Ge3Net: Inferring Continuous Population Structure Coordinates Along the Genome Richa Rastogi , Arvind S. Kumar, Helgi Hilmarsson, Carlos D. Bustamante, Daniel Mas Montserrat, Alexander G. Ioannidis. NeurIPS workshop on Learning Meaningful Representations of Life (LMRL), 2020 This work introduces a novel deep learning architecture to infer fine grain local ancestry along the genome with simple to use geographical coordinates. Additionally, we identify detecting changepoints (when local ancestry changes along the genome) as a key technical problem. To this end, our method detects these changepoints when the prediction uncertainty is high and abstains from making erroneous predictions in the corresponding segments. code bibtex