HELIX：净化与统一——重新思考大规模推荐中的特征交互与序列建模
HELIX: Purified and Unified - Rethinking Feature Interaction and Sequence Modeling for Large-Scale Recommendation
浏览论文内容
中文总结 AI 辅助
针对大规模推荐中特征交互与序列建模单独扩展受限的问题，提出HELIX统一架构，通过交错序列检索与特征交互并保持单向信息流，实现两轴联合扩展，在TikTok部署后GMV提升约6%。
中文摘要 AI 辅助
工业推荐排序模型通常沿两个建模轴进行扩展：一是在异构用户、物品、上下文和交叉特征上进行特征交互，二是在长序列、信息丰富且多类型的用户行为历史上进行序列建模。我们发现，单独扩展任一能力均不足，因为各自都表现出有限的扩展上限和次优的扩展律斜率。我们推测，要实现更有利的扩展律斜率，需要同时扩展两个轴。为此，我们提出HELIX，一种用于大规模推荐的净化且统一的架构。HELIX交错进行序列检索和特征交互，同时强制从可复用的序列状态到候选条件混合令牌的单向信息流。这种设计在保持用户侧序列计算可摊销的同时，保留了两个建模轴之间的跨深度通信，从而实现对序列建模和特征交互的灵活且不对称的扩展。在TikTok电商推荐系统中部署后，HELIX持续提升了离线CTR AUC、CVR AUC及其他排序指标。在线A/B测试中，它实现了每位用户电商视频GMV约6%的增长。
英文摘要
Industrial recommendation ranking models typically scale along two modeling axes: feature interaction over heterogeneous user, item, context, and cross features, and sequence modeling over long, informative, and multi-type user behavior histories. We find that scaling either capability in isolation is insufficient, as each exhibits a limited scaling ceiling and a suboptimal scaling-law slope. We conjecture that achieving a more favorable scaling-law slope requires jointly scaling both axes. To support this, we present HELIX, a purified and unified architecture for large-scale recommendation. HELIX interleaves sequence retrieval and feature interaction while enforcing one-way information flow from reusable sequence states to candidate-conditioned mix-tokens. This design preserves cross-depth communication between the two modeling axes while keeping user-side sequence computation amortizable, enabling flexible and asymmetric scaling of sequence modeling and feature interaction. Deployed in TikTok's e-commerce recommendation system, HELIX consistently improves offline CTR AUC, CVR AUC, and other ranking metrics. In online A/B tests, it achieves an approximately 6% increase in e-commerce video GMV per user.
发表机构
- Global E-Commerce Recommendation Video Team（全球电商推荐视频团队）
机构由 AI 辅助整理，请以论文原文为准。
