1 paper
Bijan Mazaheri, Siddharth Jain, Matthew Cook +1
We explore "omitted label contexts," in which training data is limited to a subset of the possible labels. This setting is standard among specialized human experts or specific, foc…