Introduction to Machine Learning for Game Audio

Presenters:
Krystle Becknauld
John Byrd, Gigantic Software - Santa Ana, CA, USA

In the past four years, every tech industry has put machine learning systems into production. Virtual assistants, cybersecurity systems, and self-driving cars are all built on machine learning. However, none of these novel algorithms are being used in the interactive entertainment industries.

This talk summarizes the most important new research in this exciting new field and describes novel applications of machine learning, specifically for interactive audio. You'll see demonstrations and code for: automatic feature extraction, classification, silence removal, tempo extraction, automatic speaker detection, emotion detection, convolutional and recurrent neural networks, variational autoencoders, and WaveNet. Additionally, this talk will show you the state of the art in interactive audio, and it will also tell you what will be possible, until the year 2030.