Community Question Answering on SemEvalCQA (test)
80.1MAPAI-CNN + Feats
Evaluation results
| Method | Links | ||
|---|---|---|---|
| AI-CNN + Featsnumber of parameters=140K, training time per epoch=3250s, external features=true2017.07 | 80.1 | 76.9 | |
| HyperQAnumber of parameters=45K, training time per epoch=10s2017.07 | 79.5 | 80.9 | |
| Kelp2017.07 | 79.2 | 75.1 | |
| AI-CNNnumber of parameters=140K, training time per epoch=3250s2017.07 | 79.2 | 76.3 | |
| ARC-II2017.07 | 78 | 75.3 | |
| ConvKN2017.07 | 77.7 | 75.5 | |
| ARC-I2017.07 | 77.1 | 74.1 | |
| AP-CNN2017.07 | 77.1 | 75.5 | |
| KeLPSubmission Type=contrastive12018.04 | 49 | — | |
| SimBowSubmission Type=contrastive22018.04 | 47.87 | — | |
| Reddit+SNLITraining Data=Reddit conversations + SNLI, Task Specific Training=None2018.04 | 47.42 | — | |
| SimBowSubmission Type=primary2018.04 | 47.22 | — | |
| RedditTraining Data=Reddit conversations, Task Specific Training=None2018.04 | 47.07 | — |