Johannes M. van Hulst, Faegheh Hasibi, Koen Dercksen, Krisztian Balog, Arjen P. de Vries
Entity linking is a standard component in modern retrieval system that is often performed by third-party toolkits. Despite the plethora of open source options, it is difficult to find a single system that has a modular architecture where certain components may be replaced, does not depend on external sources, can easily be updated to newer Wikipedia versions, and, most important of all, has state-of-the-art performance. The REL system presented in this paper aims to fill that gap. Building on state-of-the-art neural components from natural language processing research, it is provided as a Python package as well as a web API. We also report on an experimental comparison against both well-established systems and the current state-of-the-art on standard entity linking benchmarks.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Entity Linking | MSNBC | Micro-F1 | 72.4 | van Hulst et al. (2020) |
| Entity Linking | AIDA-CoNLL | Micro-F1 strong | 80.5 | van Hulst et al. (2020) |
| Entity Linking | Derczynski | Micro-F1 | 41.1 | van Hulst et al. (2020) |