Looking for mimicry in a snake assemblage using deep learning

Batesian mimicry is a canonical example of evolution by natural selection,
popularized by highly colorful species resembling unrelated models with
astonishing precision. However, Batesian mimicry could also occur in
inconspicuous species and rely on subtle resemblance. Although potentially
widespread, such instances have been rarely investigated, such that the
real frequency of Batesian mimicry has remained largely unknown. To fill
this gap, we developed a new approach using deep learning to quantify
visual resemblance between putative mimics and models from photographs. We
applied this method to Western Palearctic snakes. Potential nonvenomous
mimics were revealed by an excess of resemblance to sympatric venomous
snakes compared to random expectations. We found that 8% of the
non-venomous species were potential mimics, although they resembled their
models imperfectly. This study is the first to quantify the frequency of
Batesian mimicry in a whole community of vertebrates, and shows that even
concealed species can act as potential models. Our approach should prove
useful to detect mimicry in other communities, and more generally it
highlights the benefits of deep learning for quantitative studies of
phenotypic resemblance.