applied-systems

An early-warning system that models disasters as instability transitions rather than isolated events. It combines force-based instability modeling with an interpretable ML escalation-risk layer to detect when hazards become disasters due to exposure growth, response delays, and buffer collapse.