[论文解读] Deep Session Interest Network for Click-Through Rate Prediction

Yufei Feng, Fuyu Lv+5 more|arXiv (Cornell University)|May 16, 2019

Recommender Systems and Techniques参考文献 19被引用 39

一句话总结

DSIN 模型将用户行为看作多个历史会话，利用自注意力提取每个会话的兴趣，再用 Bi-LSTM 模型化其演化，并通过局部激活单元聚合以预测 CTR。

ABSTRACT

Click-Through Rate (CTR) prediction plays an important role in many industrial applications, such as online advertising and recommender systems. How to capture users' dynamic and evolving interests from their behavior sequences remains a continuous research topic in the CTR prediction. However, most existing studies overlook the intrinsic structure of the sequences: the sequences are composed of sessions, where sessions are user behaviors separated by their occurring time. We observe that user behaviors are highly homogeneous in each session, and heterogeneous cross sessions. Based on this observation, we propose a novel CTR model named Deep Session Interest Network (DSIN) that leverages users' multiple historical sessions in their behavior sequences. We first use self-attention mechanism with bias encoding to extract users' interests in each session. Then we apply Bi-LSTM to model how users' interests evolve and interact among sessions. Finally, we employ the local activation unit to adaptively learn the influences of various session interests on the target item. Experiments are conducted on both advertising and production recommender datasets and DSIN outperforms other state-of-the-art models on both datasets.

研究动机与目标

• 强调用户行为序列由在会话内行为同质、跨会话行为异质组成。
• 提出 DSIN 以利用多历史会话进行 CTR 预测。
• 开发一个使用带偏置编码的会话级兴趣提取器。
• 用 Bi-LSTM 捕捉会话兴趣之间的时序关系。
• 使用局部激活单元自适应聚合会话兴趣，然后再进行最终预测。

提出的方法

• 将用户行为序列按时间间隔分割为会话（30分钟规则）。
• 应用带偏置编码的自注意力提取每个会话的兴趣。
• 使用 Bi-LSTM 捕捉会话兴趣之间的序列交互。
• 使用局部激活单元自适应地对会话兴趣进行加权以对抗目标项目。
• 将带有会话信息的表示与用户/物品嵌入拼接，然后通过 MLP 预测 CTR。

实验结果

研究问题

• RQ1 利用多历史会话并对会话层级进行注意是否能提升 CTR 预测，相较于单序列模型？
• RQ2 会话间演变和会话兴趣的自适应聚合是否带来更好的预测性能？
• RQ3 偏置编码和会话交互建模对 CTR 精度的贡献是什么？

主要发现

模型	广告 AUC	推荐 AUC
YoutubeNet-NO-UB a	0.6239	0.6419
YoutubeNet	0.6313	0.6425
DIN-RNN	0.6319	0.6435
Wide&Deep	0.6326	0.6432
DIN	0.6330	0.6459
DIEN	0.6343	0.6473
DSIN-PE b	0.6357	0.6494
DSIN-BE-NO-SIIL c	0.6365	0.6499
DSIN-BE d	0.6375	0.6515

• DSIN 在广告数据集和推荐数据集上获得比基线更好的 AUC。
• 引入多会话并建模其演化，相较于单会话模型如 DIN/DIEN，性能有所提升。
• 偏置编码和会话交互层显著提升 DSIN 的效果。
• 局部激活单元为会话兴趣相对于目标项提供自适应权重。