A Composer's Guide to Creating with Generative Neural Networks

Explore the intersection of artificial intelligence and music composition in this 43-minute conference talk from GOTO Chicago 2023. Delve into a case study of an independent composer's journey to build, optimize, and train a deep learning model for audio generation. Learn about the challenges of data augmentation, solo coding, and limited resources faced by artists in the AI space. Discover the process of developing creative work that balances aesthetic, conceptual, and ethical considerations. Gain insights into data collection, model selection, infrastructure setup, development strategies, and composition techniques. Examine future directions for AI in music and participate in a creative exercise. This talk offers a unique perspective on the emerging field of generative music creation and catalogs areas for future research to facilitate the use of deep learning models in creative contexts.

Intro
What's in this presentation, and what not?
Outline
Background & goals
Data collection
Model selection
Infrastructure
Development
Deciding when it's done
Composition
Future directions
Creative exercise
Gratitude
Outro