Intern (relx)

Job Description

Taxonomies group synonymous terms together as concepts, and assign a hierarchy between them. Many knowledge sources use the concepts defined in a taxonomy to annotate entities such as diseases, proteins, or medical devices. Therefore, mapping concepts to each other enables obtaining a comprehensive overview of all knowledge about an entity. However, historically, many taxonomies have been developed in parallel, leading to them having (at least) partially overlapping contents. Mapping taxonomy concepts to each other could therefore save researchers large amounts of time and effort that would otherwise be required to manually combine the knowledge.

Performing all mappings manually by subject matter experts assures the highest quality, but is too costly. Existing tools that automatically map concepts are error prone, as different lexical variants of terms may be used, and if they match exactly, they may be ambiguous and therefore refer to different entities. If terms don’t match, they may require possibly error-inducing transformation to make them match, which also fail when the semantic distance is too large (e.g. “disease” and “illness”).

At Elsevier, we have developed a toolkit to support subject matter experts to construct taxonomies. It consists of several models, including a model that distinguishes terms being used ambiguously from those being used unambiguously in scientific literature, a model that assigns a category to a term, a model that identifies synonymous terms, and a term normalization toolkit. Combined with the structural information of a taxonomy, they can be used to create a rich feature set for every concept which may be used to map them to each other (1,2).

In this project, you will create an ensemble algorithm based upon such a feature set, identifying the optimal mix of features and weights to be assigned to them to achieve the highest performance in mapping concepts to each other. If succesful, your work will contribute to better interconnections between scientific knowledge contained in different sources, enabling scientists to easily and quickly obtain a comprehensive overview of all scientific knowledge about their entity of interest.

-----------------------------------------------------------------------

Elsevier is an equal opportunity employer: qualified applicants are considered for and treated during employment without regard to race, color, creed, religion, sex, national origin, citizenship status, disability status, protected veteran status, age, marital status, sexual orientation, gender identity, genetic information, or any other characteristic protected by law. We are committed to providing a fair and accessible hiring process. If you have a disability or other need that requires accommodation or adjustment, please let us know by completing our Applicant Request Support Form: https://forms.office.com/r/eVgFxjLmAK , or please contact 1-855-833-5120.

Please read our Candidate Privacy Policy.