BERT Knows Punta Cana is not just beautiful, it's gorgeous: Ranking Scalar Adjectives with Contextualised Representations
This addresses the need for improved natural language understanding and reasoning by enabling more accurate intensity detection for scalar adjectives, though it is incremental as it builds on existing BERT methods.
The paper tackled the problem of ranking scalar adjectives by intensity using a BERT-based approach, and the result showed that BERT provides better quality intensity rankings than static embeddings and previous models, as demonstrated on gold standard datasets and an Indirect Question Answering task.
Adjectives like pretty, beautiful and gorgeous describe positive properties of the nouns they modify but with different intensity. These differences are important for natural language understanding and reasoning. We propose a novel BERT-based approach to intensity detection for scalar adjectives. We model intensity by vectors directly derived from contextualised representations and show they can successfully rank scalar adjectives. We evaluate our models both intrinsically, on gold standard datasets, and on an Indirect Question Answering task. Our results demonstrate that BERT encodes rich knowledge about the semantics of scalar adjectives, and is able to provide better quality intensity rankings than static embeddings and previous models with access to dedicated resources.