Papers with document embedding methods
Learning Outside the Box: Discourse-level Features Improve Metaphor Identification (N19-1)
Copied to clipboard
| Challenge: | Current approaches to metaphor identification use restricted linguistic contexts, e.g. by only considering a verb’s arguments or the sentence containing a phrase. |
| Approach: | They propose to train simple gradient boosting classifiers on representations of an utterance and its surrounding discourse learned with a variety of document embedding methods. |
| Outcome: | The proposed classifiers obtained state-of-the-art results on the 2018 VU Amsterdam metaphor identification task without complex metaphor-specific features or deep neural architectures employed by other systems. |