A Serendipitous Recommendation System Considering User Curiosity
Zhelin Xu, Atsushi Matsumura
TL;DR
This work tackles the overemphasis on prediction accuracy in recommender systems by introducing serendipity, balancing usefulness with unexpectedness. It proposes a per-user diversive curiosity model derived from long-term and short-term preferences to weight usefulness and unexpectedness, producing personalized serendipitous recommendations. The method combines a neural usefulness predictor with a latent-space unexpectedness measure and calibrates serendipity via a per-user curiosity score, demonstrating competitive performance and higher unexpectedness on MovieLens-1M compared to a state-of-the-art baseline. The study contributes a practical framework for curiosity-aware serendipity, enabling tailored exploration and potential improvements in user satisfaction and discovery in real-world systems.
Abstract
To address the problem of narrow recommendation ranges caused by an emphasis on prediction accuracy, serendipitous recommendations, which consider both usefulness and unexpectedness, have attracted attention. However, realizing serendipitous recommendations is challenging due to the varying proportions of usefulness and unexpectedness preferred by different users, which is influenced by their differing desires for knowledge. In this paper, we propose a method to estimate the proportion of usefulness and unexpectedness that each user desires based on their curiosity, and make recommendations that match this preference. The proposed method estimates a user's curiosity by considering both their long-term and short-term interests. Offline experiments were conducted using the MovieLens-1M dataset to evaluate the effectiveness of the proposed method. The experimental results demonstrate that our method achieves the same level of performance as state-of-the-art method while successfully providing serendipitous recommendations.