Sense abstractness, semantic activation, and word sense disambiguation

Abstract

Concrete concepts are often easier to understand than abstract concepts. The notion of abstractness is thus closely tied to the organisation of our semantic memory, and more specifically our internal lexicon, which underlies our word sense disambiguation (WSD) mechanisms. Stateof-the-art automatic WSD systems often draw on a variety of contextual cues and assign word senses by an optimal combination of statistical classifiers. The validity of various lexico-semantic resources as models of our internal lexicon and the cognitive aspects pertinent to the lexical sensitivity of WSD are seldom questioned. We attempt to address these issues by examining psychological evidence of the internal lexicon and its compatibility with the information available from computational lexicons. In particular, we compare the responses from a word association task against existing lexical resources, WordNet and SUMO, to explore the relation between sense abstractness and semantic activation, and thus the implications on semantic network models and the lexical sensitivity of WSD. Our results suggest that concrete senses are more readily activated than abstract senses, and broad associations are more easily triggered than narrow paradigmatic associations. The results are expected to inform the construction of lexico-semantic resources and WSD strategies.

Annotating Concept Abstractness by Common-Sense Knowledge

“…A first evaluation of the annotation is based on internal coherence. We recorded the average abstractness of the values along all dimensions: also based on literature (see, e.g., [10,7]) we expect that concrete vectors are, on average, more connected to other concrete concepts (e.g., through the Synonym relation).…”

“…The role of abstractness has also been explored in the context of the Word Sense Disambiguation [10], leading to the finding that words with very high or very low score of abstractness are easier to disambiguate. Along this line, the association of word senses with senses of di↵erent words have been examined, finding that concrete concepts tend to be related to concrete concepts and abstract concepts tend to be related to abstract ones.…”

“…These features make conceptual abstractness matter of broad interest for computational linguistics, and the investigation on conceptual abstractness a challenging though only superficially explored field. Information on conceptual abstractness impacts on many diverse NLP tasks, such as the word sense disambiguation task [10], the semantic processing of figurative language [3,18,14], the automatic translation and simplification [22], the characterisation of web queries with di culty scores [21], the processing of social tagging information [2], and many others, as well.…”

Longitudinal Study of a Website for Assessing American Presidential Candidates and Decision Making of Potential Election Irregularities Detection

“…Studies in subjects as varied as finance and medicine have used extractive summation, sentiment classification, and corpus-based stemming techniques, which are taken from this field of study (Arroyo-Fernández, Curiel, & Méndez-Cruz 2019); (Singh & Gupta, 2019); (Xu, Huang, Zhang, & Wang, (2019); (Xing, Cambria, & Zhang, (2019). Kwong (2008) examined the chronological evidence for internal lexicons and suggested that concrete senses are more readily activated than abstract ones and that broad associations are more easily triggered than narrow paradigmatic ones. Ion and Tufis (2009) describe two different WSD systems, one of which is applicable to parallel corpora and the other of which is less knowledge rich.…”