Modeling Stage-wise Evolution of User Interests for News Recommendation

2026-03-11

AI Summary

This work addresses the challenge that existing news recommendation methods struggle to jointly model users’ long-term stable preferences and short-term dynamic interests, primarily due to their reliance on static interaction graphs that overlook the temporal evolution of user interests. To overcome this limitation, the authors propose a unified framework that integrates both global and local temporal perspectives. The global component leverages graph neural networks to capture long-term collaborative signals, while the local component partitions user behavior into sequential subgraphs, employing LSTM to track recent interest dynamics and self-attention mechanisms to model long-range dependencies. Innovatively combining stage-wise temporal subgraph segmentation with a dual-branch architecture, the proposed method significantly outperforms strong baselines on two large-scale real-world datasets, yielding recommendations that are more timely and relevant.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Recommender Systems
• Planning, Routing, and Scheduling: Temporal Planning

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Personalized news recommendation is highly time-sensitive, as user interests are often driven by emerging events, trending topics, and shifting real-world contexts. These dynamics make it essential to model not only users'long-term preferences, which reflect stable reading habits and high-order collaborative patterns, but also their short-term, context-dependent interests that change rapidly over time. However, most existing approaches rely on a single static interaction graph, which struggles to capture both long-term preference patterns and short-term interest changes as user behavior evolves. To address this challenge, we propose a unified framework that learns user preferences from both global and local temporal perspectives. A global preference modeling component captures long-term collaborative signals from the overall interaction graph, while a local preference modeling component partitions historical interactions into stage-wise temporal subgraphs to represent short-term dynamics. Within this module, an LSTM branch models the progressive evolution of recent interests, and a self-attention branch captures long-range temporal dependencies. Extensive experiments on two large-scale real-world datasets show that our approach consistently outperforms strong baselines and delivers fresher and more relevant recommendations across diverse user behaviors and temporal settings.

Problem

• news recommendation
• user interest evolution
• long-term preference
• short-term interest
• temporal dynamics

Innovation

• stage-wise temporal subgraphs
• global-local preference modeling
• LSTM with self-attention
• dynamic user interest evolution
• news recommendation