Search papers, labs, and topics across Lattice.
1
0
2
1
Modeling annotator disagreement in discourse relation recognition is more effective with label distributions than with annotator-specific models, challenging the assumption that perspectivist models are universally superior for ambiguous NLP tasks.
