Machine Comprehension on MMQA 5.2 v1.0 (test)
0.7443F1 (QE-PE)m-BERT based MMC
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| m-BERT based MMCExperimental Setting=with QH – PH Aug., Data Augmentation=mono-lingual Hindi QA augmentation2020.06 | 0.7443 | 0.6139 | 0.6506 | 0.6434 | |
| m-BERT based MMCExperimental Setting=with QE – PH Aug., Data Augmentation=cross-lingual QA augmentation2020.06 | 0.7326 | 0.6375 | 0.6437 | 0.6381 | |
| m-BERT based MMCExperimental Setting=Zero Shot2020.06 | 0.7195 | 0.5151 | 0.5233 | 0.5097 | |
| Benchmark (Gupta et al. 2019)Description=Reported scores in Gupta et al. 20192020.06 | 0.5027 | 0.3789 | 0.3702 | 0.4814 |