OxygenREC-v2: Internalizing Discrimination into Generative Recommendation
2026-07-27 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors developed OxygenREC-v2, a recommendation system that combines retrieving and ranking items using one single model. Instead of adding a separate ranking step, their model learns directly from user behaviors like clicks and purchases to guide its suggestions. They train the model in two phases: first by conditioning on specific user actions, then by using future behaviors to refine its predictions without extra reward models. When tested on a big e-commerce site, their approach improved how often users click and buy items compared to the previous version.
generative recommendationautoregressive decodingsemantic identifier sequencesdiscriminative objectiveself-distillationentropy-aware trajectory optimizationMixture of Experts (MoE)click-through conversion rateGMV (Gross Merchandise Volume)reward-model-free policy optimization
Authors
Guo Tang, Hanye Wu, Changjiang Han, Qingyang Li, Ming Zhang, Xiangyu Qian, Yanchen Qiao, Huanjie Wang, Zhi Ma, Zhen Li, Yaqiang Zang, Pinghua Gong
Abstract
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.