With the evolution of large language models (LLMs), there is growing interest in leveraging their rich semantic understanding to enhance industrial recommendation systems (RecSys). Traditional RecSys relies on ID-based embeddings for user sequence modeling in the General Search Unit (GSU) and Exact Search Unit (ESU) paradigm, which suffers from low information density, knowledge isolation, and weak generalization ability. While LLMs offer complementary strengths with dense semantic representations and strong generalization, directly applying LLM embeddings to RecSys faces critical challenges: representation unmatch with business objectives and representation unlearning end-to-end with downstream tasks. In this paper, we present QARM V2, a unified framework that bridges LLM semantic understanding with RecSys business requirements for user sequence modeling.
@article{arxiv.2602.08559,
title = {QARM V2: Quantitative Alignment Multi-Modal Recommendation for Reasoning User Sequence Modeling},
author = {Tian Xia and Jiaqi Zhang and Yueyang Liu and Hongjian Dou and Tingya Yin and Jiangxia Cao and Xulei Liang and Tianlu Xie and Lihao Liu and Xiang Chen and Shen Wang and Changxin Lao and Haixiang Gan and Jinkai Yu and Keting Cen and Lu Hao and Xu Zhang and Qiqiang Zhong and Zhongbo Sun and Yiyu Wang and Shuang Yang and Mingxin Wen and Xiangyu Wu and Shaoguo Liu and Tingting Gao and Zhaojie Liu and Han Li and Kun Gai},
journal= {arXiv preprint arXiv:2602.08559},
year = {2026}
}