arXiv:2603.28066cs.HCcs.AI2026-03

构建群体人物画像,让生成式模拟兼具个体真实与集体洞察。

Synonymix: Unified Group Personas for Generative Simulations

  • 用图抽象合并多个个人故事,生成可查询的群体表征
  • 合成角色在社会调查中行为信号保留率超基线(r=0.59)
  • 适合做群体行为研究与隐私保护型模拟实验

生成式代理模拟目前存在于两个层面:个体人物画像用于角色互动,群体模型用于集体行为分析与干预测试。本文提出第三层次——介观级模拟:通过群体层级表征进行交互,同时保持对丰富个体经验的锚定。为此,我们提出 Synonymix 流水线,利用基于图的抽象与合并技术,从多个生活故事人物画像构建统一图谱(unigraph),生成可查询的集体表示,可用于认知探索或合成人物生成。在通用社会调查项目上的评估表明,合成代理的行为信号保留优于人口统计基线(p<0.001, r=0.59),且具备可证明的隐私保障(最大来源贡献<13%)。本文邀请探讨介观级模拟带来的新交互方式,以及高保真人物画像是否可能真正捕捉生活质感。

原文摘要 · Abstract (English)

Generative agent simulations operate at two scales: individual personas for character interaction, and population models for collective behavior analysis and intervention testing. We propose a third scale: meso-level simulation - interaction with group-level representations that retain grounding in rich individual experience. To enable this, we present Synonymix, a pipeline that constructs a "unigraph" from multiple life story personas via graph-based abstraction and merging, producing a queryable collective representation that can be explored for sensemaking or sampled for synthetic persona generation. Evaluating synthetic agents on General Social Survey items, we demonstrate behavioral signal preservation beyond demographic baselines (p<0.001, r=0.59) with demonstrable privacy guarantee (max source contribution <13%). We invite discussion on interaction modalities enabled by meso-level simulations, and whether "high-fidelity" personas can ever capture the texture of lived experience.

生成模拟群体建模隐私保护

Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品;中文卡片由大模型生成,请以原文为准。