Rapsai: Rapid Application Prototyping System for AI
The project Rapsai aims to make the prototyping of machine learning (ML) based multimedia applications more efficient and accessible. The researchers conducted a formative study with ML practitioners to gather insights on common challenges in the field, which informed the development of Rapsai. Rapsai is a visual programming platform that allows for rapid and iterative development of end-to-end ML-based multimedia applications, through a node-graph editor that facilitates interactive characterization and visualization of ML model performance. It streamlines the prototyping process with interactive data augmentation and model comparison capabilities, in a no-coding environment. The evaluation of Rapsai in four real-world case studies suggests that it accelerates workflow, makes informed decisions, and allows for the holistic evaluation of model behavior with real-world input.