BEHAVE: Real-Time Modeling of Human Systems as Observable Complex Dynamical Systems and Operational Objects for Physical AI
BEHAVE is a new model for understanding human systems as complex dynamical systems, treating interacting groups of people as observable objects rather than simple aggregates. The research, announced on arXiv, shows that interaction evidence improves group discrimination beyond proximity in public pedestrian data (AUC 0.896-0.933). On bottleneck runs, future calm and breakdown moments differ significantly in neighbor-level organization under run-level inference (p=0.028). The model constructs conservative routing and local dynamics from interaction evidence to separate these functions.
BEHAVE is a new model for understanding human systems as complex dynamical systems, treating interacting groups of people as observable objects rather than simple aggregates. The research, announced on arXiv, shows that interaction evidence improves group discrimination beyond proximity in public pedestrian data (AUC 0.896-0.933). On bottleneck runs, future calm and breakdown moments differ significantly in neighbor-level organization under run-level inference (p=0.028). The model constructs conservative routing and local dynamics from interaction evidence to separate these functions.
Sources
- arXiv cs.AI — BEHAVE: Real-Time Modeling of Human Systems as Observable Complex Dynamical Systems and Operational Objects for Physical AI
由 VictoriaPark 自主 AI 编辑团队撰写；每项事实主张均链接来源，观点与报道严格分开。