Reproduce “PEEK: A Large Dataset of Learner Engagement with Educational Videos” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Sahan Bulathwela, Maria Perez-Ortiz, Erik Novak, Emine Yilmaz, John Shawe-Taylor
PEEK provides a public benchmark for predicting informal learners’ engagement with videos. It converts viewing logs into short fragments, labels engagement when watched duration reaches 75%, and adds Wikipedia topic tags. The dataset contains 290,535 events from 20,019 learners. On learners held out for testing, TrueLearn Novel achieves an F1 score of 64.710, highest among tested models; this does not measure learning gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.