Unifying Generative Recall and Multi-Objective Ranking in a Single Decoder-Only Sequence
2026-07-27 • Information Retrieval
Information Retrieval
AI summaryⓘ
The authors address problems in recommendation systems that separately handle recalling items and ranking them, which can lead to inefficiencies and loss of information. They introduce UniR², a single Transformer model that combines these two steps by processing user data, item history, and item features all together. This approach uses a special attention mechanism to let the model focus differently on each task while sharing core knowledge. Tests on large datasets and real users show that this unified method works well and improves system performance.
recommendation systemsrecallrankingTransformergenerative modelsmulti-objective optimizationattention mechanismLoRAuser contextitem features
Authors
Ruochen Yang, Shuang Wen, Pengbo Xu, Yusheng Huang, Jiangxia Cao, Shuang Yang, Zhaojie Liu, Jiawei Sheng, Tingwen Liu
Abstract
Modern industrial recommendation systems typically separate recall and ranking into two independent stages. Although this cascade supports corpus-level retrieval and fine-grained multi-objective scoring, it causes objective inconsistency, information loss at the candidate hand-off, and redundant user-side context computation. Meanwhile, the generative recall and ranking scaling share a common Transformer-based modeling philosophy, where architectural consistency creates a natural opportunity for unified integration. However, direct sharing remains challenging since the two tasks require different information visibility and optimization methods. Therefore, we propose \textbf{UniR$^2$}, a \textbf{Uni}fied decoder-only Transformer that unifies Generative \textbf{R}ecall and Multi-Objective \textbf{R}anking within a single heterogeneous sequence comprising user context, SID trajectory, and item features. Within this sequence, the generated trajectory serves as a representation bridge between recall and ranking, where Dual-Query Prefix-Causal Attention provides task-specific visibility. The two tasks share the base attention weights but retain separate optimization boundaries, with ranking-side LoRA preserving ranking adaptability without disrupting the generative backbone. Extensive offline experiments on large-scale industrial data demonstrate the effectiveness and efficiency of UniR$^2$ for both recall and ranking. Long-term online A/B tests on Kuaishou platform further show consistent positive gains, validating the practicality of unified model in large-scale recommendation systems.