Scientific Authoring Support: A Tool to Navigate in Typed Citation Graphs

Abstract

Scientific authors urgently need help in managing the fast increasing number of publications. We describe and demonstrate a tool that supports authors in browsing graphically through electronically available publications, thus allowing them to quickly adapt to new domains and publish faster. Navigation is assisted by means of typed citation graphs, i.e. we use methods and resources from computational linguistics to compute the kind of citation that is made from one paper to another (refutation, use, confirmation etc.). To verify the computed citation type, the user can inspect the highlighted citation sentence in the original PDF document. While our classification methods used to generate a realistic test data set are relatively simple and could be combined with other proposed approaches, we put a strong focus on usability and quick navigation in the potentially huge graphs. In the outlook, we argue that our tool could be made part of a community approach to overcome the sparseness and correctness dilemma in citation classification. 1

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 17 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.