3 ms·
One thing I left out: the paper compares these results against prior published methods on the same metrics, and we come out ahead of all of them by a few points
by jtewright 4mo ago
One thing I left out: the paper compares these results against prior published methods on the same metrics, and we come out ahead of all of them by a few points. However it's on our own population rather than a shared benchmark, so it's not apples to apples.