Unifying Generative Recall and Multi-Objective Ranking in a Single Decoder-Only Sequence

2026-07-27

AI Summary

This work addresses the misalignment, information loss, and computational redundancy arising from the conventional separation of retrieval and ranking in recommendation systems. The authors propose UniR², the first framework to unify generative retrieval and multi-objective ranking within a single Transformer decoder. By modeling heterogeneous sequences—comprising user context, session ID trajectories, and item features—and leveraging generated trajectories as representational bridges, UniR² introduces Dual-Query Prefix-Causal Attention to enforce task-specific visibility constraints. The architecture shares base attention weights across tasks while employing LoRA fine-tuning to balance ranking adaptability with generation stability. Extensive offline evaluations and long-term A/B tests on Kuaishou demonstrate that UniR² significantly outperforms baseline methods in both retrieval and ranking metrics, delivering consistent and substantial gains in real-world deployment.

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.

Problem

• recall
• ranking
• recommendation systems
• objective inconsistency
• information loss

Innovation

• Generative Recall
• Multi-Objective Ranking
• Decoder-Only Transformer
• Dual-Query Prefix-Causal Attention
• LoRA