EUVIMEDEUVIMED entdecken
Crossref

A scientometric study of retrieval-augmented generation research: Mapping the intellectual landscape of a transformative AI paradigm

Mahdi Shirdeli, Ali Shirdeli, Shahpour Alirezaee

Scientometrica · 2027 · 3 (2) · 107-122

DOI: 10.5267/j.ijdns.2026.24 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Scientometrica, 2027. Verfasst von Mahdi Shirdeli und 2 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 1 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This scientometric study provides a comprehensive analysis of the intellectual landscape surrounding Retrieval-Augmented Generation (RAG), a transformative paradigm that addresses fundamental limitations of Large Language Models by grounding generative outputs in external knowledge sources. Analyzing 200 highly cited Scopus-indexed publications from 2006 to 2026, this study employs bibliometric techniques including citation analysis, co-authorship network analysis, keyword co-occurrence analysis, and thematic evolution mapping to characterize the field's development trajectory, intellectual structure, and emerging frontiers. The findings reveal exponential growth in RAG research since 2020, with a 215.7% compound annual growth rate between 2020 and 2023, accelerating dramatically after 2024. The United States and China emerge as dominant contributors, supported by a collaborative “Triple Helix” network linking academia, industry, and government. Thematic analysis identifies six major research clusters: foundational RAG architectures, biomedical and healthcare applications, educational applications, legal and compliance applications, industrial and manufacturing applications, and multimodal RAG systems. Evaluation practices reveal fragmentation with over 40 distinct metrics employed across studies. Key challenges identified include hallucination mitigation, computational efficiency, retrieval quality optimization, and privacy preservation. Emerging directions include agentic RAG systems, graph-based retrieval, iterative retrieval-generation synergy, and personalized RAG applications. This study contributes a data-driven roadmap for navigating this rapidly evolving field and identifies priority areas for future investigation.

Zitieren

APA

Shirdeli, M., Shirdeli, A., & Alirezaee, S. (2027). A scientometric study of retrieval-augmented generation research: Mapping the intellectual landscape of a transformative AI paradigm. Scientometrica, 3(2), 107-122. https://doi.org/10.5267/j.ijdns.2026.24

Vancouver

Shirdeli M, Shirdeli A, Alirezaee S. A scientometric study of retrieval-augmented generation research: Mapping the intellectual landscape of a transformative AI paradigm. Scientometrica. 2027;3(2):107-122. doi:10.5267/j.ijdns.2026.24

BibTeX
@article{Shirdeli2027,
  author = {Shirdeli, Mahdi and Shirdeli, Ali and Alirezaee, Shahpour},
  title = {A scientometric study of retrieval-augmented generation research: Mapping the intellectual landscape of a transformative AI paradigm},
  journal = {Scientometrica},
  year = {2027},
  volume = {3},
  number = {2},
  pages = {107--122},
  doi = {10.5267/j.ijdns.2026.24}
}
Kostenlos starten →