MermaidLLM: Dataflow Diagrams for Explainable Skill Formalization and Real-time Support with Multimodal LLMs
Gaze tracking deviceFigure 1: MermaidLLM uses MLLMs to comprehend expert skills from egocentric vision/gaze, formalizing them as dataflow diagrams (Mermaid).It then acts as a real-world agent, comparing the learner's context with the diagram and providing verbal guidance.
Institutions
- Sony Computer Science Laboratories (JP)
- The University of Tokyo (JP)
Publication Details
- Published
- 2025-09-27
- DOI
- https://doi.org/10.1145/3746058.3758449
- Primary Topic
- Asphalt Pavement Performance Evaluation
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Moonshot Research and Development Program