Using Deep Learning to Train Deep Learning Models

A lot of times when we're getting this data, we're making some sort of bed partnership with like a hospital system. And then often the four looking hospitals, like, say, the sloane chatterings of the world, they're really focused. Slone kettering is really focused on the corner case things like that. So could be, do these corner cases also sewing the data case versus average cases? I think that's a part of it. If your data is significantly biased, that would cost problems. But more just volume. And can ye just not a lot of data. You had every radio stand ever done in the last 30 years, coupled with the patient medical record