QA Answer Selection on TREC Answer Selection (test)
0.875MAPComp-Clip + LM + LC + TL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Comp-Clip + LM + LC + TLLearning approach=Pointwise, Implementation=Ours2019.05 | 0.875 | 0.94 | |
| Comp-Clip + LM + LCLearning approach=Pointwise, Implementation=Ours2019.05 | 0.868 | 0.928 | |
| Question ClassificationImplementation=Reported2019.05 | 0.865 | 0.904 | |
| Comp-Clip + LMLearning approach=Pointwise, Implementation=Ours2019.05 | 0.85 | 0.898 | |
| Comp-Clip + LM + LC + TLLearning approach=Listwise, Implementation=Ours2019.05 | 0.848 | 0.902 | |
| MCANImplementation=Reported2019.05 | 0.838 | 0.904 | |
| Comp-ClipLearning approach=Pointwise, Implementation=Ours2019.05 | 0.835 | 0.877 | |
| Comp-Clip + LM + LCLearning approach=Listwise, Implementation=Ours2019.05 | 0.832 | 0.88 | |
| IWAN + SCARNNImplementation=Reported2019.05 | 0.829 | 0.875 | |
| Comp-Clip + LMLearning approach=Listwise, Implementation=Ours2019.05 | 0.823 | 0.868 | |
| IWANImplementation=Reported2019.05 | 0.822 | 0.899 | |
| Comp-ClipImplementation=Reported2019.05 | 0.821 | 0.899 | |
| AP-CNNarchitecture=CNN, pooling=Attentive Pooling2016.02 | 0.753 | 0.8511 | |
| Severyn & Moschitti (2015)2016.02 | 0.746 | 0.808 | |
| Wang & Ittycheriah (2015)2016.02 | 0.746 | 0.82 | |
| Comp-ClipLearning approach=Listwise, Implementation=Ours2019.05 | 0.744 | 0.791 | |
| QA-CNNarchitecture=CNN2016.02 | 0.7147 | 0.807 | |
| Wang & Nyberg (2015)2016.02 | 0.7134 | 0.7913 | |
| AP-BILSTMarchitecture=biLSTM, pooling=Attentive Pooling2016.02 | 0.7132 | 0.8032 | |
| bigram + countTraining Set=TRAIN-ALL2014.12 | 0.7113 | 0.7846 | |
| Yih et al.Year=2013, Classifier=LCLR2014.12 | 0.7092 | 0.77 | |
| bigram + countTraining Set=TRAIN2014.12 | 0.7058 | 0.78 | |
| Yih et al.Year=2013, Classifier=BDT2014.12 | 0.694 | 0.7894 | |
| Yih et al.Year=2013, Classifier=LR2014.12 | 0.6818 | 0.7616 | |
| Severyn and MoschittiYear=20132014.12 | 0.6781 | 0.7358 | |
| QA-BILSTMarchitecture=biLSTM2016.02 | 0.675 | 0.7723 | |
| Yao et al.Year=20132014.12 | 0.6307 | 0.7477 | |
| Heilman and SmithYear=20102014.12 | 0.6091 | 0.6917 | |
| Wang et al.Year=20072014.12 | 0.6029 | 0.6852 | |
| Wgt Word Count2014.12 | 0.5961 | 0.6515 | |
| Wang and ManningYear=20102014.12 | 0.5951 | 0.6951 | |
| Word Count2014.12 | 0.5707 | 0.6266 | |
| Random2014.12 | 0.3965 | 0.4929 |