This poster introduces the novel idea of using Natural Language Processing (NLP) tools on unpublished articles, thereby alerting writers and editors to the potential problems that these articles may pose to NLP tools, postpublication, and offering them the option to nip such problems in the bud. To demonstrate the feasibility of this approach, the poster takes on the NLP tool Wikify! and uses it in the aforementioned manner, introduces an automated feature that might be deployed alongside it using WordNet and analyzes the results thus generated.

This poster proposes the novel idea of semantic stability and hypothesizes that the relative stability of words, as they undergo algorithmic manipulations, can be measured. It attempts to move toward a proof-of-concept for measuring the relative stability of words, by putting forward a preliminary Semantic Stability Index (SSI) and speculates upon the potentials, next steps, and future directions of such an approach.

No abstract