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 2604.07513cs.LGcs.AIcs.CLcs.CY
收藏

SYN-DIGITS: 一种合成控制框架用于校准数字孪生模拟

SYN-DIGITS: A Synthetic Control Framework for Calibrated Digital Twin Simulation

Finance Division, Columbia Business School(哥伦比亚商学院金融系);Department of IEOR, Columbia University(哥伦比亚大学工业工程与运筹学系);Decision, Risk, and Operations Division, Columbia Business School(哥伦比亚商学院决策、风险与运营系);Department of IEOR and Data Science Institute, Columbia University(哥伦比亚大学工业工程与运筹学系和数据科学研究所)

Grace Jiarui Fan, Chengpiao Huang, Tianyi Peng, Kaizheng Wang, Yuhang Wu

AI总结:

本文提出SYN-DIGITS框架,通过学习数字孪生响应的潜在结构,校准大语言模型预测与人类行为的一致性,提升市场研究和社交科学中的模拟可靠性。

摘要:

基于人工智能的人格模拟(常称为数字孪生模拟)在市场研究、推荐系统和社会科学中日益普及。尽管其灵活性高,但大型语言模型(LLMs)往往表现出系统性偏差和校准不足,限制了其可靠性。受因果推理中的合成控制方法启发,我们提出了SYN-DIGITS(SYNthetic Control Framework for Calibrated DIGItal Twin Simulation),一种原理明确且轻量的校准框架,通过学习数字孪生响应的潜在结构,并将其转移到与人类真实行为一致。SYN-DIGITS作为任何基于LLM的模拟器的后处理层,因此具有模型无关性。我们开发了一个潜在因子模型,正式化了校准成功的时间和原因,通过潜在空间对齐条件。我们系统评估了十种校准方法,涵盖十三种人格构建、三种LLMs和两个数据集。SYN-DIGITS支持对以前未见问题和未观测人口的个体级和分布级模拟,具有可证明的误差保证。实验表明,与未校准基线相比,SYN-DIGITS在个体级相关性上提高了高达50%,在分布级差异上减少了50-90%。

英文摘要:

Persona simulation based on large language models---often referred to as digital twin simulation---is increasingly used for market research, recommender systems, and social science research. However, these simulations often exhibit systematic bias and miscalibration relative to real human behavior. We propose SYN-DIGITS, a model-agnostic, post-hoc calibration framework that learns relationships among questions from digital twins and transfers them to predict human responses. We focus on a cold-start setting in which no human response to the target question is observed, so calibration must rely on digital twins and human data on reference questions. Our analysis shows that, though digital twins may not reproduce individual human responses accurately, they can capture nontrivial relationships among questions that can be effectively transferred to the human side. When individual-level data is unavailable and we only have access to marginal information on reference questions, our framework also supports distribution-level calibration for predicting the distribution of human responses on the target question. Across three datasets and thirteen persona constructions, experiments show that SYN-DIGITS achieves 11% to 417% relative improvements in individual-level correlation and 46%--57% relative reductions in distributional discrepancy compared to uncalibrated baselines.

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。

↑