ProphetNet + syn. mask + localness

Reported on 3 benchmarks across 1 task · 1 paper · 1 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing3 results

Question GenerationonSQuAD1.1
ROUGE-L· 2022-09-09
52.77
best: 52.8 (ProphetNet + ASGen)
SOTA
Enhancing Pre-trained Models with Text Structure Knowledge for Question Generation arXiv:2209.04179
Question GenerationonSQuAD1.1
BLEU-4· 2022-09-09
24.37
best: 25.41 (ERNIE-GENLARGE (beam size=5))
Enhancing Pre-trained Models with Text Structure Knowledge for Question Generation arXiv:2209.04179
Question GenerationonSQuAD1.1
METEOR· 2022-09-09
26.26
best: 26.73 (BART (TextBox 2.0))
Enhancing Pre-trained Models with Text Structure Knowledge for Question Generation arXiv:2209.04179