About vak#

The vak library has two main goals:

  1. make it easier for researchers studying animal vocalizations to apply neural network algorithms to their data

  2. provide a common framework for benchmarking neural network algorithms on tasks related to animal vocalizations

Neural network algorithms in vak help answer questions about vocal behavior. We use the term “vocal behavior” to encompass related research areas including animal communication [1], cultural evolution [2], and vocal learning [3] [4]. Models in the vak library include deep learning algorithms developed for bioacoustics [5], but are designed specifically for computational studies of vocal behavior [6].

The library was developed by David Nicholson and Yarden Cohen for experiments assessing performance of TweetyNet, a neural network that automates annotation of birdsong, by segmenting spectograms into the units of song, called syllables.