Question Answering On Wikiqa

MAP

MRR

评测结果

各个模型在此基准测试上的表现结果

比较表格

模型名称	MAP	MRR
distributed-representations-of-sentences-and	0.5110	0.5160
pre-training-transformer-models-with-sentence	0.901	0.914
hyperbolic-representation-learning-for-fast	0.712	0.727
pairwise-word-interaction-modeling-with-deep	0.7090	0.7234
distributed-representations-of-sentences-and	0.5976	0.6058
baseline-needs-more-love-on-simple-word	0.6788	0.6908
neural-variational-inference-for-text	0.682	0.6988
deep-learning-for-answer-sentence-selection	0.6520	0.6652
tanda-transfer-and-adapt-pre-trained	0.920	0.933
simple-and-effective-text-matching-with-1	0.7452	0.7618
noise-contrastive-estimation-and-negative	0.7010	0.7180
pre-training-transformer-models-with-sentence	0.887	0.899
neural-variational-inference-for-text	0.6552	0.6747
attentive-pooling-networks	0.6886	0.6957
wikiqa-a-challenge-dataset-for-open-domain	0.6520	0.6652
deep-learning-for-answer-sentence-selection	0.6190	0.6281
rlas-biabc-a-reinforcement-learning-based	0.924	0.908
neural-semantic-encoders	0.6811	0.6993
sentence-similarity-learning-by-lexical	0.7058	0.7226
a-compare-aggregate-model-with-latent	0.764	0.784
neural-variational-inference-for-text	0.6886	0.7069
key-value-memory-networks-for-directly	0.7069	0.7265
paragraph-based-transformer-pre-training-for	0.887	0.900
pre-training-transformer-models-with-sentence	0.909	0.920
structural-self-supervised-objectives-for	0.927	0.939