 |
| ДЕНЬ ДУРАКА |
Thursday, April 01, 2021
Where we live can affect male reproductive health, finds new study
New research, led by scientists at the University of Nottingham, suggests that the environment in which men live may affect their reproductive health.
The research, published in Scientific Reports, looked at the effects of geographical location on polluting chemicals found in dog testes, some of which are known to affect reproductive health. The unique research focused on dogs because, as a popular pet, they share the same environment as people and are effectively exposed to the same household chemicals as their owners.
The team also looked for signs of abnormalities in the testes. The findings showed that both the chemicals present and the extent of abnormalities in the testes were different depending on where the dog's had been living.
The researchers analysed the testes of dogs, which had been removed for routine clinical reasons, to see what polluting chemicals were present in the tissue. Samples were taken from across the UK, in the East and West Midlands, and the South East, as well as from Denmark and Finland.
Dr Rebecca Sumner, from the School of Veterinary Medicine and Science at the University, and lead author of the study, said: "For the first time, we have shown that the profile of chemical pollutants found in dog testes depends on where they are from. We have also shown that the same cohorts of dog testes also show geographic differences in testicular pathology and evidence of an imbalance in cells that are important for sperm production."
Dr Richard Lea, lead of the team, said: "Although this study suggests that there are fewer pathologies in dog testes from Finland compared to other locations, relating this to the chemicals detected is difficult, particularly as many other pollutants may also be present.
"We believe, that this study is of pivotal importance since our strategy to use the dog as a sentinel species for the human has allowed us to focus directly on the testis, where detected chemicals are likely to influence male reproductive function." Professor Gary England, Dean of School of Veterinary Medicine & Science, said "This work is significant since collectively, these findings indicate that environmental exposures are determined by location and this may underpin regional differences in male reproductive health."
Genome sequencing shows coronavirus variation drives pandemic surges
Fusing classical epidemiology and genomics is a tool for future pandemics
Genome sequencing of thousands of SARS-CoV-2 samples shows that surges of COVID-19 cases are driven by the appearance of new coronavirus variants, according to new research from the School of Veterinary Medicine at the University of California, Davis published April 1 in Scientific Reports.
"As variants emerge, you're going to get new outbreaks," said Bart Weimer, professor of population health and reproduction at the UC Davis School of Veterinary Medicine. The merger of classical epidemiology with genomics provides a tool public health authorities could use to predict the course of pandemics, whether of coronavirus, influenza or some new pathogen.
Although it has just 15 genes, SARS-CoV-2 is constantly mutating. Most of these changes make very little difference but sometimes the virus becomes more or less transmissible.
Weimer and graduate student DJ Darwin R. Bandoy initially analyzed the genomes of 150 SARS-CoV-2 strains, mostly from outbreaks in Asia prior to March 1, 2020, as well as epidemiology and transmission information for those outbreaks. They classified outbreaks by stage: index (no outbreak), takeoff, exponential growth and decline. The ease of transmission of a virus is set by the value R, or reproductive number, where R is the average number of new infections caused by each infected person.
They combined all this information into a metric called GENI, for pathogen genome identity. Comparing GENI scores with the phase of an epidemic showed that an increase in genetic variation immediately preceded exponential growth in cases, for example in South Korea in late February. In Singapore, however, bursts of variation were associated with smaller outbreaks that public health authorities were able to quickly bring under control.
20,000 virus samples
Weimer and Bandoy then looked at 20,000 sequences of SARS-CoV-2 viruses collected and from February to April 2020 in the United Kingdom and compared them with data on cases.
They found that the GENI variation score rose steadily with the number of cases. When the British government imposed a national lockdown in late March, the number of new cases stabilized but the GENI score continued to rise. This shows that measures such as banning gatherings, mask mandates and social distancing are effective in controlling spread of disease in the face of rapid virus evolution.
It could also help explain "superspreader" events when large numbers of people get infected in a single incident where precautions are relaxed.
Weimer said he hopes that public health authorities will take up the approach of measuring virus variation and linking it to the local transmission rate, R.
"In this way you can get a very early warning of when a new outbreak is coming," he said. "Here's a recipe for how to go about it."
###
Parts of the work previously appeared online as a preprint. Bandoy is sponsored by the Philippine California Advanced Research Institute.
Beetle outbreak impacts vary across Colorado forests
IMAGE: A SPRUCE BEETLE-IMPACTED FOREST IN SOUTHWESTERN COLORADO. view more
CREDIT: SARAH HART/ COLORADO STATE UNIVERSITY
It's no secret. Colorado's forests have had a tough time in recent years. While natural disturbances such as insect outbreaks and wildfires occurred historically and maintained forest health over time, multiple, simultaneous insect disturbances in the greater region over the past two decades have led to rapid changes in the state's forests.
A bird's eye view can reveal much about these changes. Annual aerial surveys conducted by the Colorado State Forest Service and USDA Forest Service have provided yearly snapshots for the state. New collaborative research led by Colorado State University and the University of Wisconsin-Madison now supplements this understanding with even greater spatial detail.
The study, "Effects of Bark Beetle Outbreaks on Forest Landscape Pattern in the Southern Rocky Mountains, U.S.A.," analyzed Landsat satellite imagery between 1997-2019 to quantify how outbreaks of three different insect species have impacted forests across high-elevation forests in Colorado, southern Wyoming, and northern New Mexico. The research team found that while these collective beetle outbreaks impacted around 40 percent of the area studied, the effects of these outbreak varied due to differences in forest structures and species composition across the region.
"In contrast to research that has examined the heterogeneous effects of wildfire on trees, there hasn't been much work on the landscape-level variation in bark beetle effects on forests, particularly across broad areas," said Sarah Hart, co-author and assistant professor in the Forest and Rangeland Stewardship department. "Heterogeneity plays an important role in how these forests will look in the future, where surviving trees will regenerate the forest, and what potential there is for future outbreaks."
Their results indicate that most forest stands affected by insects still have mature trees that can be sources for reestablishing seeds and conditions for the next generation of trees to grow. Areas with tree mortality greater than 90 percent were relatively small and isolated. Unlike severe wildfires that can kill all trees in its path, trees typically survive bark beetle outbreaks, facilitating forest recovery in upcoming decades.
High-resolution, field-level accuracy
Widespread outbreaks of three important bark beetle species have occurred in Colorado's forests since the turn of the century: mountain pine beetle, spruce beetle, and the western balsam beetle (that affects various fir tree species). These bark beetles primarily target large trees with reduced defenses due to lower precipitation amounts and higher temperature trends since the turn of the century.
This research team combined satellite imagery capable of identifying small groups of dead trees with a decade of extensive field data from nearly 250 plots to develop presence and severity maps for tree mortality caused by bark beetle attacks. Having this data combination gave the research team detailed information about how many trees have died in particular places, and helped to identify what may still be causing the death of individual trees.
"These maps give us unique insight into the effects of recent insect outbreaks because they span a large area but also show a lot of detail, and we are confident that they are showing us how many trees are dying because technicians counted trees on the ground," Kyle Rodman, lead author and post-doctoral researcher at the University of Wisconsin-Madison said.
The maps the team produced indicate that areas most impacted by bark beetles are concentrated in northern and southwestern Colorado due to higher concentrations of old lodgepole pine and spruce forests which were then infested by mountain pine beetle and spruce beetle, respectively. Western balsam beetle impacts were also widespread across the region, but these beetles tended to kill fewer trees in any single location.
"Satellite data is a crucial bridge that allows us to take detailed information from individual places and extend this localized knowledge to large areas," Rodman said. "In using these maps, we can see how the forest has changed over the past 20 years during each of these outbreaks."
Fortunately, much of the 25,000 square kilometer study area showed low to moderate levels of tree mortality, with high tree mortality being contained in small and isolated patches averaging only about nine city blocks in overall size.
"People tend to notice what has changed, rather than what has stayed the same," Rodman said. "These forests have changed a lot, but I am hopeful. It will just take a little while for them to recover, but many of these beetle-killed forests are likely to recover within a few decades."
Monkeys experience the visual world the same way people do
When humans look out at a visual landscape like a sunset or a beautiful overlook, we experience something -- we have a conscious awareness of what that scene looks like. This awareness of the visual world around us is central to our everyday existence, but are humans the only species that experiences the world consciously? Or do other non-human animals have the same sort of conscious experience we do?
Scientists and philosophers have asked versions of this question for millennia, yet finding answers -- or even appropriate ways to ask the question -- has proved elusive. But a team of Yale researchers recently devised an ingenious way to try to solve this riddle.
Writing on March 29 in the Proceedings of the National Academy of Sciences, they make the case that one non-human species -- the rhesus macaque -- also has a conscious awareness of the world around it.
"People have wondered for a long time whether animals experience the world the way we do, but it's been difficult to figure out a good way to test this question empirically," said Moshe Shay Ben-Haim, a postdoctoral fellow at Yale and first author of the paper.
Researchers have known for a long time that people can be influenced by unconscious subliminal cues -- visual stimuli presented just outside of our threshold for conscious awareness, said Laurie Santos, a professor of psychology at Yale who is co-senior author of the study along with her colleague Steve Chang, associate professor of psychology and of neuroscience, and Ran Hassin of Hebrew University.
"We tend to show different patterns of learning when presented with subliminal stimuli than we do for consciously experienced, or supraliminal stimuli," she said.
If monkeys show the same "double dissociation" pattern that humans do, it would mean that monkeys probably experience the supraliminally presented stimuli in the same way as people do -- as a conscious visual experience.
Ben-Haim, Santos, and their team thought of a novel way to explore whether macaques also exhibit a difference in learning when stimuli are experienced consciously versus non-consciously.
In a series of experiments, they had monkeys and humans guess whether a target image would appear on the left or right side of a screen. Before the target appeared, participants received a visual cue -- a small star-- on the side opposite of where the target would subsequently appear. The researchers varied whether the cue was presented supraliminally or subliminally. When the cue was presented for a few seconds, human participants successfully learned that the target would appear in the opposite location from the cue. But when the cue was presented subliminally -- quickly enough that it escaped people's conscious perception -- participants showed a different pattern of performance; they continued to choose the side that was subliminally cued, failing to learn the rule that the cue predicted the opposite side.
Surprisingly, the researchers found that monkeys showed exactly the same response patterns as the people did: like humans, macaques were able to successfully look to the target location when the cues were presented consciously, but showed the reverse pattern for subliminal cues. This striking result suggests that monkeys have two levels of processing just as humans do, one of which must be conscious.
"These results show that at least one non-human animal exhibits both non-conscious perception as well as human-like conscious visual awareness." said Ben-Haim. "We now have a new non-verbal method for assessing whether other non-human creatures experience visual awareness in the same way as humans."
###
Other authors of the paper are Yarrow Dunham, Olga Dal-monte, and Nicholas Fagan of Yale.
Faster imaging in rubber x-ray CT imaging helps tires become smarter and more efficient
IMAGE: A SCHEMATIC OF HIGH-SPEED 4D-CT IMAGING SYSTEM & RESULTING 3D IMAGE OF RUBBER FAILURE IN PROGRESS. view more
CREDIT: SUMITOMO RUBBER INDUSTRIES, LTD.
Sumitomo Rubber Industries, Ltd (SRI) and Tohoku University teamed up to increase the speed of 4-Dimensional Computed Tomography (4D-CT) a thousand-fold, making it possible to observe rubber failure in tires in real-time.
This breakthrough will accelerate the development of new tire materials to provide super wear resistance, greater environmental friendliness, and longer service life. It will also aid significantly in the advancement of smart tires.
SRI initially developed 4D-CT as part of the ADVANCED 4D NANO DESIGN, a new materials development technology unveiled in 2015 that enables highly accurate analysis and simulation of the rubber's internal structure from the micro to nanoscale. This analysis ultimately leads to enhanced rubber wear performance in terms of fuel efficiency, wet grip, and wear resistance.
Since the birth of ADVANCED 4D NANO DESIGN, SRI has undertaken new materials development using the technology available at the SPring-8 synchrotron radiation research facility. However, the existing 4D-CT technique still took several seconds to capture a single 3D image.
Wataru Yashiro, associate professor at Tohoku University, had previously developed a new method to produce higher quality CT images within milliseconds. The technology, which used intense synchrotron radiation, made the breakthrough feasible. It became possible to observe rubber failure as it occurs at varying speeds, allowing a better approximation of a tire's rubber wear.
"As a result of this research, we have increased the speed of imaging by approximately a thousand-fold, meaning that a 3D image can now be captured in around 1/100 of a second," said Yashiro.
Looking ahead, SRI hopes to use this technology in conjunction with machine learning and other advanced data processing techniques to analyze the enormous amounts of data that this imaging technique will generate. On the collaborative side, SRI and Tohoku University will continue to serve as participants in the Japan Science and Technological Agency's Core Research for Evolutionary Science and Technology Program (CREST).
###
Mystery of photosynthetic algae evolution finally solved
An evolutionary mystery that had eluded molecular biologists for decades may never have been solved if it weren't for the COVID-19 pandemic.
"Being stuck at home was a blessing in disguise, as there were no experiments that could be done. We just had our computers and lots of time," says Professor Paul Curmi, a structural biologist and molecular biophysicist with UNSW Sydney.
Prof. Curmi is referring to research published this month in Nature Communications that details the painstaking unravelling and reconstruction of a key protein in a single-celled, photosynthetic organism called a cryptophyte, a type of algae that evolved over a billion years ago.
Up until now, how cryptophytes acquired the proteins used to capture and funnel sunlight to be used by the cell had molecular biologists scratching their heads. They already knew that the protein was part of a sort of antenna that the organism used to convert sunlight into energy. They also knew that the cryptophyte had inherited some antenna components from its photosynthetic ancestors - red algae, and before that cyanobacteria, one of the earliest lifeforms on earth that are responsible for stromatolites.
But how the protein structures fit together in the cryptophyte's own, novel antenna structure remained a mystery - until Prof. Curmi, PhD student Harry Rathbone and colleagues from University of Queensland and University of British Columbia pored over the electron microscope images of the antenna protein from a progenitor red algal organism made public by Chinese researchers in March 2020.
Unravelling the mystery meant the team could finally tell the story of how this protein had enabled these ancient single-celled organisms to thrive in the most inhospitable conditions - metres under water with very little direct sunlight to convert into energy.
Prof. Curmi says the major implications of the work are for evolutionary biology.
"We provide a direct link between two very different antenna systems and open the door for discovering exactly how one system evolved into a different system - where both appear to be very efficient in capturing light," he says.
"Photosynthetic algae have many different antenna systems which have the property of being able to capture every available light photon and transferring it to a photosystem protein that converts the light energy to chemical energy."
By working to understand the algal systems, the scientists hope to uncover the fundamental physical principles that underlie the exquisite photon efficiency of these photosynthetic systems. Prof. Curmi says these may one day have application in optical devices including solar energy systems.
EATING FOR TWO
To better appreciate the significance of the protein discovery, it helps to understand the very strange world of single-celled organisms which take the adage "you are what you eat" to a new level.
As study lead author, PhD student Harry Rathbone explains, when a single-celled organism swallows another, it can enter a relationship of endosymbiosis, where one organism lives inside the other and the two become inseparable.
"Often with algae, they'll go and find some lunch - another alga - and they'll decide not to digest it. They'll keep it to do its bidding, essentially," Mr Rathbone says. "And those new organisms can be swallowed by other organisms in the same way, sort of like a matryoshka doll."
In fact, this is likely what happened when about one and a half billion years ago, a cyanobacterium was swallowed by another single-celled organism. The cyanobacteria already had a sophisticated antenna of proteins that trapped every photon of light. But instead of digesting the cyanobacterium, the host organism effectively stripped it for parts - retaining the antenna protein structure that the new organism - the red algae - used for energy.
And when another organism swallowed a red alga to become the first cryptophyte, it was a similar story. Except this time the antenna was brought to the other side of the membrane of the host organism and completely remoulded into new protein shapes that were equally as efficient at trapping sunlight photons.
EVOLUTION
As Prof. Curmi explains, these were the first tiny steps towards the evolution of modern plants and other photosynthetic organisms such as seaweeds.
"In going from cyanobacteria that are photosynthetic, to everything else on the planet that is photosynthetic, some ancient ancestor gobbled up a cyanobacteria which then became the cell's chloroplast that converts sunlight into chemical energy.
"And the deal between the organisms is sort of like, I'll keep you safe as long as you do photosynthesis and give me energy."
One of the collaborators on this project, Dr Beverley Green, Professor Emerita with the University of British Columbia's Department of Botany says Prof. Curmi was able to make the discovery by approaching the problem from a different angle.
"Paul's novel approach was to search for ancestral proteins on the basis of shape rather than similarity in amino acid sequence," she says.
"By searching the 3D structures of two red algal multi-protein complexes for segments of protein that folded in the same way as the cryptophyte protein, he was able to find the missing puzzle piece."
###
Shining, colored LED lighting on microalgae for next-generation biofuel
Combined blue and red illumination enhances growth, biosynthesis of D. Salina microalgae
IMAGE: RESEARCHERS SHOW HOW A COMBINATION OF MONOCHROMATIC RED AND BLUE LED ILLUMINATION ON ONE TYPE OF MICROALGA CAN ENHANCE ITS GROWTH AND INCREASE THE BIOSYNTHESIS OF CRITICAL COMPONENTS. HERE, LED... view more
CREDIT: XIAOJIAN ZHOU/YANGZHOU UNIVERSITY
WASHINGTON, March 30, 2021 -- As ethanol, biodiesel, and other biofuels continue to present challenges, such as competing with food security or lacking the technology for more efficient and low-cost production, microalgae are gaining momentum as a biofuel energy crop.
In their paper, published in the Journal of Renewable and Sustainable Energy, by AIP Publishing, Yangzhou University researchers in China show how a combination of monochromatic red and blue LED illumination on one type of microalga can enhance its growth and increase the biosynthesis of critical components, such as lipids, for microalgae feedstock development.
The researchers focused on Dunaliella salina (D. salina), typically extracted from sea salt fields and found in salt lakes. Easily cultivated and known for its bright pink color because of its high levels of carotenoid, D. salina is widely used in foods and cosmetics.
Microalgae tend to accumulate higher amounts of lipids (fatty acids that make up natural oils and waxes) than other biomass feedstocks do, which means a much higher percentage of the organisms can be turned into usable biofuel. In the case of D. salina, the lack of a cell wall could make algae biofuel production easier to pursue.
LEDs, as tunable single-color light sources, are already used to optimize plant growth, particularly in greenhouse cultivation. All parts of the visible spectrum are used in photosynthesis, but light also influences plant development. Adding more blue or red light, for instance, affects different plants in different ways. Optimal illumination conditions for microalgae growth and lipid production yield remain unknown.
In their study, the researchers applied red, blue, or combined red-blue illumination to D. salina culture. They found different intensities of blue light did not significantly enhance microalgae growth but boosted the lipid, protein, and carbohydrate levels. Red light, on the other hand, reduced algae growth and lipid formation, compared to blue and white light.
However, when red and blue lights were simultaneously applied in various ratios, the microalgae showed a major boost in growth and lipid productivity. The optimal 4-to-3 ratio of red and blue light significantly improved lipid productivity by more than 35% and increased dry biomass yield by more than 10% compared to the white light control.
The researchers are planning to analyze the composition of fatty acids synthesized in the algae under the favorable combined lighting for increased lipid production.
"Biodiesel performance is dependent on the composition of fatty acids, so we want to determine how the combined monochromatic lights would affect the quality of microalgae biodiesel," author Xiaojian Zhou said.
###
The article "Impact of combined monochromatic light on the bio-component productivity of Dunaliella salina" is authored by Cuili Jin, Binqi Yu, Shouyuan Qian, Qing Liu, and Xiaojian Zhou. The article will appear in Journal of Renewable and Sustainable Energy on March 30, 2021 (DOI: 10.1063/5.0041330). After that date, it can be accessed at https:/
ABOUT THE JOURNAL
Journal of Renewable and Sustainable Energy is an interdisciplinary journal that publishes across all areas of renewable and sustainable energy relevant to the physical science and engineering communities. Topics covered include solar, wind, biofuels and more, as well as renewable energy integration, energy meteorology and climatology, and renewable resourcing and forecasting. See https:/
Bottom-up' approach needed to study freshwater blooms
A national research team urges more complete study of harmful cyanobacteria
IMAGE: CYANOBACTERIA BLOOMS DAMAGE THE QUALITY OF LAKE WATER, AFFECT LAKE COMMUNITIES, AND CAN THREATEN HUMAN HEALTH. view more
CREDIT: MIDGE ELIASSEN
HANOVER, N.H. - March 30, 2021 - Cyanobacteria living at the bottom of lakes may hold important, under-researched clues about the threat posed by these harmful organisms, according to a Dartmouth-led study.
The research, published in the Journal of Plankton Research, urges a more comprehensive approach to cyanobacteria studies in order to manage the dangerous blooms during a time of global climate change.
"Most studies of cyanobacteria focus on the times when they are visible in the water column," said Kathryn Cottingham, the Dartmouth Professor in the Arts and Sciences, and a professor of biology. "By concentrating on this part of the life cycle, we may be missing important clues about how these harmful organisms are responding to ongoing global change."
Also known as blue-green algae, blooms of cyanobacteria are increasing in many freshwater systems worldwide, damaging the quality of lake water and affecting lake communities. The blooms also threaten human health through toxins that can damage organ systems.
Cyanobacteria are complex, but researchers are bringing together clues on how they respond to changing seasons.
Freshwater cyanobacteria live either suspended in the water column or at the lake bottom depending, in part, on water temperature. During warmer months, suspended "pelagic" cyanobacteria thrive in warm, well-lit surface waters. In the fall, they sink to the bottom and spend the winter in a resting or fully dormant state.
The Dartmouth-led study focuses on how cyanobacteria behave around the bottom-dwelling "overwinter" period. The research stresses that the sediment-dwelling stage returns to the water column with higher water temperatures, often following the mixing of the water column. Climate change is decreasing some types of mixing but increasing others - such as that caused by extreme precipitation events.
According to the paper, if mixing is constrained and cyanobacteria are left at the bottom, blooms could decrease. "A more complete understanding of all stages of the cyanobacterial life cycle will enable plankton researchers to better predict how ongoing climate change will affect the frequency, intensity and duration of cyanobacterial blooms," the study said.
Land-use changes--such as deforestation, fertilizer use, and development--and climate change are considered the main drivers of cyanobacteria outbreaks. Although the precise causes of the blooms are still being studied, researchers believe that they come from ongoing increases in nutrient loading, temperature and precipitation.
"Our work indicates that cyanobacterial blooms could either increase or decrease as a result of climate change, necessitating preventative lake management to limit human health risks," said Cayelan Carey, associate professor of biological sciences at Virginia Tech and co-author on the study. "Avoiding fertilizer use and installing waterfront buffers can help decrease cyanobacteria, thereby providing 'insurance' against potential cyanobacteria increases due to warmer temperatures in the future."
The study focused on temperate lakes, but the research team stresses that other waterbodies should be include in the proposed research agenda.
Kathleen Weathers from the Cary Institute, and Holly Ewing and Meredith Greer from Bates College also contributed to the study.
 |
| STARE AT THE WHITE SPOT CLOSE YOUR EYES OPEN THEM TO SEE TRUE CYAN |
New early warning system for self-driving cars
AI recognizes potentially critical traffic situations seven seconds in advance
A team of researchers at the Technical University of Munich (TUM) has developed a new early warning system for vehicles that uses artificial intelligence to learn from thousands of real traffic situations. A study of the system was carried out in cooperation with the BMW Group. The results show that, if used in today's self-driving vehicles, it can warn seven seconds in advance against potentially critical situations that the cars cannot handle alone - with over 85% accuracy.
To make self-driving cars safe in the future, development efforts often rely on sophisticated models aimed at giving cars the ability to analyze the behavior of all traffic participants. But what happens if the models are not yet capable of handling some complex or unforeseen situations?
A team working with Prof. Eckehard Steinbach, who holds the Chair of Media Technology and is a member of the Board of Directors of the Munich School of Robotics and Machine Intelligence (MSRM) at TUM, is taking a new approach. Thanks to artificial intelligence (AI), their system can learn from past situations where self-driving test vehicles were pushed to their limits in real-world road traffic. Those are situations where a human driver takes over - either because the car signals the need for intervention or because the driver decides to intervene for safety reasons.
Pattern recognition through RNN
The technology uses sensors and cameras to capture surrounding conditions and records status data for the vehicle such as the steering wheel angle, road conditions, weather, visibility and speed. The AI system, based on a recurrent neural network (RNN), learns to recognize patterns with the data. If the system spots a pattern in a new driving situation that the control system was unable to handle in the past, the driver will be warned in advance of a possible critical situation.
"To make vehicles more autonomous, many existing methods study what the cars now understand about traffic and then try to improve the models used by them. The big advantage of our technology: we completely ignore what the car thinks. Instead we limit ourselves to the data based on what actually happens and look for patterns," says Steinbach. "In this way, the AI discovers potentially critical situations that models may not be capable of recognizing, or have yet to discover. Our system therefore offers a safety function that knows when and where the cars have weaknesses."
Warnings up to seven seconds in advance
The team of researchers tested the technology with the BMW Group and its autonomous development vehicles on public roads and analyzed around 2500 situations where the driver had to intervene. The study showed that the AI is already capable of predicting potentially critical situations with better than 85 percent accuracy - up to seven seconds before they occur.
Collecting data with no extra effort
For the technology to function, large quantities of data are needed. After all, the AI can only recognize and predict experiences at the limits of the system if the situations were seen before. With the large number of development vehicles on the road, the data was practically generated by itself, says Christopher Kuhn, one of the authors of the study: "Every time a potentially critical situation comes up on a test drive, we end up with a new training example." The central storage of the data makes it possible for every vehicle to learn from all of the data recorded across the entire fleet.
Subscribe to:
Posts (Atom)