MelBERT: Metaphor Detection via Contextualized Late Interaction using Metaphorical Identification Theories
This work addresses the problem of identifying metaphorical expressions in text for natural language processing applications, representing an incremental improvement over existing methods.
The paper tackled automated metaphor detection by proposing MelBERT, a model that integrates contextualized word representations with linguistic metaphor identification theories, achieving state-of-the-art performance on four benchmark datasets including VUA-18, VUA-20, MOH-X, and TroFi.
Automated metaphor detection is a challenging task to identify metaphorical expressions of words in a sentence. To tackle this problem, we adopt pre-trained contextualized models, e.g., BERT and RoBERTa. To this end, we propose a novel metaphor detection model, namely metaphor-aware late interaction over BERT (MelBERT). Our model not only leverages contextualized word representation but also benefits from linguistic metaphor identification theories to distinguish between the contextual and literal meaning of words. Our empirical results demonstrate that MelBERT outperforms several strong baselines on four benchmark datasets, i.e., VUA-18, VUA-20, MOH-X, and TroFi.