A Unified Feature Representation for Lexical Connotations
This work addresses the challenge of recognizing cultural and emotional perspectives in text, which is important for applications like stance detection, but it is incremental as it builds on existing methods for lexical representation.
The paper tackled the problem of understanding subtle ideological and emotional connotations in language by creating a new lexical resource using distant labeling, which aligns well with human judgments, and showed that using embeddings from this resource provides a statistically significant improvement in stance detection with limited data.
Ideological attitudes and stance are often expressed through subtle meanings of words and phrases. Understanding these connotations is critical to recognizing the cultural and emotional perspectives of the speaker. In this paper, we use distant labeling to create a new lexical resource representing connotation aspects for nouns and adjectives. Our analysis shows that it aligns well with human judgments. Additionally, we present a method for creating lexical representations that captures connotations within the embedding space and show that using the embeddings provides a statistically significant improvement on the task of stance detection when data is limited.