Understanding how animals interact socially is crucial for studying their behavior and its implications in neuroscience and ecology. Researchers at the Brain Cognition and Brain Disease Institute (BCBDI) of the Shenzhen Institute of Advanced Technology of the Chinese Academy of Sciences have introduced a new method called the Social Behavior Atlas (SBeA), which uses artificial intelligence (AI) for tracking and analyzing the behavior of multiple animals in three dimensions (3D).
Published in Nature Machine Intelligence, this study presents a novel approach to multi-animal behavior analysis. Unlike traditional methods that require predefined categories of social behavior, the SBeA framework utilizes a few-shot learning AI algorithm. This algorithm can accurately identify similar-looking animals with over 90 percent accuracy, allowing for the discovery of previously undefined differences in animal social behavior.
One of the key advantages of the SBeA framework is its ability to synthesize large amounts of data and train models with higher accuracy. According to Wei Pengfei, the corresponding author of the study, this results in more precise estimation of 3D social gestures.
The SBeA technology is particularly effective for calculating 3D social posture, identifying individual animals, and analyzing subtle social interactions in various species such as mice, birds, and domestic dogs. Furthermore, its potential for cross-species applications opens up new avenues for studying social behavior across different types of animals.
