I teach a Digital Signal Processing Lab to Electrical Engineering undergraduate level students. We have most Labs on MATLAB and some Labs on a dsp kit tms320c6713. I am planning to incorporate topics and contents related to Machine Learning, and I am wondering if it's ok? Anyone else doing the same (teaching Machine Learning topics in Signal Processing lab)?
$\begingroup$
$\endgroup$
1
-
$\begingroup$ You're the teacher, so that's a question you only should answer in my opinion. If it helps, from my experience with audio-related Machine Learning, Signal Processing is a must-have skill for such things as feature extraction. I don't know how deep you want to go in your teaching of machine learning, but if anything it can't hurt to show students what Signal Processing brings to that field! $\endgroup$– JdipCommented Aug 24, 2022 at 13:51
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
ML can be incorporated into SP as a useful complement. ML has killed SP in many areas, and it's informative to teach how and why - also SP remains a useful framework from which ML can be understood. Further, SP is fundamental in science as the "study of the measured and the finite" (if taught right).
SP has also learned from ML. The best example I can think of is Wavelet Scattering - this lecture explains the success of conv-nets in terms of signal processing, and it's my favorite in terms of rigor and intuition. I provide a visual overview here.
That said, I'd not recommend "ML + SP" as in "Fourier transform week 1, dense networks math week 2". Keeping ML mostly high-level should avoid making it a distraction.
answered Aug 24, 2022 at 14:11