Perhaps I misunderstood but it calculates the FDE of query and looks for a similar FDE in the dataset of the model. Doesnt this require calculating all the equivalent sized FDEs in the model?