|
Descripción
|
|
|---|---|
| Scholarly publications are the primary scientific output. The hypotheses that form the basis of these publications are based on scholarly assertions from prior publications. Worryingly, evidence shows that cited assertions may "drift" in their meaning, transforming (superficially or deeply) the original claim. This effect that can worsen through chains of citations. The assertion, upon which a novel hypothesis is based, therefore, may be significantly distinct from the scientific assertion backed by evidence and stated in the conclusion of a scholarly paper. How common is this drift, how significant, and how much can it influence the design of novel hypothesis? Is it possible to automatically detect if our research hypotheses are well-founded? The goal of our work is to investigate automated methods of tracking scientific assertions through the scholarly literature, and assess the degree of divergence in the claim at each step in the citation chain, back to the original publication of that claim. This will be accomplished by, first, generating several dozen citation networks through database queries against the Web of Knowledge, and two proprietary datasets we have been given access to. We will then match assertions between these papers by text/concept mining, using two scholarly-assertion grammar ontologies developed by researchers at Elsevier. We will design metrics to quantify the degree to which the assertion has "drifted", and evaluate how commonly this occurs in the scholarly literature in general. Finally, we will explore approaches to machine-readable publication that can simplify automated validation by more precisely representing the claim. | |
|
Internacional
|
Si |
|
Nombre congreso
|
Workshop New Frontiers in Plant Biology |
|
Tipo de participación
|
970 |
|
Lugar del congreso
|
Centro de Biotecnología y Genómica de Plantas |
|
Revisores
|
Si |
|
ISBN o ISSN
|
0000-0000 |
|
DOI
|
|
|
Fecha inicio congreso
|
15/06/2016 |
|
Fecha fin congreso
|
17/06/2016 |
|
Desde la página
|
30 |
|
Hasta la página
|
30 |
|
Título de las actas
|
Approaches to formally tracking, representing, and validating scientific argumentation |