Personalized multi-head self-attention network for news recommendation.

Area of Science:

• Computer Science
• Artificial Intelligence
• Information Retrieval

Background:

• Personalized news recommender systems combat information overload.
• Existing methods struggle to systematically model user latent interests and explicit preferences.
• Focus on accuracy often overlooks interpretability and efficient resource utilization.

Purpose of the Study:

• To develop an effective framework for personalized news recommendation.
• To enhance the modeling of user latent interests and explicit preferences.
• To improve recommendation accuracy and interpretability without complex designs.

Main Methods:

• Designed a Personalized Multi-Head Self-Attention Network (PMSN).
• Integrated multi-head self-attention for modeling word and news interactions.
• Employed a personalized attention mechanism using user IDs for enhanced interpretability.

Main Results:

• PMSN demonstrated superior performance over state-of-the-art methods on real-world datasets.
• Achieved high recommendation accuracy.
• Validated the interpretability enhancement through attention mechanisms via case studies.

Conclusions:

• PMSN offers an efficient and interpretable solution for personalized news recommendation.
• The model effectively captures user latent interests and explicit preferences.
• Outperforms existing methods without requiring complex structures or extensive resources.