Computational Pipeline

This project conceived a novel computational data-driven approach that extracts conceptually relevant formative feedback, called “formative fugues,” suitable for guiding learners exploring complex systems concepts in open-ended learning environments. The approach learns common patterns of explorations by extracting scientifically meaningful sequences from a corpus of data of prior learners’ explorations of a particular complex system. These common patterns, dubbed “fugues”, can be reused, repurposed and reassembled into longer chains much like musical fugues. The computational approach leverages causal modeling followed by pattern matching to identify the formative fugues from among multiple simultaneous causal chains and can be used in real-time during a given enactment.