Researchers from the Bavarian State Collections of Natural History (SNSB) and LMU München have developed an AI-based method to accurately determine the sex of sheep using only linear measurements of their talus bones. This breakthrough could revolutionize the field of archaeozoology, making it much easier to identify the sex of ancient animal remains. The study, which was recently presented at the 20th IEEE International Conference on e-Science in Osaka, Japan, is a significant step forward in the application of machine learning to the analysis of archaeological finds. Archaeozoology is the study of animal remains found in archaeological sites, and this new AI-powered approach could greatly streamline the process of identifying the sex of sheep and other livestock.
Revolutionizing Archaeozoological Sex Determination
Traditionally, zoologists have relied on morphological analysis to determine the sex of animals, such as assessing the pelvic bone in sheep and cattle. However, this approach is often problematic for archaeological finds, where only fragments of the bones are typically preserved. This makes it challenging even for experts to confidently distinguish between male and female animals.
The new AI-based method developed by the SNSB and LMU München researchers offers a solution to this long-standing challenge. By training various machine-learning algorithms on a dataset of over 240 sheep ankle bones with known sex, the team was able to create a system that can accurately identify the sex of sheep based on just four different measurements of the talus bone, also known as the ankle bone. This technique was then successfully applied to 170 previously unidentified sheep ankle bones from an archaeological site in Mongolia, demonstrating its effectiveness in practical applications.
Harnessing the Power of AI for Archaeozoological Insights
The use of AI in archaeozoology is a relatively new and untapped field, but the potential benefits are significant. As Nadine Schüler, the first author of the study and a scientist at the State Collection for Paleonanatomy Munich (SNSB-SPM) and LMU München, explains, “Machine learning could be the solution, but it is rarely used in archaeozoology. In medicine, AI is already used to classify human bones. Our study is a first step towards applying the methodology to archaeozoological data.”
The AI-based sexing method developed by the research team provides archaeozoologists with a quick and cost-effective way to assess their findings, without the need for invasive DNA analysis or reliance on the subjective interpretations of skeletal morphology. The accuracy rates of the algorithms are impressive, with some variants achieving up to 90% accuracy in correctly identifying the sex of the sheep. This level of precision is a significant improvement over traditional methods, which can be prone to errors, especially when dealing with fragmentary remains.
Unlocking the Secrets of the Talus Bone
The choice of the talus bone as the focus of this study is not arbitrary. As the researchers explain, the ankle bones of sheep and other livestock are relatively small and compact, and they are often well-preserved in archaeological sites. This makes the talus an ideal target for analysis, as it can provide valuable insights into the sex and other characteristics of ancient animal populations.
By leveraging the power of AI, the researchers have demonstrated that the talus bone can serve as a reliable indicator of sex, even when other skeletal elements are not available. This opens up new avenues for archaeozoological research, allowing scientists to extract more detailed information from the limited animal remains that are typically found at excavation sites. As the field of archaeozoology continues to evolve, the integration of AI-based methods like this one will undoubtedly become an increasingly important tool in the quest to unravel the mysteries of the past.
