Algorithm finds domain experts by matching papers and clarifying author names
Retrieve-then-Disambiguate: Scalable Reviewer Discovery
Information RetrievalDigital Libraries
Summary
The peer review system for scientific papers struggles because it misses many relevant experts who could review submissions. The authors created a tool that finds researchers by first finding similar scientific papers and then figuring out who the authors really are, even when names are confusing. They tested this on 500 astronomy and physics papers and successfully identified many connected researchers. Their tool helps find reviewers more accurately by linking papers to the right experts.
What this means in practice
- •For conference organizers: Automatically generate candidate reviewer lists by matching submission topics to relevant, uniquely identified researchers in large domains.
- •For digital library managers: Improve author identity resolution to enhance user search experiences and researcher profiling in digital repositories.
Authors
Vicente Amado Olivo, Joe Petrecca, Wolfgang E. Kerzendorf
Abstract
The growing number of scientific submissions is straining the peer review process, while a large pool of relevant researchers remains unidentified and is therefore not considered as potential reviewers. We formulate the discovery of these researchers as a retrieval problem: given a submission, we retrieve and rank researchers based on domain similarity. We present retrieve-then-disambiguate, an open-source algorithm that identifies domain relevant researchers for a query submission by retrieving similar publications from the scientific literature and resolving ambiguous author names into unique researcher identities. We evaluate the algorithm on 500 astronomy and physics publications. We use ORCID-linked first authors, coauthors, and cited researchers as proxy labels to evaluate the recovery of researchers with known connections to each query publication. When retrieving the top 500 most similar papers, the algorithm recovers roughly 68\% of first authors, 60\% of coauthors, and 45\% of cited researchers. Our code is available at https://github.com/deepthought-initiative/Retrieve-then-Disambiguate.