Predicting Catastrophes: The Role of Criticality
This event is part of the Departmental Seminars.
William Klein, Harvey Gould, Plamen Ivanov, Alex Sushkov and Chris Grant.
Is prediction feasible in systems at criticality? While conventional scale-invariant argument suggests a negative answer, evidences from simulation of driven-dissipative systems and real systems such as ruptures in material and crashes in the financial market have suggested otherwise.
In this seminar, I will address the question of predictability at criticality by investigating a non-equilibrium system, a driven-dissipative system called the OFC model which is used to describe earthquake faults, and a non-equilibrium process, damage spreading in the Ising model. Both systems display a phase transition at the critical point. Using machine learning, I show that in the OFC model, scaling events are indistinguishable from one another and only the large, non-scaling events are distinguishable from the small, scaling events. This implies that large earthquakes (catastrophes) off the Gutenberg-Richter scaling are distinguishable from the smaller earthquakes. I also show that as the critical point is approached, predictability falls. For damage spreading in the Ising model, an opposite behavior is seen: the accuracy of predicting whether damage will spread indefinitely (catastrophe) or heal increases as the critical point is approached. I will also use machine learning to identify useful precursors to the 'catastrophic' events.