Zhuosheng Zhang, Jiangtong Li, Hai Zhao, Bingjie Tang
This paper describes a hypernym discovery system for our participation in the SemEval-2018 Task 9, which aims to discover the best (set of) candidate hypernyms for input concepts or entities, given the search space of a pre-defined vocabulary. We introduce a neural network architecture for the concerned task and empirically study various neural network models to build the representations in latent space for words and phrases. The evaluated models include convolutional neural network, long-short term memory network, gated recurrent unit and recurrent convolutional neural network. We also explore different embedding methods, including word embedding and sense embedding for better performance.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Hypernym Discovery | General | MAP | 5.77 | SJTU BCMI |
| Hypernym Discovery | General | MRR | 10.56 | SJTU BCMI |
| Hypernym Discovery | General | P@5 | 5.96 | SJTU BCMI |
| Hypernym Discovery | Music domain | MAP | 4.71 | SJTU BCMI |
| Hypernym Discovery | Music domain | MRR | 9.15 | SJTU BCMI |
| Hypernym Discovery | Music domain | P@5 | 4.91 | SJTU BCMI |
| Hypernym Discovery | Medical domain | MAP | 11.69 | SJTU BCMI |
| Hypernym Discovery | Medical domain | MRR | 25.95 | SJTU BCMI |
| Hypernym Discovery | Medical domain | P@5 | 11.69 | SJTU BCMI |
| Taxonomy Learning | General | MAP | 5.77 | SJTU BCMI |
| Taxonomy Learning | General | MRR | 10.56 | SJTU BCMI |
| Taxonomy Learning | General | P@5 | 5.96 | SJTU BCMI |
| Taxonomy Learning | Music domain | MAP | 4.71 | SJTU BCMI |
| Taxonomy Learning | Music domain | MRR | 9.15 | SJTU BCMI |
| Taxonomy Learning | Music domain | P@5 | 4.91 | SJTU BCMI |
| Taxonomy Learning | Medical domain | MAP | 11.69 | SJTU BCMI |
| Taxonomy Learning | Medical domain | MRR | 25.95 | SJTU BCMI |
| Taxonomy Learning | Medical domain | P@5 | 11.69 | SJTU BCMI |