Multi-Granular Sequence Encoding via Dilated Compositional Units for Reading Comprehension
Open paper
Benchmark
Model
Rank
Results
open-domain-question-answering-on-searchqa
Bi-Attention + DCU-LSTM
–
N-gram F1: 59.5
Unigram Acc: 49.4
question-answering-on-narrativeqa
BiAttention + DCU-LSTM
–
Rouge-L: 41.44
BLEU-1: 36.55
BLEU-4: 19.79
METEOR: 17.87
Rank counts only results with a code link.