Search for: All records

Here we present a machine learning framework and model implementation that can learn to simulate a wide variety of challenging physical domains, involving fluids, rigid solids, and deformable materials interacting with one another. Our framework—which we term “Graph Network-based Simulators” (GNS)—represents the state of a physical system with particles, expressed as nodes in a graph, and computes dynamics via learned message-passing. Our results show that our model can generalize from single-timestep predictions with thousands of particles during training, to different initial conditions, thousands of timesteps, and at least an order of magnitude more particles at test time. Our model was robust to hyperparameter choices across various evaluation metrics: the main determinants of long-term performance were the number of message-passing steps, and mitigating the accumulation of error by corrupting the training data with noise. Our GNS framework advances the state-of-the-art in learned physical simulation, and holds promise for solving a wide range of complex forward and inverse problems. 

Southern flounder (Paralichthys lethostigma) exhibit environmental sex determination (ESD), where environmental factors can influence phenotypic sex during early juvenile development but only in the presumed XX female genotype. Warm and cold temperatures masculinize fish with mid-range conditions producing at most 50% females. Due to sexually dimorphic growth, southern flounder fisheries are dependent upon larger females. Wild populations could be at risk of masculinization from ESD due to globally increasing water temperatures. We evaluated the effects of habitat and temperature on wild populations of juvenile southern flounder in North Carolina, USA. While northern habitats averaged temperatures near 23 °C and produced the greatest proportion of females, more southerly habitats exhibited warmer temperatures (>27 °C) and consistently produced male-biased sex ratios (up to 94% male). Rearing flounder in the laboratory under temperature regimes mimicking those of natural habitats recapitulated sex ratio differences observed across the wild populations, providing strong evidence that temperature is a key factor influencing sex ratios in nursery habitats. These studies provide evidence of habitat conditions interacting with ESD to affect a key demographic parameter in an economically important fishery. The temperature ranges that yield male-biased sex ratios are within the scope of predicted increases in ocean temperature under climate change.