Welcome to Music and AI

ally b
3 min readNov 1, 2020

My name is Ally, and I’m a sophomore at the Lakeside School in Seattle, Washington! I play the violin, participate in competitive math contests, and, when it’s not raining, play both indoor and beach volleyball. But, I’ve always loved coding: the freedom to construct nearly anything, projects sometimes consuming me for hours at a time. From there, it was natural that I became hooked in the AI field.

Sheet music for Paganini’s 24th Caprice — Paganini’s Caprices are famously some of the hardest pieces for violin ever written

In May, I began researching possible camps that could teach me the basics of AI. Enter the Inspirit AI camp. The cumulative project I worked on, after reviewing the basics of coding and Python libraries, was detecting emotions in a human face. Our instructor Rohan Badlani was vital to our group’s success, providing images for our slides, knowledge about certain algorithms or functions, and reviewing the final slideshow for accuracy. Unfortunately, our model had a maximum of only 67% accuracy (we used a Convolutional Neural Network to do so). I had a wonderful time, though, so I decided to apply for the Inspirit AI Ambassador program.

An output from our Emotion Recognition model

Strangely enough, my first experience with AI was when I saw the infamous clip of HAL 9000, the robot from 2001: Space Odyssey, lock Dave in the unforgiving void of space. “I’m sorry, Dave,” HAL mumbled, shutting the pod bay doors. The whole crew ends up dead, as HAL prioritizes executing his commands over the crew’s safety. It left me wondering — what more can AI do?

I’m currently working on a project surrounding the intersection of music and AI, so I will updates on how it’s going, along with current AI technologies that work with musical input. For instance, take Google’s program that selects a composer and emulates their style of music, spitting out a piece that’s similar, but not quite there. How can AI analyze something widely deemed distinctly human and break it down into just frequencies? Where do we lose the humanity, the expression in music? And, how would one match performances of the same piece to each other, even though they may be played differently: tempo-wise, dynamics, volume, and even adding extra notes (like trills and mordents for embellishment?)?

As a newer member of the AI community myself, I hope to introduce new students to AI and make my blog accessible to all. Stay tuned for the next installment of this blog!

Add me on LinkedIn: https://www.linkedin.com/in/ally-bush-3499281b8

  • - — — — — — Ally is a Student Ambassador in the Inspirit AI Student Ambassadors Program. Inspirit AI is a pre-collegiate enrichment program that exposes curious high school students globally to AI through live online classes. Learn more at https://www.inspiritai.com/. — — — — —

--

--