Domain-Grounded Tool Orchestration for LLM-Guided Scientific Analysis
2026-08-31 • Computational Engineering, Finance, and Science
Computational Engineering, Finance, and ScienceHuman-Computer Interaction
AI summaryⓘ
The authors explain that scientific workflows in fields like fluid dynamics rely on complex steps where mistakes in tool use or order can lead to wrong results. They developed a system that divides the process into clear roles: understanding what the user wants with an AI, running trusted analysis tools, and then explaining the results again with AI. This method uses domain-specific rules to avoid common errors seen in AI-generated scripts. Tested in two scientific fields, their approach improved how accurately the system interprets results, especially when information is provided in focused chunks rather than all at once.
Computational Fluid DynamicsScientific VisualizationLarge Language ModelsDomain OntologiesParaViewModel Context ProtocolTopology ToolKitPost-processingWorkflow AutomationScript Generation Errors
Authors
Jeff Lee, Sebastien Jourdain, Cory Quammen, Patrick O'Leary, Berk Geveci
Abstract
Scientific analysis workflows encode deep domain knowledge through sequences of tightly coupled operations where correctness depends on tool selection, execution order, and parameterization. A CFD engineer investigating flow separation must extract wall shear stress, identify zero-crossings in skin friction, and confirm with boundary-layer profiles: a chain that requires both domain expertise and proficiency with visualization tools. Current approaches to LLM-assisted scientific visualization generate scripts that encode this knowledge implicitly, and often incorrectly, producing code that executes but yields wrong results. We present an architecture that separates intent interpretation (LLM) from execution (deterministic domain tools) from explanation (LLM), connected by the Model Context Protocol (MCP) and grounded by domain ontologies that constrain planning to valid analysis chains. We instantiate the architecture in two domains on the same ParaView server infrastructure: computational fluid dynamics post-processing and topological data analysis via the Topology ToolKit (TTK). Adding the second domain required only an ontology and tool wrappers around existing filters, with no change to the architecture, protocol, or deployment. By construction the design removes whole classes of failure that affect script generation (such as API hallucination and missing pipeline stages) and narrows the strategic errors that remain. An ablation across both domains locates the ontology's empirical effect: it does not change which tools the planner selects, which is already reliable, but corrects how the model interprets results, raising interpretation accuracy from 0.41 to 0.91, and only when the relevant fact is retrieved in scoped rather than bulk form. ParaView's client-server model carries analysis to production-scale datasets through a thin browser client.