1 paper · 2 filters
Aryeh Kontorovich, Ariel Avital
We revisit the classic problem of aggregating binary advice from conditionally independent experts, also known as the Naive Bayes setting. Our quantity of interest is the error pro…