Behavior is the key to the mind, but can machines recognize and classify behavior by automated means? This is what biopsychologists and computer scientists could demonstrate by using the deep-learning-based pose-estimation tool DeepLabCut. By establishing supervised machine learning predictive classifiers for pigeon behaviors, they created an open-source software package that recognizes eating, walking, head shaking, tail shaking, preening, and fluffing with high precision and could thus analyze time-series analyses of behavioral sequences.

Wittek, N., Wittek, K., Keibel, C., Güntürkün, O., Supervised machine learning aided behavior classification of pigeons, Behavior Research Methods, 2022, https://doi.org/10.3758/s13428-022-01881-w