You Only Evaluate Once: A Tree-based Rerank Method at Meituan

TL;DR

YOLOR is a novel one-stage reranking method for recommender systems that eliminates the traditional two-stage process, improving effectiveness and efficiency by using hierarchical context aggregation and feature reuse, and has been successfully deployed in industry.

Contribution

YOLOR introduces a one-stage reranking approach that removes the GSU, utilizing hierarchical context extraction and feature caching for better effectiveness and efficiency.

Findings

YOLOR outperforms traditional two-stage methods on public and industry datasets.

YOLOR achieves list-level effectiveness and permutation-level efficiency.

Successfully deployed in Meituan's food delivery platform.

Abstract

Reranking plays a crucial role in modern recommender systems by capturing the mutual influences within the list. Due to the inherent challenges of combinatorial search spaces, most methods adopt a two-stage search paradigm: a simple General Search Unit (GSU) efficiently reduces the candidate space, and an Exact Search Unit (ESU) effectively selects the optimal sequence. These methods essentially involve making trade-offs between effectiveness and efficiency, while suffering from a severe \textbf{inconsistency problem}, that is, the GSU often misses high-value lists from ESU. To address this problem, we propose YOLOR, a one-stage reranking method that removes the GSU while retaining only the ESU. Specifically, YOLOR includes: (1) a Tree-based Context Extraction Module (TCEM) that hierarchically aggregates multi-scale contextual features to achieve "list-level effectiveness", and (2) a Context Cache Module (CCM) that enables efficient feature reuse across candidate permutations to achieve "permutation-level efficiency". Extensive experiments across public and industry datasets validate YOLOR's performance, and we have successfully deployed YOLOR on the Meituan food delivery platform.