Johnson's team are working on turning lab results into the kind of predictive power geologists have been looking for. They're going to build a stigil simulation of a fault and then mix in real world data by training their system on it. One particular target he mentioned was the park field quake, which was quite a large event. So they 're going to see whether, given thistype of training, their system could have predicted that. If they can do that, then they know they will have hit gold.