1 paper
Julian Risch, Timo Möller, Julian Gutsch +1
The evaluation of question answering models compares ground-truth annotations with model predictions. However, as of today, this comparison is mostly lexical-based and therefore mi…