-演講時間: 114年12月10日（三） 14:00~16:00

-演講地點: E6-A207教室

-演講者: 王科植 助理教授 (國立臺灣師範大學資訊工程學系)

-演講主題: Human-Centered AI Integration in Scientific Simulation Data Exploration Workflow

-演講摘要：Computer simulations and the analysis of their outputs are one of the primary ways scientists develop understanding of complex phenomena. In the AI era, researchers are increasingly exploring how AI models can be integrated into traditional simulation data analysis pipelines to create workflows that are more efficient and effective. While adding AI to scientific data analysis promises benefits, it also raises challenges. These include managing massive training scientific datasets, producing inference results that are reliable, uncertainty-aware, and scientifically trustworthy, building practical engineering tools that enable non–computer science domain scientists to evaluate the AI model incooperated to their existing pipelines, and exploring how knowledge learned by AI models can feed back into scientific reasoning, hypothesis formation, and interactive visual exploration. This talk will share the vision, challenges, and experiences in integrating AI models to support simulation data processing and analysis. This will also highlight emerging research opportunities and future directions for AI-assisted scientific data analysis workflows.