Interpreting Verbal Metaphors by Paraphrasing
This work addresses the problem of processing metaphorical expressions for NLP tasks, offering a domain-specific improvement in machine translation.
The paper tackles the challenge of interpreting verbal metaphors in NLP by developing an unsupervised method using BERT and WordNet hypernyms and synonyms, which significantly outperforms state-of-the-art baselines and improves machine translation accuracy for English metaphors across 8 target languages.
Metaphorical expressions are difficult linguistic phenomena, challenging diverse Natural Language Processing tasks. Previous works showed that paraphrasing a metaphor as its literal counterpart can help machines better process metaphors on downstream tasks. In this paper, we interpret metaphors with BERT and WordNet hypernyms and synonyms in an unsupervised manner, showing that our method significantly outperforms the state-of-the-art baseline. We also demonstrate that our method can help a machine translation system improve its accuracy in translating English metaphors to 8 target languages.