Presentation Details
| Using Computation To Accelerate Research on the Cultural Evolution of Vocal Communication Mason Youngblood. Stony Brook University, Stony Brook, NY, USA |
Abstract
The study of the cultural evolution of vocal communication often involves categorizing units into types (e.g. syllables in songbirds, or notes in humpback whales). While this approach is useful in many cases, it necessarily flattens the complexity and nuance present in real communication systems. In this talk, I argue that modern computational methods reveal vocal communication to be far more continuous than previously assumed. Unsupervised machine learning methods, like variational autoencoders and vision transformers, can project vocalizations into high-dimensional latent spaces, capturing fine-grained variation without manual categorization. Open-source tools like chatter make these techniques accessible, extracting continuous measures of complexity, predictability, and similarity from vocal sequences. These representations can also ground mechanistic studies of cultural evolution: agent-based models can be fit to empirical data via simulation-based inference, enabling rigorous hypothesis testing. Collectively, these approaches, which are becoming widely used to study our own species, have the potential to accelerate research on the cultural evolution of vocal communication.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.