腾讯广告中长行为序列建模的实践
Practice on Long Behavior Sequence Modeling in Tencent Advertising
AI总结:
针对广告域行为稀疏问题,提出跨域构建长行为序列的两阶段框架,通过分层硬搜索、解耦嵌入软搜索、解耦TIN及目标解耦编码等方法,在腾讯广告平台实现GMV显著提升。
摘要:
长序列建模已成为推荐系统中捕捉用户长期偏好的不可或缺的前沿方向。然而,广告域内的用户行为本质上是稀疏的,这给仅使用单一广告域数据构建长行为序列带来了重大障碍。这促使我们不仅跨不同广告场景收集用户行为,还跨越广告域的边界进入内容域,从而构建统一的商业行为轨迹。这种跨域或跨场景的整合带来了以下挑战:(1)不同场景和域之间的特征分类体系差距;(2)不相关特征字段对引起的跨字段干扰;(3)针对不同广告目标优化时,在时间和语义模式上的目标级干扰。为解决这些挑战,我们在两阶段长序列建模框架内提出了若干实用方法。在第一(搜索)阶段,我们设计了一种分层硬搜索方法来处理复杂的特征分类体系层级,同时采用一种基于解耦嵌入的软搜索来缓解注意力机制与特征表示之间的冲突。在第二(序列建模)阶段,我们引入了:(a)解耦侧信息时间兴趣网络(TIN)以减轻跨字段冲突;(b)目标解耦位置编码和目标解耦SASRec以解决目标级干扰;(c)堆叠TIN以建模高阶行为相关性。在腾讯大规模广告平台上生产部署后,我们的创新带来了显著的性能提升:微信视频号整体GMV提升4.22%,微信朋友圈整体GMV提升1.96%。
英文摘要:
Long-sequence modeling has become an indispensable frontier in recommendation systems for capturing users' long-term preferences. However, user behaviors within advertising domains are inherently sparse, posing a significant barrier to constructing long behavioral sequences using data from a single advertising domain alone. This motivates us to collect users' behaviors not only across diverse advertising scenarios, but also beyond the boundaries of the advertising domain into content domains-thereby constructing unified commercial behavior trajectories. This cross-domain or cross-scenario integration gives rise to the following challenges: (1) feature taxonomy gaps between distinct scenarios and domains, (2) inter-field interference arising from irrelevant feature field pairs, and (3) target-wise interference in temporal and semantic patterns when optimizing for different advertising targets. To address these challenges, we propose several practical approaches within the two-stage framework for long-sequence modeling. In the first (search) stage, we design a hierarchical hard search method for handling complex feature taxonomy hierarchies, alongside a decoupled embedding-based soft search to alleviate conflicts between attention mechanisms and feature representation. In the second (sequence modeling) stage, we introduce: (a) Decoupled Side Information Temporal Interest Networks (TIN) to mitigate inter-field conflicts; (b) Target-Decoupled Positional Encoding and Target-Decoupled SASRec to address target-wise interference; and (c) Stacked TIN to model high-order behavioral correlations. Deployed in production on Tencent's large-scale advertising platforms, our innovations delivered significant performance gains: an overall 4.22% GMV lift in WeChat Channels and an overall 1.96% GMV increase in WeChat Moments.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。