Classifying the Unknown: Identification of Insects by Deep Zero-shot Bayesian Learning

Insects represent a large majority of biodiversity on Earth, yet so few species are described. Describing new species typicallyrequires specific taxonomic expertise to identify morphological characters that distinguish it from other known species andDNA-based methods have aided in providing additional evidence of separate species. Machine learning (ML) provides apowerful method in identifying new species given its analytical processing is more sensitive to subtle physical differencesin images humans may not process. Existing ML algorithms are limited by image repositories that only contain describedspecies, leaving out the possibility of identifying new species. We develop a Bayesian deep learning method for zero-shotclassification of species. The proposed approach forms a Bayesian hierarchy of species around corresponding genera anduses deep embeddings of images and DNA barcodes to identify insects to the lowest taxonomic level possible. To demonstratethis proof of concept, we use a database of 32,848 insect images from 1,040 described species split into training and test datawherein the test data includes 243 species not present in the training data. Our results demonstrate that using DNA sequencesand images together, known insects can be classified with 96.66% accuracy while unknown (to the database) insects have anaccuracy of 81.39% in identifying the correct genus. The proposed deep zero-shot Bayesian model demonstrates a powerfulnew approach that can be used for the gargantuan task of identifying new insect species.