Table 4 F1-score of sentences with different numbers of triples.
From: A hybrid attention and dilated convolution framework for entity and relation extraction and mining
Baseline models | NYT | WebNLG | ||||||
|---|---|---|---|---|---|---|---|---|
\(N=1\) | \(N=2\) | \(N=3\) | \(N=4\) | \(N=1\) | \(N=2\) | \(N=3\) | \(N=4\) | |
CopyR\(_{OneDecoder}\) | 66.6 | 52.6 | 49.7 | 48.7 | 65.2 | 33.0 | 22.2 | 14.2 |
CopyR\(_{MultiDecoder}\) | 67.1 | 58.6 | 52.0 | 53.6 | 59.2 | 42.5 | 31.7 | 24.2 |
GraphRel\(_{1p}\) | 69.1 | 59.5 | 54.4 | 53.9 | 63.8 | 46.3 | 34.7 | 30.8 |
GraphRel\(_{2p}\) | 71.0 | 61.5 | 57.4 | 55.1 | 66.0 | 48.3 | 37.0 | 32.1 |
CopyR\(_{RL}\) | 71.7 | 72.6 | 72.5 | 77.9 | 63.4 | 62.2 | 64.4 | 57.2 |
HADNet(Ours) | 73.6 | 74.1 | 75.9 | 81.1 | 68.1 | 68.9 | 70.9 | 69.9 |
