Papers by Da Huo
Inducing Universal Semantic Tag Vectors (2020.lrec-1)
Copied to clipboard
| Challenge: | Existing semantic tags are useful for syntactically oriented downstream NLP tasks . but their size is limited and many words are out-of-vocabulary words . |
| Approach: | They propose to tagging words with semantic distinctions that are likely to be useful across semantic tasks. |
| Outcome: | The proposed semantic tagging scheme can predict unseen words with high accuracy . it distinguishes privative attributes from subsective ones, making it easier to discern fake detectives . |