We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.SD

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Computer Science > Sound

Title: Generative Deep Learning for Virtuosic Classical Music: Generative Adversarial Networks as Renowned Composers

Abstract: Current AI-generated music lacks fundamental principles of good compositional techniques. By narrowing down implementation issues both programmatically and musically, we can create a better understanding of what parameters are necessary for a generated composition nearly indistinguishable from that of a master composer.
Comments: 13 pages, 6 figures Update: Revised format to align closer to IEEE standards
Subjects: Sound (cs.SD); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Audio and Speech Processing (eess.AS)
ACM classes: I.2; I.5; I.7; J.5
Cite as: arXiv:2101.00169 [cs.SD]
  (or arXiv:2101.00169v3 [cs.SD] for this version)

Submission history

From: Daniel Szelogowski [view email]
[v1] Fri, 1 Jan 2021 05:40:12 GMT (1006kb)
[v2] Mon, 19 Apr 2021 22:40:56 GMT (375kb)
[v3] Sat, 13 Nov 2021 01:26:04 GMT (375kb)

Link back to: arXiv, form interface, contact.