Suchen und Finden von Forschungsinformationen leichter gemacht

Abstract

"In this article, we would like to introduce the newly developed search infrastructure that we have built at the *Institute for Employment Research (IAB)* to create optimized access to the metadata and background information of the research database. The new search function is also intended to contribute to a potential increase in research productivity. The infrastructure is based on Elasticsearch with an Angular front end. We extend the basic functionality using Flask microservices, which use word2vec models to promote search creativity and transformer models for abstract searching, even across different languages. In this way, we aim to show how easily and effectively modern NLP models can be combined with traditional search ideas, while simultaneously providing a blueprint for appropriate implementation." (Author's abstract, IAB-Doku) ((en))

Cite article

Dittgen, M., Metzger, L. & Thomsen, U. (2024): Suchen und Finden von Forschungsinformationen leichter gemacht. (FDZ-Methodenreport 02/2024 (de)), Nürnberg, 16 p. DOI:10.5164/IAB.FDZM.2402.de.v1