NextGen: A Multi-Objective Generative Re-ranking Framework for Taobao Recommendation

Abstract

Reranking is a critical stage in e-commerce recommendation that reorders items to optimize the final list as a whole. Existing generative reranking methods suffer from two limitations: (1) they rely on next-item exposure prediction as supervision, neglecting page-level multi-objective signals such as total clicks or transactions; (2) autoregressive decoding incurs prohibitive latency, making full-candidate scoring infeasible under real-time constraints. We propose NextGen, a generative reranking framework for page-aware multi-objective recommendation under strict latency budgets. NextGen introduces three key designs: (1) a full-candidate autoregressive generative framework via a global context-aware Encoder-Decoder architecture that jointly optimizes page-level multiple objectives while autoregressively selecting items across the entire candidate set, rigorously constraining inference latency for real-time deployment; (2) a residual connection scheme for incremental reranking that feeds upstream ranking scores into both the item embedding layer and the candidate selection module, so the model only needs to predict nonlinear gains relative to ranking scores, improving convergence stability; (3) LLM-enhanced semantic embeddings from Taobao’s multimodal foundation model that enrich item representations with domain knowledge at no additional inference cost. Online A/B tests on Taobao Miaosha show that NextGen achieves +18.85% GMV and +12.35% order volume over the strongest baseline. The system is deployed in production, serving hundreds of millions of users daily.