1 paper · 1 filter
Lars van der Laan, Mark Van Der Laan
We study semisupervised mean estimation with a small labeled sample, a large unlabeled sample, and a black-box prediction model whose output may be miscalibrated. A standard approa…