All news

Digital twins of respondents distilled from survey histories into compact personas

The study asks whether long survey histories can be compressed into compact, task-specific personas that stay predictive for unseen respondents. An iterative pipeline builds persona representations and learns how persona evidence shapes simulated responses. It is evaluated on the Product Preferences-Pricing task of Twin-2K-500, training on 20 respondents and 20 products and testing on 30 unseen respondents and 20 unseen products.

Why it matters

The method offers a measurable way to build behavioral models of agents and validate them on unseen data, which matters for agent-based modeling of large systems.

Relevant to DT products

Original headline
From Survey Histories to Interpretable Digital Twins: Auto-Discovering Task-Specific Personas and Behavioral Mechanisms
Read the original: arXiv — Computers and Society (cs.CY)

Translation and summary are machine-generated from the source. The full article is not republished; its rights belong to the source.