View in EDS

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

Saved in:
Bibliographic Details
Title: Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs.
Authors: Zhao, Guanghui1,2 (AUTHOR), Wang, Zhe2,3 (AUTHOR), Dong, Yu3 (AUTHOR), Li, Guan2,3 (AUTHOR), Shan, Guihua1,2,3 (AUTHOR) sgh@cnic.cn
Source: Information Visualization. Jul2026, Vol. 25 Issue 3, p373-390. 18p.
Subjects: Scientific visualization, Language models, Workflow management, Code generators, Interactive computer systems
Abstract: 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. [ABSTRACT FROM AUTHOR]
Copyright of Information Visualization is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 194357162
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Zhao%2C+Guanghui%22">Zhao, Guanghui</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Wang%2C+Zhe%22">Wang, Zhe</searchLink><relatesTo>2,3</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Dong%2C+Yu%22">Dong, Yu</searchLink><relatesTo>3</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Li%2C+Guan%22">Li, Guan</searchLink><relatesTo>2,3</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Shan%2C+Guihua%22">Shan, Guihua</searchLink><relatesTo>1,2,3</relatesTo> (AUTHOR)<i> sgh@cnic.cn</i>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Information+Visualization%22">Information Visualization</searchLink>. Jul2026, Vol. 25 Issue 3, p373-390. 18p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Scientific+visualization%22">Scientific visualization</searchLink><br /><searchLink fieldCode="DE" term="%22Language+models%22">Language models</searchLink><br /><searchLink fieldCode="DE" term="%22Workflow+management%22">Workflow management</searchLink><br /><searchLink fieldCode="DE" term="%22Code+generators%22">Code generators</searchLink><br /><searchLink fieldCode="DE" term="%22Interactive+computer+systems%22">Interactive computer systems</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: 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. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Information Visualization is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=194357162
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1177/14738716261434848
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 18
        StartPage: 373
    Subjects:
      – SubjectFull: Scientific visualization
        Type: general
      – SubjectFull: Language models
        Type: general
      – SubjectFull: Workflow management
        Type: general
      – SubjectFull: Code generators
        Type: general
      – SubjectFull: Interactive computer systems
        Type: general
    Titles:
      – TitleFull: Toward reliable scientific visualization pipeline construction with structure-aware retrieval-augmented LLMs.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Zhao, Guanghui
      – PersonEntity:
          Name:
            NameFull: Wang, Zhe
      – PersonEntity:
          Name:
            NameFull: Dong, Yu
      – PersonEntity:
          Name:
            NameFull: Li, Guan
      – PersonEntity:
          Name:
            NameFull: Shan, Guihua
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 07
              Text: Jul2026
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-print
              Value: 14738716
          Numbering:
            – Type: volume
              Value: 25
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Information Visualization
              Type: main
ResultId 1