From Bayesian to deep knowledge tracing — upgrading NumPath's student model with a PyTorch LSTM

NumPath is moving from Bayesian Knowledge Tracing to a Deep Knowledge Tracing model built in PyTorch with an LSTM.

The new student simulator learns from skill-and-correctness attempt sequences and predicts the next answer using the standard DKT setup.

Unlike BKT, the model can capture relationships across skills, not just isolated knowledge components.

That should help NumPath model shared error patterns more accurately and improve mastery prediction for students with dyscalculia.