AI 中文总结
研究工业搜索和排名系统中CTR预测从DLRM转向Transformer架构的问题，提出天猫GS可扩展排名架构，含分层分布校准令牌化等五个关键组件，经实验和测试，该架构提高训练吞吐量，提升UCTCVR和GMV。
AI 中文摘要
在工业搜索和排名系统中，点击率（CTR）预测正从传统深度学习推荐模型（DLRM）转向统一的、计算密集型的Transformer架构。这种转变是由提高模型浮点运算利用率（MFU）和通过缩放定律实现可预测收益的需求驱动的。然而，现有方法在采用大语言模型（LLM）架构时往往采用全令牌化策略，忽略了排名特征的异构性。我们提出了天猫GS，一种用于天猫搜索的可扩展排名架构。天猫GS包括五个关键组件：分层分布校准令牌化，结合逐字段显著性重新加权（FSR）和分布校准投影（DCP）将不同特征映射到优化子空间；具有逐字段QKV投影和噪声自适应门控的字段自适应门控Transformer主干，用于精细语义交互；解耦FiLM后期融合以保留显式高频信号；上下文感知偏差网络，将系统偏差与用户意图解耦；以及具有动态加权损失的误差感知渐进训练，用于稳健学习。在天猫搜索上进行的广泛离线实验和在线A/B测试表明，天猫GS提高了训练吞吐量，并在UCTCVR和GMV方面取得了显著收益。
英文摘要
In industrial search and ranking systems, Click-Through Rate (CTR) prediction is shifting from traditional Deep Learning Recommendation Models (DLRM) toward unified, compute-intensive Transformer architectures. This transition is driven by the need to improve Model FLOPs Utilization (MFU) and achieve predictable gains through scaling laws. However, existing approaches such as OneTrans and Climber often adopt an all-in-tokenization strategy when adapting Large Language Model (LLM) architectures, overlooking the heterogeneous nature of ranking features. We propose TmallGS, a scalable ranking architecture for Tmall search. TmallGS includes five key components: (1) Hierarchical Distribution-Calibrated Tokenization, which combines Field-wise Saliency Reweighting (FSR) and Distribution-Calibrated Projection (DCP) to map diverse features into optimized subspaces; (2) a Field-Adaptive Gated Transformer Backbone with per-field QKV projections and noise-adaptive gating for refined semantic interaction; (3) Decoupled FiLM Late Fusion to preserve explicit high-frequency signals; (4) a Context-Aware Bias Net to decouple systemic bias from user intent; and (5) Error-Aware Progressive Training with dynamically weighted losses for robust learning. Extensive offline experiments and online A/B tests on Tmall Search show that TmallGS improves training throughput and achieves substantial gains in UCTCVR and GMV.
