LLM Agents for Interactive Workflow Provenance: Reference Architecture and Evaluation Methodology

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Souza, Renan, Poteet, Timothy, Etz, Brian, Rosendo, Daniel, Gueroudji, Amal, Shin, Woong, Balaprakash, Prasanna, da Silva, Rafael Ferreira
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916963204202496
author Souza, Renan
Poteet, Timothy
Etz, Brian
Rosendo, Daniel
Gueroudji, Amal
Shin, Woong
Balaprakash, Prasanna
da Silva, Rafael Ferreira
author_facet Souza, Renan
Poteet, Timothy
Etz, Brian
Rosendo, Daniel
Gueroudji, Amal
Shin, Woong
Balaprakash, Prasanna
da Silva, Rafael Ferreira
contents Modern scientific discovery increasingly relies on workflows that process data across the Edge, Cloud, and High Performance Computing (HPC) continuum. Comprehensive and in-depth analyses of these data are critical for hypothesis validation, anomaly detection, reproducibility, and impactful findings. Although workflow provenance techniques support such analyses, at large scale, the provenance data become complex and difficult to analyze. Existing systems depend on custom scripts, structured queries, or static dashboards, limiting data interaction. In this work, we introduce an evaluation methodology, reference architecture, and open-source implementation that leverages interactive Large Language Model (LLM) agents for runtime data analysis. Our approach uses a lightweight, metadata-driven design that translates natural language into structured provenance queries. Evaluations across LLaMA, GPT, Gemini, and Claude, covering diverse query classes and a real-world chemistry workflow, show that modular design, prompt tuning, and Retrieval-Augmented Generation (RAG) enable accurate and insightful LLM agent responses beyond recorded provenance.
format Preprint
id arxiv_https___arxiv_org_abs_2509_13978
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LLM Agents for Interactive Workflow Provenance: Reference Architecture and Evaluation Methodology
Souza, Renan
Poteet, Timothy
Etz, Brian
Rosendo, Daniel
Gueroudji, Amal
Shin, Woong
Balaprakash, Prasanna
da Silva, Rafael Ferreira
Distributed, Parallel, and Cluster Computing
Artificial Intelligence
Databases
68M14, 68M20, 68T07
C.2.4; D.1.3; I.2.0
Modern scientific discovery increasingly relies on workflows that process data across the Edge, Cloud, and High Performance Computing (HPC) continuum. Comprehensive and in-depth analyses of these data are critical for hypothesis validation, anomaly detection, reproducibility, and impactful findings. Although workflow provenance techniques support such analyses, at large scale, the provenance data become complex and difficult to analyze. Existing systems depend on custom scripts, structured queries, or static dashboards, limiting data interaction. In this work, we introduce an evaluation methodology, reference architecture, and open-source implementation that leverages interactive Large Language Model (LLM) agents for runtime data analysis. Our approach uses a lightweight, metadata-driven design that translates natural language into structured provenance queries. Evaluations across LLaMA, GPT, Gemini, and Claude, covering diverse query classes and a real-world chemistry workflow, show that modular design, prompt tuning, and Retrieval-Augmented Generation (RAG) enable accurate and insightful LLM agent responses beyond recorded provenance.
title LLM Agents for Interactive Workflow Provenance: Reference Architecture and Evaluation Methodology
topic Distributed, Parallel, and Cluster Computing
Artificial Intelligence
Databases
68M14, 68M20, 68T07
C.2.4; D.1.3; I.2.0
url https://arxiv.org/abs/2509.13978