📄 中文摘要
研究提出了一种轻量级的双塔模型，以解决传统的受欢迎程度排名在餐厅发现中的不足。该模型通过结合用户和餐厅的特征，利用嵌入技术实现个性化推荐。与传统方法相比，双塔模型在处理用户偏好和餐厅特征方面具有更高的灵活性和准确性。实验结果表明，该模型显著提高了用户对餐厅的发现率和满意度，为餐饮行业的个性化推荐提供了新的思路和方法。
基于双塔嵌入变体的个性化餐厅排名
出处: Personalized Restaurant Ranking with a Two-Tower Embedding Variant
发布: 2026年3月13日
📄 English Summary
Personalized Restaurant Ranking with a Two-Tower Embedding Variant
A lightweight two-tower model has been proposed to address the shortcomings of traditional popularity ranking in restaurant discovery. By integrating user and restaurant features, the model employs embedding techniques for personalized recommendations. Compared to conventional methods, the two-tower model offers greater flexibility and accuracy in handling user preferences and restaurant characteristics. Experimental results indicate that this model significantly enhances users' discovery rates and satisfaction with restaurants, providing new insights and methodologies for personalized recommendations in the dining industry.