DBLPLink: An Entity Linker for the DBLP Scholarly Knowledge Graph

TL;DR

DBLPLink is a web tool that links scholarly text to the DBLP knowledge graph using advanced language models and entity embeddings, enabling comparison of different linking methods.

Contribution

Introduces DBLPLink, a novel entity linking system combining T5 models and various KG embeddings for scholarly knowledge graphs.

Findings

Effective entity linking with T5 models demonstrated.

Comparison of different KG embeddings shows varying performance.

User interface allows easy comparison of linking results.

Abstract

In this work, we present a web application named DBLPLink, which performs entity linking over the DBLP scholarly knowledge graph. DBLPLink uses text-to-text pre-trained language models, such as T5, to produce entity label spans from an input text question. Entity candidates are fetched from a database based on the labels, and an entity re-ranker sorts them based on entity embeddings, such as TransE, DistMult and ComplEx. The results are displayed so that users may compare and contrast the results between T5-small, T5-base and the different KG embeddings used. The demo can be accessed at https://ltdemos.informatik.uni-hamburg.de/dblplink/.