Models trained on large data sets of seismic events can estimate the number of aftershocks better than conventional models do.
Models trained on large data sets of seismic events can estimate the number of aftershocks better than conventional models do.
You think adding a few words about how much much more precise they are would hurt? Someone should need to go to the sources to find out how this works exactly or how they designed it or whatever, but simply stating the actual improvement should be part of the article.