Supervisor: Torsten Möller, Lu Ying
We have developed an AI-based workflow for automatically generating animated transitions between data visualizations. The workflow involves multiple steps and AI agents for analyzing visualizations, planning transitions, and generating animations.
A current issue is that the whole workflow can take a relatively long time to complete. This may be caused by how different agents and processing steps are organized, such as sequential execution, repeated analysis, unnecessary agent calls, or recomputation of intermediate results. This project will explore where these bottlenecks come from and how the workflow can be made faster, for example through better agent coordination, reuse of intermediate results, or parallel execution.
Prerequisites: VIS
Contact: Torsten Möller, Lu Ying