Perspektywy muzycznej sztucznej inteligencji

  • Author: Wojciech Krzyżanowski
  • Institution: Uniwersytet im. Adama Mickiewicza w Poznaniu
  • ORCID: https://orcid.org/0000-0003-3576-5735
  • Year of publication: 2022
  • Source: Show
  • Pages: 67-76
  • DOI Address: https://doi.org/10.15804/CDM.2022.1.07
  • PDF: cdm/1/cdm107.pdf

The Musical Perspectives of Artificial Intelligence

Recently we have been witnessing an amazing outburst of visual art done by Artificial Intelligence based on texts prompts. Will it ever be able to exhibit such a creative approach towards music? Before we answer, we should learn some basics of these systems. Artificial Intelligence powered by machine learning becomes more effective through exposure to empirical data. By analysing similarities and differences between hundreds of thousands of photographs of violin, it understands what a violin can look like, without being given any other explicit instruction. Machine learning systems specialize in recognizing objects and GAN networks can also generate new objects based on the analysed data. As for now there is no highly effective and diverse system to produce music based only on informal text prompts. But its existence is theoretically possible. We can even envision more complex systems, which automate such output around biological user feedback.

REFERENCES:

artificial intelligence GAN networks DALL·E biomusicology machine learning

Wiadomość do:

 

 

© 2017 Adam Marszałek Publishing House. All rights reserved.

Projekt i wykonanie Pollyart