Results on the ultra-fine entity typing task. LITE series are equipped with the Taxonomic Statement template. “w/o label dependency” is applied to the “NLI+L” setting. The F1 result by LITENLI +L is statistically significant (p-value ¡ 0.01 in t-test) in comparison with the best baseline result by MLMET.
| Model . | P . | R . | F1 . | |
|---|---|---|---|---|
| UFET-biLSTM (Choi et al., 2018) | 48.1 | 23.2 | 31.3 | |
| LabelGCN (Xiong et al., 2019) | 50.3 | 29.2 | 36.9 | |
| LDET (Onoe and Durrett, 2019) | 51.5 | 33.0 | 40.1 | |
| Box4Types (Onoe et al., 2021) | 52.8 | 38.8 | 44.8 | |
| LRN (Liu et al., 2021) | 54.5 | 38.9 | 45.4 | |
| MLMET (Dai et al., 2021) | 53.6 | 45.3 | 49.1 | |
| LITE | NLI | 1.5 | 7.1 | 2.5 |
| L | 48.7 | 45.8 | 47.2 | |
| D+L | 27.5 | 56.4 | 37.0 | |
| NLI+D+L | 45.4 | 49.9 | 47.4 | |
| NLI+L | 52.4 | 48.9 | 50.6 | |
| –w/o label dependency | 53.3 | 46.6 | 49.7 | |
| Model . | P . | R . | F1 . | |
|---|---|---|---|---|
| UFET-biLSTM (Choi et al., 2018) | 48.1 | 23.2 | 31.3 | |
| LabelGCN (Xiong et al., 2019) | 50.3 | 29.2 | 36.9 | |
| LDET (Onoe and Durrett, 2019) | 51.5 | 33.0 | 40.1 | |
| Box4Types (Onoe et al., 2021) | 52.8 | 38.8 | 44.8 | |
| LRN (Liu et al., 2021) | 54.5 | 38.9 | 45.4 | |
| MLMET (Dai et al., 2021) | 53.6 | 45.3 | 49.1 | |
| LITE | NLI | 1.5 | 7.1 | 2.5 |
| L | 48.7 | 45.8 | 47.2 | |
| D+L | 27.5 | 56.4 | 37.0 | |
| NLI+D+L | 45.4 | 49.9 | 47.4 | |
| NLI+L | 52.4 | 48.9 | 50.6 | |
| –w/o label dependency | 53.3 | 46.6 | 49.7 | |