The Argument Classifier
Egyptian fruit bat · A machine-learning classifier trained on ~15,000 real bat vocal interactions could tell what the bats were arguing about, well above chance. Can you?
The samples here are pitched-shifted synthesized fingerprints — not raw hydrophone audio — whose spectral traits mirror the canonical characteristics of each context in Prat et al. 2016. Same shapes, audible pitch. Cite them like data, not like recordings.
call 1 · what are they arguing about?
guessed 0
right 0
accuracy —