7
$\begingroup$

I am interested in others' experience using Mathematica for machine learning and bioinformatics course curricula. Let me first say that I have been programming with MMA for almost 10 years now (out of the 40 I have been programming total!). I love the aesthetics and the language structure. I am putting together a bioinformatics course for graduate students in biological sciences, with an emphasis on rigor and reproducibility, but also with sections on high dimensional data analysis, network science, and an introduction to machine learning concepts. I would love to do this with MMA notebooks for course notes, etc.

My problem is that MMA is not even on the charts for current usage in machine learning: https://rpubs.com/cvrajesh/kagglesurvey2018. My colleagues have suggested Jupiter notebooks and Python, which are getting better, but nowhere near MMA's elegance.

I would be interested in other's experience putting together course materials in bioinformatics and machine learning with MMA, and their reception among colleagues and students.

My questions are:

1) Have others done this, and would they be willing to share examples of their syllabi?

2) Seems like the future of ML (at least for now) is heavily Python weighted; is there any role for MMA here? I hope so, but am concerned.

If this is not the appropriate forum for this, please suggest where I should post.

Thanks!

$\endgroup$

closed as primarily opinion-based by Carl Lange, MarcoB, eyorble, LCarvalho, Öskå Sep 8 at 17:17

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ @ GraphMan, I am not sure whether you had a chance to look at this post: mathematica.stackexchange.com/questions/86643/… $\endgroup$ – ramesh Jul 23 at 14:52
  • 1
    $\begingroup$ @ramesh - Thanks. I had seen that post a while ago, and it has some good comparisons and comments, but from 4 years ago. Given how things continue to evolve in ML, and trying to be responsible to students, I am looking for examples of how others have used MMA to teach ML. None of the students want to learn MMA, they all say it will be impossible for them to get jobs, post-docs, be published etc. unless their code is written in Python or R. So...looking for some like minded MMA educators in bioinformatics and ML to talk to. Anybody out there? $\endgroup$ – GraphMan Jul 23 at 16:55
  • 1
    $\begingroup$ I could be totally wrong, but it might also be worth talking to Wolfram? Idk if this is something that they do, but they might love the opportunity to gain a few more users and also demonstrate the usefulness of the Wolfram Language for machine learning - something they seem to be pushing hard. $\endgroup$ – MassDefect Jul 24 at 0:36
  • $\begingroup$ "also with sections on ... network science" I can only speak about this part, as I do use Mathematica for network analysis. I think that since v12 (which had many fixes for Graph), it can be pretty good for this purpose. But I could not live without the IGraph/M package anymore. At this point, it has become critical for me for network analysis. Of course, as the author of IGraph/M, I can hardly comment on this without bias. ;-) $\endgroup$ – Szabolcs Jul 24 at 11:54
  • $\begingroup$ What makes Mathematica so nice for working with networks for me is the combination of built-in Graph functionality and IGraph/M. I would not be nearly as happy either with plain Mathematica or using igraph from a different language (R or Python). $\endgroup$ – Szabolcs Jul 24 at 11:55