2 results
Search Results
Now showing 1 - 2 of 2
- ItemClimatic windows for human migration out of Africa in the past 300,000 years([London] : Nature Publishing Group UK, 2021) Beyer, Robert M.; Krapp, Mario; Eriksson, Anders; Manica, AndreaWhilst an African origin of modern humans is well established, the timings and routes of their expansions into Eurasia are the subject of heated debate, due to the scarcity of fossils and the lack of suitably old ancient DNA. Here, we use high-resolution palaeoclimate reconstructions to estimate how difficult it would have been for humans in terms of rainfall availability to leave the African continent in the past 300k years. We then combine these results with an anthropologically and ecologically motivated estimate of the minimum level of rainfall required by hunter-gatherers to survive, allowing us to reconstruct when, and along which geographic paths, expansions out of Africa would have been climatically feasible. The estimated timings and routes of potential contact with Eurasia are compatible with archaeological and genetic evidence of human expansions out of Africa, highlighting the key role of palaeoclimate variability for modern human dispersals.
- ItemEvolutionary design of explainable algorithms for biomedical image segmentation([London] : Nature Publishing Group UK, 2023) Cortacero, Kévin; McKenzie, Brienne; Müller, Sabina; Khazen, Roxana; Lafouresse, Fanny; Corsaut, Gaëlle; Van Acker, Nathalie; Frenois, François-Xavier; Lamant, Laurence; Meyer, Nicolas; Vergier, Béatrice; Wilson, Dennis G.; Luga, Hervé; Staufer, Oskar; Dustin, Michael L.; Valitutti, Salvatore; Cussat-Blanc, SylvainAn unresolved issue in contemporary biomedicine is the overwhelming number and diversity of complex images that require annotation, analysis and interpretation. Recent advances in Deep Learning have revolutionized the field of computer vision, creating algorithms that compete with human experts in image segmentation tasks. However, these frameworks require large human-annotated datasets for training and the resulting “black box” models are difficult to interpret. In this study, we introduce Kartezio, a modular Cartesian Genetic Programming-based computational strategy that generates fully transparent and easily interpretable image processing pipelines by iteratively assembling and parameterizing computer vision functions. The pipelines thus generated exhibit comparable precision to state-of-the-art Deep Learning approaches on instance segmentation tasks, while requiring drastically smaller training datasets. This Few-Shot Learning method confers tremendous flexibility, speed, and functionality to this approach. We then deploy Kartezio to solve a series of semantic and instance segmentation problems, and demonstrate its utility across diverse images ranging from multiplexed tissue histopathology images to high resolution microscopy images. While the flexibility, robustness and practical utility of Kartezio make this fully explicable evolutionary designer a potential game-changer in the field of biomedical image processing, Kartezio remains complementary and potentially auxiliary to mainstream Deep Learning approaches.