#MuseNet has learned the “fingerprints” of hundreds of composers, trained on thousands and thousands of scores. This is machine learning – specifically a Transformer – so everything it knows about music is “self-taught”, so to speak! 2/6
-
-
Prikaži ovu nit
-
I took some Bach (Prelude from 4th English Suite) and cut holes in it – like the swiss cheese we saw earlier in the lecture! Then I instructed MuseNet to fill in these holes in the style of Bach. I got about 100 bits of music from it which I had to edit and select from! 3/6
Prikaži ovu nit -
This became more exciting when
@RI_Science asked me to arrange it for@chineke4change. Transforming the piece from a keyboard work to orchestra let me make creative decisions – turning it into a properly creative exercise involving me, some AI-generated material and Bach. 4/6Prikaži ovu nit -
I first tried the experiment with
@MarcusduSautoy at@Barbican concert performed by the fantastic@MahanEsfahani on harpsichord who spoke about the “feel” of performing Bach vs#AI material. 5/6Prikaži ovu nit -
I always thought of Bach as a little like code – both have clarity, elegance and purpose – but comparing them let me consider what creativity is, or can be, & really helped me in my own composition. 6/6
Prikaži ovu nit
Kraj razgovora
Novi razgovor -
-
-
So great to meet you the other week Robert! Let’s fuse AI music with AI animation one day

-
Yes!! I'll keep my feelers out for the right project
Kraj razgovora
Novi razgovor -
-
-
Thank you for being absolutely incredibly throughout the whole process!
Hvala. Twitter će to iskoristiti za poboljšanje vaše vremenske crte. PoništiPoništi
-
-
-
Are Machines offended if there is an AI Critic App that calls them derivative!?
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.