1 paper · 1 filter
Jinghan Jia, Abi Komma, Timothy Leffel +5
In task-oriented conversational AI evaluation, unsupervised methods poorly correlate with human judgments, and supervised approaches lack generalization. Recent advances in large l…