Understanding Semantic Search on Scientific Repositories: Steps towards Meaningful Findability

Open Science has become one of the most important movements for leveraging scientific collaboration through sharing and reuse of research outputs, such as publications, data, software and methods, which are stored and preserved in public repositories. The emergence of FAIR principles as mandatory for effective openness has made it clear that it is not enough to publish such outputs – ”findability” (our focus) is a key issue. It can be leveraged by good documentation and, e.g., PIDs (persistent identifiers), but search mechanisms have yet to adapt to this new context. Many such mechanisms – in particular semantic search – have risen as a means towards findability. Nevertheless, implementing these mechanisms and integrating them into scientific repositories presents many challenges. This paper presents a systematic literature review of research efforts on mechanisms designed to support search for research outputs – publications, data and processes. Our analysis is based on processing the entire collection of papers stored in Scopus, IEEE Xplore and arXiV. We identified 299 papers related to semantic search in scientific repositories. Their analysis provided a categorization of existing literature and unexplored gaps, pointing out new research challenges for semantic search mechanisms to support Open Science