arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.00700cs.IR
收藏

SODA: 基于语义的分布对齐用于生成式推荐

Distribution-Level Contrastive Supervision for Generative Recommendation

University of Science and Technology of China(中国科学技术大学);Upwork;The Chinese University of Hong Kong(香港中文大学);National University of Singapore(新加坡国立大学)

Ziqi Xue, Dingxian Wang, Yimeng Bai, Shuai Zhu, Jialei Li, Xiaoyan Zhao, Frank Yang, Andrew Rabinovich, Yang Zhang, Pablo N. Mendes

AI总结:

SODA通过分布层面的监督和对比学习,提升生成式推荐系统的性能,实现端到端可微训练。

摘要:

生成式推荐作为一种可扩展的替代传统检索-排序流程的方法,通过在紧凑的token空间中操作得以实现。然而,现有方法主要依赖于离散代码层面的监督,导致信息丢失,并限制了分词器与生成式推荐器之间的联合优化。在本文中,我们提出了一种分布层面的监督范式,利用多层代码本上的概率分布作为软而信息丰富的表示。基于这一想法,我们引入了语义导向的分布对齐(SODA),一种基于贝叶斯个性化排序的即插即用对比监督框架,通过负KL散度对语义丰富的分布进行对齐,同时实现端到端可微训练。在多个真实世界数据集上的广泛实验表明,SODA一致提升了各种生成式推荐器backbone的性能,验证了其有效性和通用性。代码将在接受后提供。

英文摘要:

Recent generative recommenders improve scalability by retrieving items through token generation instead of traditional ranking over large candidate sets. Yet their training signals are still dominated by discrete code prediction, which overlooks the soft assignment information naturally produced by the tokenizer. This mismatch limits semantic transfer from the tokenizer to the recommender and may hurt overall optimization. We tackle this limitation by introducing a distribution-based supervision scheme for generative recommendation, where multi-level codebook probabilities are treated as soft semantic targets. On top of this design, we develop SODA, a plug-and-play alignment framework that adopts a BPR-style contrastive objective to align recommender representations with target-side distributional representations against negative ones. The proposed method enriches training with finer semantic cues while leaving the decoding stage unchanged. Experimental studies on multiple real-world benchmarks demonstrate that SODA consistently strengthens diverse generative recommendation architectures. Code is available at https://github.com/freyasa/SODA

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑