OxygenREC-v2：将歧视内化到生成式推荐中
OxygenREC-v2: Internalizing Discrimination into Generative Recommendation
浏览论文内容
中文总结 AI 辅助
研究针对生成式推荐中纳入行为信号的挑战，提出OxygenREC-v2，通过用记录行为条件生成及监督训练，在预训练和训练后分别采用不同方式，保持统一主干，部署在电商平台，相比OxygenREC-v1提升了用户点击率转化率和商品交易总额。
中文摘要 AI 辅助
生成式推荐通过自回归解码语义标识符（SID）序列，在单个模型中统一了检索和排序。然而，可靠地纳入来自点击、添加到购物车和订单的行为信号仍然具有挑战性。现有方法要么联合优化生成式和判别式目标，需要微妙的权衡，要么使用单独的排序器作为事后强化学习奖励，存在分布外评分和奖励不一致的风险。我们提出了OxygenREC-v2，一种将歧视内化到生成式推荐（IDGR）中的生成式推荐器。OxygenREC-v2不是添加单独的判别式目标，而是使用记录的行为来条件生成并监督训练。在预训练期间，行为指令根据目标行为来条件生成。在训练后，未来的交互行为在我们的熵感知轨迹优化自蒸馏框架中被用作特权知识，实现无奖励模型的策略优化。在两个训练阶段，OxygenREC-v2都保持一个统一的主干。我们将OxygenREC-v2实现为一个具有3B参数、1B激活的混合专家（MoE），并将其部署在该论文的大型电子商务平台上。在多个在线A/B测试中，OxygenREC-v2比OxygenREC-v1将用户点击率转化率（UCTCVR）提高了1.6 - 4.4%，商品交易总额（GMV）提高了2.8 - 6.8%。
英文摘要
Generative recommendation unifies retrieval and ranking within a single model by autoregressively decoding semantic identifier (SID) sequences. Yet reliably incorporating behavior signals from clicks, cart additions, and orders remains challenging. Existing approaches either jointly optimize generative and discriminative objectives, requiring delicate trade-offs, or use a separate ranker as a post-hoc reinforcement-learning reward, risking out-of-distribution scoring and reward misalignment. We propose OxygenREC-v2, a generative recommender that Internalizes Discrimination into Generative Recommendation (IDGR). Rather than adding a separate discriminative objective, OxygenREC-v2 uses logged behavior to condition generation and supervise training. During pre-training, a behavior instruction conditions generation on the target behavior. During post-training, future interaction behaviors are exploited as privileged knowledge in our entropy-aware trajectory optimization self-distillation framework, enabling reward-model-free policy optimization. Throughout both training stages, OxygenREC-v2 maintains a single unified backbone. We implement OxygenREC-v2 as a 3B-parameter, 1B-activated MoE and deploy it on JD.com's large-scale e-commerce platform. Across multiple online A/B tests, OxygenREC-v2 improves user click-through conversion rate (UCTCVR) by 1.6--4.4% and GMV by 2.8--6.8% over OxygenREC-v1.
