Use case · Idea to test
Music and playlists
Learn what you like to hear next.
- Rate short clips
- Learn your taste
- Suggest another track
A listener likes sparse piano and skips busy percussion. A music app could rank existing tracks or suggest new clips with that pattern in mind. A future version could start from a playlist you already like, then refine its suggestions with new ratings. Game soundtracks and meditation playlists were also original ideas.
Where the small model fits
A small personal scorer learns from ratings. With a compatible generator, it could score or adjust the internal inputs before rendering a full audio clip. Another option is to represent existing tracks with an audio model and rank those tracks.
What would need to work
Compare blind listener choices against an ordinary playlist, including variety, unwanted repeats and the full cost of making audio. The lab has not validated an audio generator.
Try the related work
This demo learns from simple drawings. It does not run the application described above.
Related experiments and evidence
Earlier proposal and example code
The earlier page is preserved as a historical proposal; its claims are not new validation.