction of interpretable, time-aware user descriptions. Experiments demonstrate significant improvements in recommendation performance on dense datasets (e.g., Movies&TV), whereas gains are limited on sparse domains (e.g., Video Games), revealing data density as a critical constraint on the effectiveness of LLM-generated user profiles.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Effective recommender systems demand dynamic user understanding, especially in complex, evolving environments. Traditional user profiling often fails to capture the nuanced, temporal contextual factors of user preferences, such as transient short-term interests and enduring long-term tastes. This paper presents an assessment of Large Language Models (LLMs) for generating semantically rich, time-aware user profiles. We do not propose a novel end-to-end recommendation architecture; instead, the core contribution is a systematic investigation into the degree of LLM effectiveness in capturing the dynamics of user context by disentangling short-term and long-term preferences. This approach, framing temporal preferences as dynamic user contexts for recommendations, adaptively fuses these distinct contextual components into comprehensive user embeddings. The evaluation across Movies&TV and Video Games domains suggests that while LLM-generated profiles offer semantic depth and temporal structure, their effectiveness for context-aware recommendations is notably contingent on the richness of user interaction histories. Significant gains are observed in dense domains (e.g., Movies&TV), whereas improvements are less pronounced in sparse environments (e.g., Video Games). This work highlights LLMs' nuanced potential in enhancing user profiling for adaptive, context-aware recommendations, emphasizing the critical role of dataset characteristics for practical applicability.

Problem

• Assessing LLMs for time-aware user profiling in recommendations
• Disentangling short-term and long-term user preferences dynamically
• Evaluating LLM effectiveness in sparse vs. dense interaction domains

Innovation

• LLMs generate time-aware user profiles
• Disentangle short-term and long-term preferences
• Adaptively fuse contextual components into embeddings