OneTrans-V2：用一个Transformer统一工业推荐系统中的召回、粗排与精排
OneTrans-V2: Unifying Retrieval, Pre-rank, and Fine-rank with One Transformer in Industrial Recommender
- ByteDance Global E-Commerce Recommendation Foundation Team（字节跳动全球电商推荐基础团队）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
OneTrans-V2用一个Transformer统一工业推荐系统的召回、粗排和精排，通过联合训练、稀疏MoE和决策条件生成式召回，提升GMV 9.74%并实现3.2倍吞吐量。
AI中文摘要：
工业推荐系统通常以召回、粗排和精排的级联方式运行，但这些阶段通常作为独立模型进行训练和服务，导致用户序列的重复编码、孤立的优化以及重复的工程工作。在OneTrans模型级统一的基础上，我们提出了OneTrans-V2，一个统一整个级联流程的Transformer。它将用户行为序列编码一次作为共享上下文，同时保留各阶段特有的候选特征和计算。联合训练使三个阶段相互增强，并支持从精排到粗排的模型内知识蒸馏。我们使用稀疏混合专家（MoE）扩展共享主干，在受限激活计算下增加容量，并通过μP风格参数化稳定扩展。为整合目标特定的召回通道，我们引入了决策条件生成式召回（DCGR）。DCGR预测描述即将发生交互的决策前缀，并基于该前缀生成物品，使业务目标能够引导单一生成过程。最后，序列原生训练（SNT）围绕每个用户的终身行为序列组织训练，并在多次曝光中摊销其编码。OneTrans-V2已部署在一个大规模工业推荐系统的所有三个阶段，将商品交易总额（GMV）提升了9.74%，并通过协同设计的服务栈，在相同硬件预算下提供了其替代级联系统3.2倍的吞吐量。
英文摘要：
Industrial recommendation systems typically operate as a \emph{cascade} of retrieval, pre-rank, and fine-rank, but these stages are usually trained and served as separate models, causing repeated user-sequence encoding, isolated optimization, and duplicated engineering effort. Building on OneTrans' model-level unification, we present OneTrans-V2, one Transformer that unifies the entire cascade. It encodes the user behavior sequence once as a shared context while preserving stage-specific candidate features and computation. Joint training lets the three stages reinforce one another and enables in-model knowledge distillation from fine-rank to pre-rank. We scale the shared backbone with sparse mixture-of-experts (MoE), which increases capacity with bounded activated computation, and stabilize scaling with $μ$P-style parameterization. To consolidate objective-specific retrieval channels, we introduce Decision-Conditioned Generative Retrieval (DCGR). DCGR predicts a decision prefix describing the upcoming interaction and generates items conditioned on it, allowing business objectives to steer a single generative process. Finally, Sequence-Native Training (SNT) organizes training around each user's lifelong behavior sequence and amortizes its encoding across exposures. Deployed across all three stages of a large-scale industrial recommendation system, OneTrans-V2 improves gross merchandise value (GMV) by 9.74\% and, with a co-designed serving stack, delivers $3.2\times$ the throughput of the cascade it replaces under the same hardware budget.
