Large-Scale Concept Mining in Scientific Publications - Einzelansicht

Scientific publications can be regarded as
contributions to a complex asynchronous and multi-threaded conversation
about specific research questions. Scientists develop and describe their
own ideas and relate them to and compare them with previous related
ideas and concepts in the field. In this project, you will learn how to
apply neural natural language processing technologies and parallel
computing to make conceptual relationships explicit in a datastructure
called a concept graph. In a concept graph, two concepts are connected
by a directed edge if one relates to the other. It can further be
developed into a concept prerequisite graph with directed edges if the
concepts build upon another. During the project, we would like to create
and publish a concept graph as a resource for CORD-19, an open dataset
of currently 280K scholarly articles about the coronavirus.

time and place: t.b.a.

Abschlusspräsentation und Ausarbeitung

B.Sc. Medieninformatik

M.Sc. Computer Science for Digital Media /Computer Science and Media

M.Sc. Human-Computer Interaction