CAP-Mix：用软原型混合桥接硬语义ID与连续生成式检索
原标题：Item Representations Matter: Bridging Hard Semantic IDs and Soft Prototype Mixtures for Generative Retrieval
AI 摘要
论文提出 CAP-Mix，一种连续生成式推荐方法，将每个物品表示为共享原型的软混合加上物品专属向量，并在推理时保留软表示。在三个数据集上，CAP-Mix 在稀疏 Amazon 数据集上优于扁平表，仅给定离线 SID 时 HR@5 比硬表提升 4.45%–14.90%，六个基准平均 HR@5 和 NDCG@5 分别提升 6.9% 和 13.6%，推理吞吐为离散基线的 1.4–161 倍。
正文节选
Item Representations Matter: Bridging Hard Semantic IDs and Soft Prototype Mixtures for Generative Retrieval Abstract Modern recommender systems rely on generative retrieval techniques to predict the next item either as discrete codes or as a continuous vector. On the one hand, Semantic IDs (SIDs) give discrete models shared parameters and pretrained semantic structure, but discretizing items into hard codes discards some information from the original content embeddings. On the other hand, conti