EUVIMEDEUVIMED entdecken
Crossref

Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs

Guanghui Zhao, Zhe Wang, Yu Dong, Guan Li, Guihua Shan

Information Visualization · 2026-07 · 25 (3) · 373-390

DOI: 10.1177/14738716261434848 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Information Visualization, online erschienen am 09.04.2026. Verfasst von Guanghui Zhao und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 42 Literaturangaben. Gefördert von: Chinese Academy of Sciences, Computer Network Information Center, Chinese Academy of Sciences, natural science foundation of beijing municipality.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Scientific visualization pipelines encode domain-specific procedural knowledge with strict execution dependencies, making their construction sensitive to missing stages, incorrect operator usage, or improper ordering. Thus, generating executable scientific visualization pipelines from natural-language descriptions remains challenging for large language models, particularly in web-based environments where visualization authoring relies on explicit code-level pipeline assembly. In this work, we investigate the reliability of LLM-based scientific visualization pipeline generation, focusing on vtk.js as a representative web-based visualization library. We propose a structure-aware retrieval-augmented generation workflow that provides pipeline-aligned vtk.js code examples as contextual guidance, supporting correct module selection, parameter configuration, and execution order. We evaluate the proposed workflow across multiple multi-stage scientific visualization tasks and LLMs, measuring reliability in terms of pipeline executability and human correction effort. To this end, we introduce correction cost as metric for the amount of manual intervention required to obtain a valid pipeline. Our results show that structured, domain-specific context substantially improves pipeline executability and reduces correction cost. We additionally provide an interactive analysis interface to support human-in-the-loop inspection and systematic evaluation of generated visualization pipelines.

Förderung

  • Chinese Academy of Sciences
  • Computer Network Information Center, Chinese Academy of Sciences
  • natural science foundation of beijing municipality

Zitieren

APA

Zhao, G., Wang, Z., Dong, Y., Li, G., & Shan, G. (2026). Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs. Information Visualization, 25(3), 373-390. https://doi.org/10.1177/14738716261434848

Vancouver

Zhao G, Wang Z, Dong Y, Li G, Shan G. Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs. Information Visualization. 2026;25(3):373-390. doi:10.1177/14738716261434848

BibTeX
@article{Zhao2026,
  author = {Zhao, Guanghui and Wang, Zhe and Dong, Yu and Li, Guan and Shan, Guihua},
  title = {Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs},
  journal = {Information Visualization},
  year = {2026},
  volume = {25},
  number = {3},
  pages = {373--390},
  doi = {10.1177/14738716261434848}
}
Kostenlos starten →