Recent advances in computational ethology

Recent advances in computational ethology

New advances in computer vision and machine learning have led to a great boom of new techniques to track animals from video data and analyze their behavior. For a quick overview of the number of different techniques developed in computational ethology I recommend a paper compilation on Github by An Liang.

New Techniques

Some of the most popular methods are:

These methods all differ in their user-friendliness for non-software developers, but also in the techniques used for marker-based or pixel-based pose estimation, overall model architecture, individual or social scene complexity, behavior clustering method, use of supervised or unsupervised machine learning, and the degree of self-sufficiency.

Fields of research

These methods have been used so far to analyze behavior (ethology) and the link between behavior and neural activity (neuroethology) in animals such as humans, chimpanzees, macaques, cats, mice, rats, bats, octopus, lizards, ants, and many many more. See for yourself:

Exercise

Choose one of the Tweets above and try to find out its background. Who posted it? Where does she work? What is her role? Can you find her research online? How has she used DeepLabCut, and what for?

Skim the web for 10 minutes and present your chosen tweet to the group.