Predicting Earthquakes with AI
This is a problem I've been intermittently looking at for a few years now. I'm strongly convinced it is feasible to get useful notification ahead of dangerous seismic events...with some major caveats.
There is a lot of reliable, historic and real-time seismic data available. It's typically expressed as motion over time of points on the surface of the planet. It's possible to correlate between data from different monitoring locations to pinpoint with reasonable accuracy the source of particular sets of measurements, after the fact.
AI tools have, in recent years, got very good at making predictions about the behaviour of physical systems over time. Synthetic videos can be generated that convincingly show apparent real-world events.
But the system in which earthquakes occur is massive, extremely complex and chaotic.
Radio Precursors
of the behaviour of multidimensional time series data
But t
to be in condition to receive radio seismic precursors it takes very strong earthquakes and we must be very close to them