Comment by gotoeleven

At a high level, it's very unintuitive to me that transforming to the frequency domain for analyzing long sequences of tokens--when there is in general no expectation that the sequences have a periodic structure--can improve efficiency.

Stated another way, how can it be possible that it is more efficient to translate the sequence into a series of N variables, where the nth variable is the sum of every nth term of the sequence, if it is unlikely that any relationship between these variables holds for any fixed period? If I combine the 1st 4th 7th 10th .... elements of the sequence, how do we expect the addition of anything beyond the first two elements to add anything but noise?

Stated another another way, if I'm going to approximate a function as a sum of sine waves, this is most efficient when the function is periodic and requires more and more sine waves in the sum to approximate the function on a larger and larger domain when the function is not periodic.