Optimizing Generative Ranking Relevance via Reinforcement Learning in Xiaohongshu Search

TL;DR

This paper introduces a reinforcement learning framework to improve generative relevance models in Xiaohongshu search by grounding reasoning in business-specific criteria, leading to better relevance and business outcomes.

Contribution

It proposes a novel RL-based training method with Stepwise Advantage Masking for relevance modeling, enhancing interpretability and performance in industrial search systems.

Findings

Significant improvements in relevance metrics

Enhanced robustness and interpretability

Effective model distillation for deployment

Abstract

Ranking relevance is a fundamental task in search engines, aiming to identify the items most relevant to a given user query. Traditional relevance models typically produce scalar scores or directly predict relevance labels, limiting both interpretability and the modeling of complex relevance signals. Inspired by recent advances in Chain-of-Thought (CoT) reasoning for complex tasks, we investigate whether explicit reasoning can enhance both interpretability and performance in relevance modeling. However, existing reasoning-based Generative Relevance Models (GRMs) primarily rely on supervised fine-tuning on large amounts of human-annotated or synthetic CoT data, which often leads to limited generalization. Moreover, domain-agnostic, free-form reasoning tends to be overly generic and insufficiently grounded, limiting its potential to handle the diverse and ambiguous cases prevalent in open-domain search. In this work, we formulate relevance modeling in Xiaohongshu search as a reasoning task and introduce a Reinforcement Learning (RL)-based training framework to enhance the grounded reasoning capabilities of GRMs. Specifically, we incorporate practical business-specific relevance criteria into the multi-step reasoning prompt design and propose Stepwise Advantage Masking (SAM), a lightweight process-supervision strategy which facilitates effective learning of these criteria through improved credit assignment. To enable industrial deployment, we further distill the large-scale RL-tuned model to a lightweight version suitable for real-world search systems. Extensive offline evaluations and online A/B tests demonstrate that our approach consistently delivers significant improvements across key relevance and business metrics, validating its effectiveness, robustness, and practicality for large-scale industrial search systems.