Scientists have recorded the first footage of a giant squid hunting in the wild. The elusive creatures are notoriously difficult to film, as their habitat is housands of feet under the sea, where it's dark and the crushing pressure of water requires specialist equipment. While several dead specimens have washed up on shore, the first still images of a living giant squid in the wild weren't recorded until 2004 and video wasn't obtained until 2012. Now marine biologists have, for the first time, captured footage of Architeuthis dux hunting prey in the wild. The footage was captured in 2019, but researchers have now released analysis of the creature's behavior.

Effective conservation actions require effective population monitoring. However, accurately counting animals in the wild to inform conservation decision-making is difficult. Monitoring populations through image sampling has made data collection cheaper, wide-reaching and less intrusive but created a need to process and analyse this data efficiently. Counting animals from such data is challenging, particularly when densely packed in noisy images. Attempting this manually is slow and expensive, while traditional computer vision methods are limited in their generalisability. Deep learning is the state-of-the-art method for many computer vision tasks, but it has yet to be properly explored to count animals. To this end, we employ deep learning, with a density-based regression approach, to count fish in low-resolution sonar images. We introduce a large dataset of sonar videos, deployed to record wild mullet schools (Mugil liza), with a subset of 500 labelled images. We utilise abundant unlabelled data in a self-supervised task to improve the supervised counting task. For the first time in this context, by introducing uncertainty quantification, we improve model training and provide an accompanying measure of prediction uncertainty for more informed biological decision-making. Finally, we demonstrate the generalisability of our proposed counting framework through testing it on a recent benchmark dataset of high-resolution annotated underwater images from varying habitats (DeepFish). From experiments on both contrasting datasets, we demonstrate our network outperforms the few other deep learning models implemented for solving this task. By providing an open-source framework along with training data, our study puts forth an efficient deep learning template for crowd counting aquatic animals thereby contributing effective methods to assess natural populations from the ever-increasing visual data.

Genghis Khan got his dying wish: despite attempts by archaeologists and scientists to find the Mongolian ruler's final resting place, the location remains a secret 800 years after his death. The search for his tomb, though, has inspired an innovative project that could help protect polar bears. "I randomly tuned into the radio one night and heard an expert talking about the use of synthetic aperture radar [SAR] to look for Genghis Khan's tomb," says Tom Smith, associate professor in plant and wildlife sciences at Brigham Young University (BYU) in Utah. "They were using SAR to penetrate layers of forest canopy in upper Mongolia, looking for the ruins of a burial structure." Talking to engineers, including BYU's Dr David Long, Smith learned that SAR is used by the military to detect enemy camps, tanks and vehicles hidden beneath camouflage and is being studied as a potential tool for finding avalanche survivors.

Progress within physical oceanography has been concurrent with the increasing sophistication of tools available for its study. The incorporation of machine learning (ML) techniques offers exciting possibilities for advancing the capacity and speed of established methods and also for making substantial and serendipitous discoveries. Beyond vast amounts of complex data ubiquitous in many modern scientific fields, the study of the ocean poses a combination of unique challenges that ML can help address. The observational data available is largely spatially sparse, limited to the surface, and with few time series spanning more than a handful of decades. Important timescales span seconds to millennia, with strong scale interactions and numerical modelling efforts complicated by details such as coastlines. This review covers the current scientific insight offered by applying ML and points to where there is imminent potential. We cover the main three branches of the field: observations, theory, and numerical modelling. Highlighting both challenges and opportunities, we discuss both the historical context and salient ML tools. We focus on the use of ML in situ sampling and satellite observations, and the extent to which ML applications can advance theoretical oceanographic exploration, as well as aid numerical simulations. Applications that are also covered include model error and bias correction and current and potential use within data assimilation. While not without risk, there is great interest in the potential benefits of oceanographic ML applications; this review caters to this interest within the research community.

SpaceX successfully launches NASA astronauts from Kennedy Space Center into space. The International Space Station's population swelled to 11 on Saturday with the jubilant arrival of SpaceX's third crew capsule in less than a year. All of the astronauts -- representing the U.S., Russia, Japan and France -- managed to squeeze into camera view for a congratulatory call from the leaders of their space agencies. This image provided by NASA, astronauts from SpaceX join the astronauts of the International Space Station for an interview on Saturday, April 24, 2021. A recycled SpaceX capsule carrying four astronauts has arrived at the International Space Station, a day after launching from Florida.

If there's one thing the United States military gets right, it's lethality. Yet even once the US military has you in its sights, it may not know who you actually are -- such are, these so-called "signature strikes" -- even as that wrathful finger of God is called down from upon on high. As Kate Crawford, Microsoft Research principal and co-founder of the AI Now Institute at NYU, lays out in this fascinating excerpt from her new book, Atlas of AI, the military-industrial complex is alive and well and now leveraging metadata surveillance scores derived by IBM to decide which home/commute/gender reveal party to drone strike next. And if you think that same insidious technology isn't already trickling down to infest the domestic economy, I have a credit score to sell you. Excerpted from Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence by Kate Crawford, published by Yale University Press.

By sending Mississippi waters on a new course, engineers hope to build new land—and test ways to save a retreating coast. In a swamp at the edge of Louisiana's Barataria Bay, plastic-capped GPS antennas sprout like oversize mushrooms from four small wooden platforms. The gear, which helps scientists monitor changes in the surrounding marsh, is easy to miss in this expanse of water and swampland the size of Delaware. But it represents something even bigger: the beginnings of a grand ecosystem engineering experiment that has been 50 years in the making and could ultimately cost some $50 billion. If all goes as planned, 2 years from now engineers will punch a massive hole in a nearby levee that holds back the Mississippi River. A 3.5-kilometer-long canal will carry sand and muck from the muddy river into the bay, helping rebuild vast wetlands eroded by sinking land and rising seas. Over 5 decades, researchers forecast that the project—formally known as the Mid-Barataria Sediment Diversion—could move enough sediment to bury the island of Manhattan under 3 meters of muck and create at least 54 square kilometers of new wetlands. The diversion, expected to cost $2 billion, is a critical part of a much larger effort aimed at preventing coastal Louisiana, and the human and wild communities it supports, from slipping beneath the sea. “There's is nothing like [it] anywhere in the world,” says coastal geoscientist Torbjörn Törnqvist of Tulane University, one of a small army of researchers who have helped shape the project through years of fieldwork, computer models, and even the use of a giant replica of the Mississippi River in Massachusetts. “It's going to be completely new.” The project, which will be financed with money paid by oil giant BP after the 2010 Deepwater Horizon oil spill, last month reached a key regulatory milestone and could begin by 2023. But it is not universally loved. Fishers who make their living pulling oysters and shrimp from nearby waters fear the influx of freshwater could harm their livelihoods. Local officials are questioning the cost. And some scientists are skeptical it will achieve the desired results, given the complexity of trying to tinker with one of the world's largest and most dynamic rivers. But others are thrilled that the experiment may finally get underway, saying it could help heal the Louisiana coast and shape similar restoration projects elsewhere. “There's no question that this is huge,” says John Lopez, a coastal scientist who was an early architect of coastwide restoration proposals for Louisiana and now runs a private consulting firm that advises environmental groups about the issue. He hopes it will become “a model for other areas in the world.” THE PROBLEM the Barataria diversion aims to address has been 150 years in the making. For millennia, the Mississippi built coastal Louisiana, depositing dirt and sand as it flooded and flipped from one channel to another. This land building explains why the river's delta pokes into the Gulf of Mexico, a swampy snout teeming with wildlife, marshy grasslands, and watery bayous darkened by ancient cypress trees. In the late 1800s, people started to build a network of levees along the river's southern reaches to reduce flooding and ease navigation. Today, the Mississippi is little more than a massive canal, straitjacketed by earthen walls and traversed by massive freighters moving cargo bound for places as far north as Wisconsin. As the river pours into the Gulf of Mexico, just a trickle of its waters still flow into surrounding bays such as Barataria to the west and Breton Sound to the east. Cut off from their steady supply of sediment, these marshlands have been sinking for 100 years. By 2016, the state had lost an estimated 4800 square kilometers of coastland since the 1930s. Nearly one-quarter of that loss occurred in Barataria Bay, and nearly 500 square kilometers in nearby Breton Sound, according to the U.S. Geological Survey. ![Figure][1] CREDITS: (MAP) K. FRANKLIN/ SCIENCE ; (DATA) MID-BARATARIA SEDIMENT DIVERSION ENVIRONMENTAL IMPACT STATEMENT , U.S. ARMY CORPS OF ENGINEERS, 2021 Decades ago, scientists first suggested trying to restore the marshland by re-engineering part of the Mississippi. In 1975, Sherwood “Woody” Gagliano, a geographer and private environmental consultant who pioneered research into Louisiana's coastal erosion, mapped an elaborate system of levees and canals that could funnel freshwater and sediment into Barataria Bay. “The mistakes that we have made in our coastal areas in the past were largely through ignorance,” wrote Gagliano, who died in 2020. “If we … [continue] to make the same errors, destroying the invaluable renewable resources, the action is inexcusable.” For decades, however, little happened. Then, in 2005, Hurricane Katrina slammed into southern Louisiana, killing more than 1800 people, causing $161 billion in damage, and devastating New Orleans, just north of Barataria Bay. The loss of wetlands that had once blunted storms was partly to blame, scientists agreed. And restoration plans previously confined to scientific conferences and bureaucratic meetings suddenly took center stage. In 2007, state lawmakers endorsed the first iteration of what has become known as the coastal master plan, a 50-year initiative to salvage the state's coastline from the forces eating away at it: sinking land, rising seas, and channels dug for the oil and gas industry. The plan includes a variety of measures: rebuilding barrier islands with new sand, hauling dredged muck to replenish drowning marshes, strengthening levees and flood barriers, and raising buildings above projected flood levels. But the most novel and ambitious piece is an updated version of Gagliano's vision of harnessing the Mississippi's power to build new land. In its latest iteration, the plan calls for creating 11 diversions along the Louisiana coastline. The first and biggest would be the massive mid-Barataria diversion, which would puncture the river's western bank. From the beginning, money posed a major barrier. Completing all the projects would cost at least $50 billion, the state estimates, with the 11 diversions alone costing $5 billion. Then BP's Deepwater Horizon drilling platform in the Gulf of Mexico exploded, unleashing the world's largest oil spill. The resulting fines promise to pour more than $7 billion into Louisiana's coffers, with much of the money earmarked for environmental restoration. At last, planners not only had a vision, but also the money to start to realize it. Last month, the plans for the mid-Barataria diversion passed a key landmark with the release of the preliminary draft of a massive, 7780-page environmental impact study by the U.S. Army Corps of Engineers (USACE). Although the study found the project could have some negative impacts on fisheries and wildlife, diversion proponents welcomed the finding that it would build significant new marshlands. That could prove critical for winning federal permits. IN THE MID-1980s , wildlife biologist Andy Nyman witnessed one of the first small experiments to examine whether punching a hole in a Mississippi levee could help rebuild bay wetlands. The state Department of Wildlife and Fisheries excavated gaps in land separating the river from several bays near where the river flows into the Gulf. At first, he recalls, he was unimpressed. After 3 months, there was little evidence of success. “I was like, ‘Man, this thing hasn't done anything.’ I thought it was a tremendous waste of time.” Twenty years later, however, Nyman, who is now at Louisiana State University (LSU), Baton Rouge, took a boat out to the area near the river's mouth. He discovered more than 100 hectares of reeds, cattails, and seagrasses now blanketed what had been open water. “It was so cool.” Many scientists predict the mid-Barataria diversion will conjure such scenes on a much bigger scale. But other experiments have shown forecasts are tricky. Just downstream of Barataria Bay, near the river's mouth, an effort called the West Bay project has helped create a string of sandy islands covering some 160 hectares. But when work began there in 2003, researchers expected nearly 10 times more land to form by now. What happened? One problem was that water poured through the diversion so quickly that it flushed much of the new sediment into deeper water, says Alex Kolker, a coastal geologist at the Louisiana Universities Marine Consortium who has studied West Bay. On top of that, the land beneath the bay was sinking unusually fast because of the underlying geology, making it hard for the new sediment to compensate. Land did start to emerge after USACE built artificial islands in the bay with dredged sand. The islands then acted like speed bumps, slowing the diverted water and allowing more sediment to accumulate. Such complexities make it hard to predict exactly how much land the new Barataria diversion will create. Last year, officials at Louisiana's Coastal Protection and Restoration Authority released a figure of up to 121 square kilometers of land over 50 years, based on hydrologic models developed by scientists in the Netherlands, a country renowned for its water management. In its recent environmental study, however, USACE cut that number by more than half, to just 54 square kilometers. The reduction reflected higher projections for sea level rise. Even the latest estimate could be overly optimistic, data from the antenna-sprouting platforms have suggested. In 2016, a team led by Törnqvist drilled three core samples, each as wide as a saucer, from the muck beneath the platforms. The deepest plunged almost 40 meters, through 10,000 years of sediments. Scientists then mounted the GPS units on poles cemented into the holes at different depths, to measure whether the marsh is sinking and by how much. Analyses of sediment cores extracted from the deepest hole suggest buried layers, particularly peat, will compress when weight is added. As a result, new sediment might cause the marsh to sink, erasing much of the gains. Studies of the core samples suggest the land gain over 50 years could be as little as 1 centimeter, or as much as 1.75 meters, says Molly Keogh, a wetland geologist who led the research while earning a Ph.D. at Tulane. The range is large in part because of uncertainty about how much sediment the river will deliver. (Keogh's estimates don't include sea level rise, which could further erode the gains.) Biology adds an uncertainty of its own. Eugene Turner, a coastal ecologist at LSU, notes that the Mississippi is often loaded with nutrients washed from farm fields farther upstream. As the new canal dumps those nutrients into the bay, he fears they could cause marsh plants to grow too fast to develop root systems hefty enough to hold sediment in place, leading to marsh erosion. “My concern is [the additional nutrients] will overwhelm the benefits,” he says. Tracy Quirk, a wetland plant ecologist at LSU, has been testing that idea. She has been adding different amounts of nutrients and sediments to more than 130 wetland plots in Barataria Bay, each 1 square meter in size. Two years of observations suggest even low levels of added nutrients catalyze root growth, she says. But, so far, she isn't seeing signs that those roots are weaker. Far from Louisiana, engineers are running experiments aimed at reducing other uncertainties. In a warehouselike building in Massachusetts, they have built a model of the Mississippi and the planned Barataria diversion canal at 1/65th scale. It even includes canoe-size models of the giant freighters that ply the waterway. “Numeric models are good, but they don't get everything perfect,” says Dan Gessler, a civil engineer and vice president at the Alden Research Laboratory, the firm that built the model in Holden, Massachusetts. Using grains of plastic to simulate sand and silt, the researchers have studied how the water will move sediment from the river to the bay. One goal is to fine-tune the design of the canal to cope with the region's relatively flat topography, which could cause the water to slow and drop sediment too soon, clogging the waterway. They also want to be sure the diversion doesn't cause sand bars to form in the main river, disrupting its flow and blocking ships. “When you're spending over a billion dollars,” Gessler says, “you want to be really confident that it's going to work exactly the way [you] expect it to.” But only the full-scale experiment will really dispel the questions, Törnqvist says. “The real interesting thing is going to be when the diversion actually gets going,” he says. “We just can't wait.” THE UNCERTAINTIES OF POLITICS , however, could get in the way. The overall coastal plan enjoys bipartisan support from state and federal lawmakers. Major environmental groups, including the Environmental Defense Fund and the National Audubon Society, have endorsed the diversions. “This is a critical part of restoring the natural processes that built and sustained Louisiana,” says Alisha Renfro, a sedimentologist with the National Wildlife Federation. But some local governments and fishing interests warn that the mid-Barataria diversion threatens their way of life. This month, the governing council for Plaquemines parish, on the eastern edge of Barataria Bay, voted eight to zero to protest the project. “The whole … concept is going to devastate our fishing and seafood industry,” says George Ricks, founder and president of the Save Louisiana Coalition, an alliance of local seafood interests. “There won't be nothing left once they do the diversion.” Scientists don't dispute that adding more freshwater to the bay's ecosystem will bring changes. Marsh plants that like brackish water, and some species of birds and fish, will likely benefit, Nyman says. “But there will be immediate costs” to species that thrive in saltier waters, he adds. For example, oysters and brown shrimp, mainstays of the local fishing industry, could be pushed out of their current ranges. The salinity drop could also cause skin diseases in the bay's 2000 dolphins, reducing their survival by as much as one-third, according to scientists at the National Oceanic and Atmospheric Administration. Other critics question the diversion's cost, given the potential economic upheaval. “Holy mackerel! We're going to spend $2 billion for 12,000 acres?” says coastal oceanographer Joe Suhayda, who retired from LSU in 2002 and now consults for several local governments, including Plaquemines parish. Many observers doubt those downsides will derail the diversion. But even if the project works as promised, they note a sobering reality: Even the most optimistic restoration scenarios see the state losing far more coastline than it will gain. At the state's coastal restoration agency, officials often point to “the big red map.” It forecasts what Louisiana's coastline will look like in 50 years, even if the state completes everything on its $50 billion to-do list. Small patches of green indicate wetlands it expects to save or restore, whereas a thick red fringe along the coast highlights marshes it expects to lose. All told, the state could protect or restore about 2000 square kilometers of wetland, the agency estimates, but other parts of the coast could lose more than 3700 square kilometers. In Barataria Bay alone, wetlands will shrink from 1500 square kilometers today to just 346 square kilometers by 2070 even with the diversion, according to USACE. “We stopped promising a long time ago that we were ever going to be able to recreate a historical coastline,” says Jim Pahl, a plant ecologist and senior scientist at the state coastal authority. Yet diversions can play an important role, he says. Although pumping in sediment with heavy equipment can rebuild patches of marshland, the land will continue to sink and disappear. Because the diversions keep delivering new sand and mud each year, Pahl says, they are critical for the state to “maintain anything resembling a sustainable system.” Törnqvist agrees, even though he concluded in a 2019 study published in Science that efforts to save the Louisiana coast are probably doomed in the very long run if seas continue to rise. The mid-Barataria diversion and similar projects, he says, will be worth the cost if they buy time for vulnerable cities in the delta, including New Orleans, to adapt to climate change. “A few more decades can make the difference in the long run,” he says, “between managed retreat and complete chaos.” [1]: pending:yes

Robotic vehicles have been used in dangerous environments for decades, from decommissioning the Fukushima nuclear power plant or inspecting underwater energy infrastructure in the North Sea. More recently, autonomous vehicles from boats to grocery delivery carts have made the gentle transition from research centres into the real world with very few hiccups. Yet the promised arrival of self-driving cars has not progressed beyond the testing stage. And in one test drive of an Uber self-driving car in 2018, a pedestrian was killed by the vehicle. Although these accidents happen every day when humans are behind the wheel, the public holds driverless cars to far higher safety standards, interpreting one-off accidents as proof that these vehicles are too unsafe to unleash on public roads.

An interpretable system for complex, open-domain reasoning needs an interpretable meaning representation. Natural language is an excellent candidate -- it is both extremely expressive and easy for humans to understand. However, manipulating natural language statements in logically consistent ways is hard. Models have to be precise, yet robust enough to handle variation in how information is expressed. In this paper, we describe ParaPattern, a method for building models to generate logical transformations of diverse natural language inputs without direct human supervision. We use a BART-based model (Lewis et al., 2020) to generate the result of applying a particular logical operation to one or more premise statements. Crucially, we have a largely automated pipeline for scraping and constructing suitable training examples from Wikipedia, which are then paraphrased to give our models the ability to handle lexical variation. We evaluate our models using targeted contrast sets as well as out-of-domain sentence compositions from the QASC dataset (Khot et al., 2020). Our results demonstrate that our operation models are both accurate and flexible.

The past decade has witnessed a groundbreaking rise of machine learning for human language analysis, with current methods capable of automatically accurately recovering various aspects of syntax and semantics - including sentence structure and grounded word meaning - from large data collections. Recent research showed the promise of such tools for analyzing acoustic communication in nonhuman species. We posit that machine learning will be the cornerstone of future collection, processing, and analysis of multimodal streams of data in animal communication studies, including bioacoustic, behavioral, biological, and environmental data. Cetaceans are unique non-human model species as they possess sophisticated acoustic communications, but utilize a very different encoding system that evolved in an aquatic rather than terrestrial medium. Sperm whales, in particular, with their highly-developed neuroanatomical features, cognitive abilities, social structures, and discrete click-based encoding make for an excellent starting point for advanced machine learning tools that can be applied to other animals in the future. This paper details a roadmap toward this goal based on currently existing technology and multidisciplinary scientific community effort. We outline the key elements required for the collection and processing of massive bioacoustic data of sperm whales, detecting their basic communication units and language-like higher-level structures, and validating these models through interactive playback experiments. The technological capabilities developed by such an undertaking are likely to yield cross-applications and advancements in broader communities investigating non-human communication and animal behavioral research.