MusicLM: The Future of AI-Powered Music Creation — an Overview

Author(s): Deltan Lobo

Originally published on Towards AI.

This member-only story is on us. Upgrade to access all of Medium.

When I was 12 and started playing the piano, it took me a while to become handy with the instrument. Now that I’m 22, after learning a few techniques, music theory(which was a nightmare for me at that time), and in addition to the piano, I play guitar, violin, and harmonica. Most things eventually changed when ChatGPT came along.

AI has completely transformed most of the creative professions, and music is no exception. One of the most recent breakthroughs in the field was a new AI model called MusicLM, produced by researchers from Google. By using MusicLM, good coherent music can be generated directly from text descriptions such as “a smooth violin melody accompanied by a distorted riff of a guitar.” When I listened to this though it was AI-generated, it was hard at first to identify the player and whether human or AI.

Using advanced methodologies like hierarchical sequence modeling, multimodal conditioning, and large-scale training data, MusicLM opens endless possibilities for artists, educators, and media creators. In looking into the methodology, techniques, applications, and broader implications of MusicLM, this paper gives an in-depth look… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI