Query-focused Summarization on Debatepedia (test)
0.4126ROUGE-1SD2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SD2Description=Soft LSTM based diversity attention model2017.04 | 0.4126 | 0.1875 | 0.4043 | |
| D2Description=LSTM based diversity attention model2017.04 | 0.3812 | 0.1676 | 0.3731 | |
| D1Description=Diversity attention model2017.04 | 0.3385 | 0.1365 | 0.3299 | |
| M1Description=Baseline model operating on context vector2017.04 | 0.3306 | 0.1335 | 0.3217 | |
| SD1Description=Soft diversity attention model2017.04 | 0.3136 | 0.1123 | 0.305 | |
| Query_attDescription=Model with query attention in addition to query encoder2017.04 | 0.2928 | 0.1024 | 0.2821 | |
| B1Description=Diversity cell replaced by basic LSTM cell2017.04 | 0.2318 | 0.0646 | 0.2203 | |
| Query_encDescription=Model with query encoder2017.04 | 0.2087 | 0.0339 | 0.1938 | |
| M2Description=Baseline model operating on attention weights and context vector2017.04 | 0.1842 | 0.0447 | 0.1745 | |
| Vanilla e-a-dDescription=Basic encoder-attention-decoder model2017.04 | 0.1373 | 0.0206 | 0.1284 |