How Co-ML Helps Beginners Learn Data Design Practices for Inclusive ML Models
Machine learning (ML) models are shaped by data, and building inclusive ML systems requires considerations around how to design representative datasets. However, few novice-oriented ML modeling tools foster hands-on learning of dataset design practices.
Data Design Practices for Inclusive ML Models
Co-ML, a tablet-based application, outlines a set of four data design practices (DDPs) and fosters the learning of DDPs through a collaborative ML model. Beginners can build image classifiers through a distributed experience where data is synchronized across multiple devices, enabling multiple users to refine ML datasets in discussion with their peers.
Using Co-ML in an Educational Setting
Co-ML was deployed in a 2-week AIML Summer Camp, where youth ages 13-18 worked in groups to build custom ML-powered mobile applications. The analysis revealed how multi-user model-building with Co-ML supported the development of DDPs and empowered learners to actively engage in exploring the role of data in ML systems.