0
$\begingroup$

I'm on my $4^{th}$ year now and searching for a theme for my diploma in mathematics & computer science specialization.

Is there are interesting problems or fields for researches that i can explore? Probably i'd wanted to work in fields connected to art - music reccomendations e.g. (but we have Spotify sure).

I was suggested to take theme "Authorship attribution in classical music", but i think this theme is rather explored.

Improve this question
$\endgroup$

2 Answers 2

Reset to default
2
$\begingroup$

You could begin your journey by picking up one of the topics present in the magenta project started by google. It's open-source so you could pick up some great ideas there.

Link to Magenta: https://magenta.tensorflow.org/

Another aspect you could go ahead with is trying to combine to genres of music and produce something new. There are also surveys in recommendation for music. Spend a day getting to know what's there and if you could take up a certain drawback from a paper and go ahead with it!

Finally, if nothing clicks, you could refer to chapter 8 from here: https://arxiv.org/abs/1709.01620

This paper is a survey on deep learning for music generation! The 8th chapter focuses on current problems and future prospects.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks. 1. In Magenta you ment i can take any other idea and use it in my paper? 2. What do you mean by "combining genres"? Combine music's written by networks from another genres? 3. By "drawback" you mean taking other papers mistakes and correcting it? $\endgroup$
    – taciturno
    Oct 18, 2020 at 22:51
  • $\begingroup$ So by magenta, I meant you can try to contribute open-source. Even minor contributions can be taken as a project since magenta is the state of the art. By combining genres I meant that you can mix rap and classical and create a new beat perhaps? Drawbacks don't necessarily mean mistakes. You could try to improve the accuracy or some other particular metric! But yes, mistakes count as well! $\endgroup$
    – Academic
    Oct 19, 2020 at 9:10
0
$\begingroup$

Music generation is always good fun, YouTuber Carykh has an interesting take on it using image generation where images are converted to and from MIDI for output and input respectively.

Something I've looked into in the past is taking in a track and separating it into its component instruments.

Additionally, you could do some work in converting sheet music into audio.

These are just a couple random suggestions, but I hope maybe one of them could inspire you to find an idea that you're really into!

Improve this answer
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.