Exploring bat song syllable representations in self-supervised audio encoders
This work provides initial steps for cross-species transfer learning in bat bioacoustics and improves understanding of out-of-distribution audio processing, but it is incremental as it builds on existing encoder models without introducing new methods.
The study investigated whether self-supervised audio encoders, pre-trained on human-generated sounds, can distinguish between bat song syllable types, finding that models trained on human speech produced the most distinctive representations.
How well can deep learning models trained on human-generated sounds distinguish between another species' vocalization types? We analyze the encoding of bat song syllables in several self-supervised audio encoders, and find that models pre-trained on human speech generate the most distinctive representations of different syllable types. These findings form first steps towards the application of cross-species transfer learning in bat bioacoustics, as well as an improved understanding of out-of-distribution signal processing in audio encoder models.