Breakthroughs, discoveries, and DIY tips sent six days a week. A newly uncovered tyrannosaur leg bone is shaking things up in the dinosaur world. The leg bone uncovered in New Mexico belongs to an unusually large tyrannosaur--the group of dinosaurs that includes the mighty . The shinbone is three feet long and about five inches in diameter, only slightly smaller than the largest known specimen. The giant leg bone is detailed in a study published today in the journal .

'Sharkansas' contains entire fossilized skeletons dating back 320 million years. Breakthroughs, discoveries, and DIY tips sent six days a week. Arkansas is hundreds of miles from the Gulf of Mexico, but it's home to countless sharks . A trove of the fossilized predator's remains are embedded within the Fayetteville Shale --a roughly 350-million-year-old geological formation in the state's northwestern corner. Because a shark's cartilage skeleton decomposes so quickly, they usually only leave teeth behind when they die.

Laplace approximations are popular techniques for endowing deep networks with epistemic uncertainty estimates as they can be applied without altering the predictions of the trained network, and they scale to large models and datasets.

Noisy ships and industry are impacting critically endangered Kemp's ridley sea turtles. Breakthroughs, discoveries, and DIY tips sent six days a week. For the world's smallest sea turtles, life in the ocean is getting pretty noisy. These relatively little turtles (on average they're still 75 to 100 pounds) mostly found in the Gulf of Mexico already face fishing gear accidents, seacraft collisions, plastic pollution, and habitat deterioration, and now excess noise may be harming the critically endangered and rare Kemp's ridley sea turtles (). We say because even though these sea turtles share waters with extremely busy shipping lanes, scientists know very little about their underwater hearing.

Inverse problems describe the task of recovering an underlying signal of interest given observables. Typically, the observables are related via some non-linear forward model applied to the underlying unknown signal. Inverting the non-linear forward model can be computationally expensive, as it often involves computing and inverting a linearization at a series of estimates. Rather than inverting the physics-based model, we instead train a surrogate forward model (emulator) and leverage modern auto-grad libraries to solve for the input within a classical optimization framework. Current methods to train emulators are done in a black box supervised machine learning fashion and fail to take advantage of any existing knowledge of the forward model. In this article, we propose a simple learned weighted average model that embeds linearizations of the forward model around various reference points into the model itself, explicitly incorporating known physics. Grounding the learned model with physics based linearizations improves the forward modeling accuracy and provides richer physics based gradient information during the inversion process leading to more accurate signal recovery. We demonstrate the efficacy on an ocean acoustic tomography (OAT) example that aims to recover ocean sound speed profile (SSP) variations from acoustic observations (e.g.