NLGR: Utilizing Neighbor Lists for Generative Rerank in Personalized Recommendation Systems

2025-02-10

AI Summary

In multi-stage recommendation, re-ranking faces two key challenges: (1) misalignment between the evaluator and generator objectives, leading to suboptimal local solutions; and (2) autoregressive item-by-item generation, which neglects global combinatorial optimization. To address these, we propose the Neighborhood List Generative Re-ranking (NLGR) model. Its core contributions are: (1) a neighborhood list augmentation training strategy that explicitly models local candidate structure to mitigate objective mismatch; and (2) a sampling-based non-autoregressive generation mechanism enabling cross-position jumps and parallel combinatorial search. Extensive experiments on multiple public and industrial datasets demonstrate significant improvements in NDCG and other metrics. NLGR has been deployed in Meituan Waimai’s production system, achieving an 18% reduction in real-time re-ranking latency and a 2.3% increase in GMV—validating both the effectiveness of combinatorial optimization modeling and its practical deployment value.

Technology Category

• Machine Learning: Learning Preferences or Rankings
• Natural Language Processing: Generation
• Search and Optimization: Algorithm Configuration

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Reranking plays a crucial role in modern multi-stage recommender systems by rearranging the initial ranking list. Due to the inherent challenges of combinatorial search spaces, some current research adopts an evaluator-generator paradigm, with a generator generating feasible sequences and an evaluator selecting the best sequence based on the estimated list utility. However, these methods still face two issues. Firstly, due to the goal inconsistency problem between the evaluator and generator, the generator tends to fit the local optimal solution of exposure distribution rather than combinatorial space optimization. Secondly, the strategy of generating target items one by one is difficult to achieve optimality because it ignores the information of subsequent items. To address these issues, we propose a utilizing Neighbor Lists model for Generative Reranking (NLGR), which aims to improve the performance of the generator in the combinatorial space. NLGR follows the evaluator-generator paradigm and improves the generator's training and generating methods. Specifically, we use neighbor lists in combination space to enhance the training process, making the generator perceive the relative scores and find the optimization direction. Furthermore, we propose a novel sampling-based non-autoregressive generation method, which allows the generator to jump flexibly from the current list to any neighbor list. Extensive experiments on public and industrial datasets validate NLGR's effectiveness and we have successfully deployed NLGR on the Meituan food delivery platform.

Problem

• Addresses goal inconsistency in recommender evaluator-generator systems.
• Improves combinatorial space optimization in reranking algorithms.
• Introduces non-autoregressive generation for flexible sequence generation.

Innovation

• Neighbor Lists for Reranking
• Non-autoregressive Generation Method
• Enhanced Combinatorial Space Training