Hierarchical Multi-Persona Induction from User Behavioral Logs: Learning Evidence-Grounded and Truthful Personas

Hierarchical Multi-Persona Induction from User Behavioral Logs: Learning Evidence-Grounded and Truthful Personas

从用户行为日志中进行分层多角色归纳:学习基于证据且真实的角色模型

Abstract: Behavioral logs provide rich signals for user modeling, but are noisy and interleaved across diverse intents. Recent work uses LLMs to generate interpretable natural-language personas from user logs, yet evaluation often emphasizes downstream utility, providing limited assurance of persona quality itself. 摘要: 行为日志为用户建模提供了丰富的信号,但这些数据往往存在噪声,且交织着多种不同的意图。近期的研究利用大语言模型(LLMs)从用户日志中生成可解释的自然语言角色,然而目前的评估往往侧重于下游任务的效用,对角色模型本身的质量保障有限。

We propose a hierarchical framework that aggregates user actions into intent memories and induces multiple evidence-grounded personas by clustering and labeling these memories. 我们提出了一种分层框架,将用户行为聚合为意图记忆,并通过对这些记忆进行聚类和标注,归纳出多个基于证据的角色模型。

We formulate persona induction as an optimization problem over persona quality—captured by cluster cohesion, persona-evidence alignment, and persona truthfulness—and train the persona model using a groupwise extension of Direct Preference Optimization (DPO). 我们将角色归纳建模为一个关于角色质量的优化问题,通过聚类内聚性、角色与证据的一致性以及角色真实性来衡量质量,并使用直接偏好优化(DPO)的组级扩展来训练该角色模型。

Experiments on a large-scale service log and two public datasets show that our method induces more coherent, evidence-grounded, and trustworthy personas, while also improving future interaction prediction. 在大型服务日志和两个公共数据集上的实验表明,我们的方法能够归纳出更连贯、基于证据且更可信的角色,同时还能提升对未来交互行为的预测能力。