Phenotype matching is a main mechanism of kin recognition, defined as the ability to recognize kin based on physical traits. Recognizing kin is advantageous to increase inclusive fitness through social preferences for closely related kin and, simultaneously, to avoid mating with close kin. In primates, facial cues have been shown to play a crucial role in identifying kin, but evidence for visual phenotype matching remains limited and is often confounded by familiarity. Recent advances in artificial intelligence (AI) and machine learning offer novel approaches to explore subtle phenotypic cues at a finer resolution and in an unbiased manner. The use of AI can therefore significantly improve our understanding of how individuals recognize kin, by identifying and analyzing specific facial features linked to genetic relatedness.

In my PhD project, I aim to use AI technology to investigate visual phenotype matching in free-ranging rhesus macaques living at the Cayo Santiago population in Puerto Rico (USA). By applying AI models to a large dataset of macaque facial images and combining these results with genetic, demographic, and behavioral data, I will explore the extent to which facial cues enable the inference of relatedness by AI and the monkeys themselves. I plan to further validate the facial traits identified by AI through behavioral experiments with free-ranging macaques. This validation will follow an initial pilot study, where human raters evaluate relatedness based on macaque faces. Given that comprehensive genomic data are available for the study population, I further aim to test, whether AI is able to predict realized relatedness (identity-by-descent) from facial cues and how these predictions relate to social behavior.

Education

Work & Field Experience

Funding & Awards