PEETS:基于显式树结构的偏好感知辅助商品捆绑推荐框架
PEETS: Preference-aware Embeddings with Explicit Tree Structure for Auxiliary Product Bundling
做电商或旅游推荐的可以看这篇:它专门解决捆绑推荐里"拒绝一个SKU不等于不要整个品类"的问题,用树结构对比损失加拒绝信号改训练。离线AUC比TWIN V2高0.9个点,线上GMV提升8.2%且延迟几乎没涨,离线部署成本很低。
论文提出 PEETS(Preference-aware Embeddings with Explicit Tree Structure),一个面向旅行辅助商品捆绑推荐(如机票保障、灵活改签服务)的训练期框架。该框架在现有点击-转化预测器上叠加两种条件对比损失:Hierarchical Contrastive Loss 以商品对的共同祖先节点作为参照,Negative Contrastive Loss 则将结构采样与对比偏移建立在用户历史拒绝事件之上。
两种损失都作用于训练期的共享 SKU 表征,把类目相关性(同一 SKU 类型内商品语义相似但价格与覆盖不同)与个体偏好监督分离开来,同时完全不动线上预测路径,不增加任何辅助损失计算。离线评测中 PEETS 达到 78.20% AUC,比 TWIN V2 高 0.90 个百分点,比匹配的基线模型高 1.15 个百分点。
为期 14 天的线上实验显示,每曝光用户日的辅助商品 GMV 相比 Policies 提升 8.20%,相比 Base 提升 3.06%,相比 TWIN V2 提升 0.84%,三个用户 bootstrap 95% 置信区间下限均大于零。捆绑率比 Policies 低 0.51 个百分点;取消指标的置信区间上限超出预设边际。服务延迟方面 p95 为 42.1 ms,与 Base 的 42.0 ms 基本持平。
PEETS: Preference-aware Embeddings with Explicit Tree Structure for Auxiliary Product Bundling
Auxiliary products such as flight protection and flexible-change services are recommended alongside primary travel purchases. This setting creates a granularity ambiguity: SKUs within one product type share semantics but differ in price and coverage, while rejecting one offer does not imply disinterest in the entire type. We present \textbf{PEETS} (\textbf{P}reference-aware \textbf{E}mbeddings with \textbf{E}xplicit \textbf{T}ree \textbf{S}tructure), a training-time framework for auxiliary product bundling. PEETS regularizes an existing click--conversion predictor through two conditional comparisons. Hierarchical Contrastive Loss uses the pair's common ancestor as a reference; Negative Contrastive Loss conditions structural sampling and comparison shifts on logged rejection events. Both act on shared SKU representations during training. This separates catalog relatedness from individual preference supervision and leaves the predictive serving path intact. PEETS achieves 78.20\% AUC, exceeding TWIN V2 by 0.90 percentage points and the matched base model by 1.15 points. In a 14-day online evaluation, gross auxiliary GMV per exposed user-day improves by 8.20\% over Policies, 3.06\% over Base, and 0.84\% over TWIN V2, with all three user-bootstrap 95\% intervals above zero. Bundling rate is 0.51 points below Policies. The cancellation interval's upper limit exceeds the prespecified margin. PEETS adds no auxiliary-loss computation to serving and records 42.1~ms p95 latency, compared with 42.0~ms for Base.
This report is categorized under AITOP's "Papers" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include PEETS, recommendation systems, contrastive learning, TWIN V2.
Source attribution: curated from arXiv cs.AI (@Maolei Huang, Jing Tian, Dexuan Duan, Yan Chen, Huawei Cao, Shuhan Song).