Skip to main content
QUICK REVIEW

[论文解读] PI2I: A Personalized Item-Based Collaborative Filtering Retrieval Framework

Shaoqing Wang, Yingcai Ma|arXiv (Cornell University)|Jan 23, 2026
Recommender Systems and Techniques被引用 0
一句话总结

PI2I 通过扩展候选池的松弛索引阶段和应用交互式评分阶段,实现对项目对项目筛选的个性化,提高在淘宝的检索准确性与在线表现。同时提供大规模淘宝互动数据集用于基准测试。

ABSTRACT

Efficiently selecting relevant content from vast candidate pools is a critical challenge in modern recommender systems. Traditional methods, such as item-to-item collaborative filtering (CF) and two-tower models, often fall short in capturing the complex user-item interactions due to uniform truncation strategies and overdue user-item crossing. To address these limitations, we propose Personalized Item-to-Item (PI2I), a novel two-stage retrieval framework that enhances the personalization capabilities of CF. In the first Indexer Building Stage (IBS), we optimize the retrieval pool by relaxing truncation thresholds to maximize Hit Rate, thereby temporarily retaining more items users might be interested in. In the second Personalized Retrieval Stage (PRS), we introduce an interactive scoring model to overcome the limitations of inner product calculations, allowing for richer modeling of intricate user-item interactions. Additionally, we construct negative samples based on the trigger-target (item-to-item) relationship, ensuring consistency between offline training and online inference. Offline experiments on large-scale real-world datasets demonstrate that PI2I outperforms traditional CF methods and rivals Two-Tower models. Deployed in the "Guess You Like" section on Taobao, PI2I achieved a 1.05% increase in online transaction rates. In addition, we have released a large-scale recommendation dataset collected from Taobao, containing 130 million real-world user interactions used in the experiments of this paper. The dataset is publicly available at https://huggingface.co/datasets/PI2I/PI2I, which could serve as a valuable benchmark for the research community.

研究动机与目标

  • 在传统基于项目的协同过滤和双塔模型之外,说明在大规模推荐系统中对个性化检索的需求。
  • 提出一个两阶段的 PI2I 框架,先扩展候选池(IBS)再应用交互式个性化评分(PRS)。
  • 通过一个触发-目标负采样策略和一个专门的损失函数,使离线训练与在线推理保持一致。
  • 在淘宝部署中展示离线收益与在线影响,并提供用于研究的大规模淘宝互动数据集。

提出的方法

  • IBS:通过放宽截断阈值构建项目对项目(i2i)表,以最大化命中率并在触发时保留更多候选项,使用 Swing 基于评分。
  • PRS:使用带有目标注意力和跨特征交互的交互式评分模型对候选项进行打分;采用多头目标注意力进行检索评分。
  • 触发-目标采样:以下一次点击的积极样本进行训练;从相关目标中生成困难负样本,从未触发的相关项中生成简单负样本,以对齐离线与在线过程。
  • 损失:优化负对数似然,使正 logits 大于负 logits(L^p)。
  • 推理:使用来自用户历史的触发项和 i2i_table 候选项,在一个大候选空间上进行在线异步打分(Top-K)。
Figure 1 . A general multi-stage architecture in modern recommender systems.
Figure 1 . A general multi-stage architecture in modern recommender systems.

实验结果

研究问题

  • RQ1在 IBS 阶段放宽截断是否能在不带来高计算代价的前提下提高命中率?
  • RQ2PRS 中的交互式评分方法是否优于基于内积的检索在个性化物品推荐中的表现?
  • RQ3基于触发-目标的负采样是否提升离线训练与在线推理的一致性?
  • RQ4在大规模工业数据和公开淘宝数据集上,PI2I 相对于 CF 和主流 Two-Tower 模型的表现如何?

主要发现

  • PI2I 在大规模真实数据集的离线实验中优于传统 CF 方法,并可与 Two-Tower 模型媲美。
  • 在淘宝的“猜你喜欢”板块,PI2I 实现在线交易率提升约 1.05%。
  • PI2I 在密集数据集(KuaiRec)上表现强劲,在稀疏淘宝数据集上也具有竞争力,特别是在较大 top-K 阈值下(如 Hit@4000)。
  • 消融研究表明目标注意力和多值触发对检索性能有提升;移除触发项或使用单值触发会降低准确性。
  • 参数研究确定了一个最优的 IBS 截断大小(T=1250),在提升命中率的同时保持运营效率。
  • 案例研究显示触发概率随时间的衰减,以及跨用户的触发分布个性化趋势。
Figure 2 . (a) Illustration of traditional I2I methods. (b) Illustration of the trigger-target sampling strategy. (c) The overall framework of PI2I . (d) Procedure of the PI2I Online Asynchronous Inference.
Figure 2 . (a) Illustration of traditional I2I methods. (b) Illustration of the trigger-target sampling strategy. (c) The overall framework of PI2I . (d) Procedure of the PI2I Online Asynchronous Inference.

更好的研究,从现在开始

从论文设计到论文写作,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。