S-GRec: Personalized Semantic-Aware Generative Recommendation with Asymmetric Advantage

TL;DR

S-GRec is a recommendation framework that leverages semantic signals from large language models through a decoupled, two-stage process, improving recommendation quality while maintaining scalability and alignment with business goals.

Contribution

The paper introduces a novel decoupled architecture with a semantic judge and an asymmetric policy optimization to incorporate semantic signals effectively in industrial recommendation systems.

Findings

Significant CTR improvements in benchmarks

1.19% GMV lift in online A/B tests

Effective integration of semantic signals without real-time LLM inference

Abstract

Generative recommendation models sequence generation to produce items end-to-end, but training from behavioral logs often provides weak supervision on underlying user intent. Although Large Language Models (LLMs) offer rich semantic priors that could supply such supervision, direct adoption in industrial recommendation is hindered by two obstacles: semantic signals can conflict with platform business objectives, and LLM inference is prohibitively expensive at scale. This paper presents S-GRec, a semantic-aware framework that decouples an online lightweight generator from an offline LLM-based semantic judge for train-time supervision. S-GRec introduces a two-stage Personalized Semantic Judge (PSJ) that produces interpretable aspect evidence and learns user-conditional aggregation from pairwise feedback, yielding stable semantic rewards. To prevent semantic supervision from deviating from business goals, Asymmetric Advantage Policy Optimization (A2PO) anchors optimization on business rewards (e.g., eCPM) and injects semantic advantages only when they are consistent. Extensive experiments on public benchmarks and a large-scale production system validate both effectiveness and scalability, including statistically significant gains in CTR and a 1.19\% lift in GMV in online A/B tests, without requiring real-time LLM inference.