Sarcasm Explanation Generation on MORE OCR samples (test)
52.88ROUGE-1 PrecisionTEAM
Evaluation Results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TEAMBackbone=bart-base-chinese, Knowledge=Full2023.06 | 52.88 | 43.08 | 36.81 | 32.34 | 48.46 | 49.68 | 33.83 | 49.25 | 90.9 | 90 | 90.8 | 71.93 | |
| TEAM-w/o-KnowBackbone=bart-base-chinese, Knowledge=None2023.06 | 51.91 | 41.51 | 34.85 | 29.85 | 47.53 | 49 | 32.77 | 47.94 | 90.5 | 91 | 90.7 | 71.43 | |
| ExMore2023.06 | 19.4 | 11.31 | 6.83 | 4.76 | 25.66 | 28.02 | 12.1 | 19.15 | 88.2 | 87.5 | 87.9 | 60.82 | |
| PGN2023.06 | 17.19 | 6.08 | 2.49 | 1.79 | 15.55 | 16.92 | 6.76 | 14.64 | 84.9 | 84.9 | 84.9 | 49.53 | |
| M-Transf2023.06 | 14.06 | 6.25 | 3.22 | 2.28 | 18.42 | 21.04 | 7.01 | 12.06 | 86.2 | 86.1 | 86.1 | 55.66 | |
| MFFG-TransfArchitecture=Transformer-based2023.06 | 12.87 | 4.12 | 1.69 | 0.62 | 14.2 | 15.54 | 3.53 | 9.7 | 81 | 83.6 | 82.3 | 41.13 | |
| MFFG-RNNArchitecture=RNN-based2023.06 | 12.18 | 4.92 | 1.73 | 0.88 | 14.01 | 15.18 | 4.56 | 10.64 | 81.2 | 83.7 | 82.4 | 45.91 | |
| Transformer2023.06 | 10.68 | 4.01 | 1.49 | 0.71 | 15.04 | 17.25 | 5.32 | 8.99 | 83.2 | 84.7 | 83.9 | 53.94 |