arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.04858cs.CLcs.AIcs.IRcs.LG
收藏

用于个性化语言提示的用户嵌入模型

User Embedding Model for Personalized Language Prompting

Google Research(谷歌研究院)

Sumanth Doddapaneni, Krishna Sayana, Ambarish Jash, Sukhdeep Sodhi, Dima Kuzmin

AI总结:

本研究提出 User Embedding Module (UEM),将长用户历史压缩为嵌入作为语言模型的软提示,有效提升个性化推荐的预测性能并支持更长历史处理。

摘要:

建模长历史在增强推荐系统中起着关键作用,它能够捕捉用户不断变化的偏好,从而实现更精确和个性化的推荐。在本研究中,我们致力于解决在自然语言中进行偏好理解时建模长用户历史的挑战。具体而言,我们引入了一个新的 User Embedding Module (UEM),它通过将自由格式的用户历史压缩并表示为嵌入,来高效处理这些历史,并将其作为 LM 的软提示。我们的实验表明,与传统的基于文本的提示方法相比,该方法在处理明显更长的历史方面具有卓越的能力,在预测性能上取得了显著提升。本研究的主要贡献是证明了利用表示为嵌入的用户信号来引导语言模型的能力。

英文摘要:

Modeling long histories plays a pivotal role in enhancing recommendation systems, allowing to capture user's evolving preferences, resulting in more precise and personalized recommendations. In this study we tackle the challenges of modeling long user histories for preference understanding in natural language. Specifically, we introduce a new User Embedding Module (UEM) that efficiently processes user history in free-form text by compressing and representing them as embeddings, to use them as soft prompts to a LM. Our experiments demonstrate the superior capability of this approach in handling significantly longer histories compared to conventional text based prompting methods, yielding substantial improvements in predictive performance. The main contribution of this research is to demonstrate the ability to bias language models with user signals represented as embeddings.

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。

↑