Sociohydrodynamics: data-driven modelling of social behavior

arxiv(2023)

引用 0|浏览4
暂无评分
摘要
Living systems display complex behaviors driven not only by physical forces, but also decision-making guided by information processing and molded by cultural and/or biological evolution. Hydrodynamic theories hold promise for simplified, universal descriptions of these collective behaviors. However, incorporating the individual preferences of decision-making organisms into a hydrodynamic theory is an open problem. Here, we develop a data-driven pipeline that links micromotives to macrobehavior by augmenting hydrodynamics with utility functions that describe individual preferences in microeconomics. We show how to systematically validate the hypotheses underlying this construction from data using statistical tools based on neural networks. We illustrate this pipeline on the case study of human residential dynamics in the United States, for which census and sociological data is available, and show how trends in sociological surveys can be related to trends seen in racial segregation. In particular, we highlight that a history-dependence in the segregation-integration transition can arise even when agents have no memory. Beyond residential segregation, our work paves the way for systematic investigations of social-driven motility in real space from micro-organisms to humans, as well as fitness-mediated motion in more abstract genomic spaces.
更多
查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要