LLM Framework for Virtual Lab Procedure Authoring
A new framework aims to manage uncertainty in LLM-generated procedural knowledge for virtual laboratory planning. Educational virtual laboratories offer scalable, adaptive, and accessible experimental training, especially when physical lab access is limited. However, authoring simulated procedures is costly, requiring descriptions of equipment, interactions, and valid procedural flows. LLMs can assist by generating detailed experimental procedures, but their output may omit actions, misorder steps, or produce logically incorrect instructions. The proposed prototype framework addresses these issues to enable reliable use of LLMs in virtual lab authoring.
Key facts
- Educational virtual laboratories make experimental training more scalable, adaptive, and accessible.
- Authoring simulated laboratory procedures is costly.
- LLMs can generate detailed experimental procedures but their output may be unreliable.
- LLM output may omit necessary actions, arrange steps in wrong order, or produce logically incorrect instructions.
- A prototype framework for managing uncertainty in LLM-generated procedural knowledge is presented.
- The framework aims to enable reliable use of LLMs in virtual laboratory planning.
Entities
—