Connect with us

TOP SCEINCE

Animal brain inspired AI game changer for autonomous robots

Published

on

Animal brain inspired AI game changer for autonomous robots


A team of researchers at Delft University of Technology has developed a drone that flies autonomously using neuromorphic image processing and control based on the workings of animal brains. Animal brains use less data and energy compared to current deep neural networks running on GPUs (graphic chips). Neuromorphic processors are therefore very suitable for small drones because they don’t need heavy and large hardware and batteries. The results are extraordinary: during flight the drone’s deep neural network processes data up to 64 times faster and consumes three times less energy than when running on a GPU. Further developments of this technology may enable the leap for drones to become as small, agile, and smart as flying insects or birds. The findings were recently published in Science Robotics.

Learning from animal brains: spiking neural networks

Artificial intelligence holds great potential to provide autonomous robots with the intelligence needed for real-world applications. However, current AI relies on deep neural networks that require substantial computing power. The processors made for running deep neural networks (Graphics Processing Units, GPUs) consume a substantial amount of energy. Especially for small robots like flying drones this is a problem, since they can only carry very limited resources in terms of sensing and computing.

Animal brains process information in a way that is very different from the neural networks running on GPUs. Biological neurons process information asynchronously, and mostly communicate via electrical pulses called spikes. Since sending such spikes costs energy, the brain minimizes spiking, leading to sparse processing.

Inspired by these properties of animal brains, scientists and tech companies are developing new, neuromorphic processors. These new processors allow to run spiking neural networks and promise to be much faster and more energy efficient.

“The calculations performed by spiking neural networks are much simpler than those in standard deep neural networks.,” says Jesse Hagenaars, PhD candidate and one of the authors of the article, “Whereas digital spiking neurons only need to add integers, standard neurons have to multiply and add floating point numbers. This makes spiking neural networks quicker and more energy efficient. To understand why, think of how humans also find it much easier to calculate 5 + 8 than to calculate 6.25 x 3.45 + 4.05 x 3.45.”

This energy efficiency is further boosted if neuromorphic processors are used in combination with neuromorphic sensors, like neuromorphic cameras. Such cameras do not make images at a fixed time interval. Instead, each pixel only sends a signal when it becomes brighter or darker. The advantages of such cameras are that they can perceive motion much more quickly, are more energy efficient, and function well both in dark and bright environments. Moreover, the signals from neuromorphic cameras can feed directly into spiking neural networks running on neuromorphic processors. Together, they can form a huge enabler for autonomous robots, especially small, agile robots like flying drones.

First neuromorphic vision and control of a flying drone

In an article published in Science Robotics on May 15, 2024, researchers from Delft University of Technology, the Netherlands, demonstrate for the first time a drone that uses neuromorphic vision and control for autonomous flight. Specifically, they developed a spiking neural network that processes the signals from a neuromorphic camera and outputs control commands that determine the drone’s pose and thrust. They deployed this network on a neuromorphic processor, Intel’s Loihi neuromorphic research chip, on board of a drone. Thanks to the network, the drone can perceive and control its own motion in all directions.

“We faced many challenges,” says Federico Paredes-Vallés, one of the researchers that worked on the study, “but the hardest one was to imagine how we could train a spiking neural network so that training would be both sufficiently fast and the trained network would function well on the real robot. In the end, we designed a network consisting of two modules. The first module learns to visually perceive motion from the signals of a moving neuromorphic camera. It does so completely by itself, in a self-supervised way, based only on the data from the camera. This is similar to how also animals learn to perceive the world by themselves. The second module learns to map the estimated motion to control commands, in a simulator. This learning relied on an artificial evolution in simulation, in which networks that were better in controlling the drone had a higher chance of producing offspring. Over the generations of the artificial evolution, the spiking neural networks got increasingly good at control, and were finally able to fly in any direction at different speeds. We trained both modules and developed a way with which we could merge them together. We were happy to see that the merged network immediately worked well on the real robot.”

With its neuromorphic vision and control, the drone is able to fly at different speeds under varying light conditions, from dark to bright. It can even fly with flickering lights, which make the pixels in the neuromorphic camera send great numbers of signals to the network that are unrelated to motion.

Improved energy efficiency and speed by neuromorphic AI

“Importantly, our measurements confirm the potential of neuromorphic AI. The network runs on average between 274 and 1600 times per second. If we run the same network on a small, embedded GPU, it runs on average only 25 times per second, a difference of a factor ~10-64! Moreover, when running the network, , Intel’s Loihi neuromorphic research chip consumes 1.007 watts, of which 1 watt is the idle power that the processor spends just when turning on the chip. Running the network itself only costs 7 milliwatts. In comparison, when running the same network, the embedded GPU consumes 3 watts, of which 1 watt is idle power and 2 watts are spent for running the network. The neuromorphic approach results in AI that runs faster and more efficiently, allowing deployment on much smaller autonomous robots.,” says Stein Stroobants, PhD candidate in the field of neuromorphic drones.

Future applications of neuromorphic AI for tiny robots

“Neuromorphic AI will enable all autonomous robots to be more intelligent,” says Guido de Croon, Professor in bio-inspired drones, “but it is an absolute enabler for tiny autonomous robots. At Delft University of Technology’s Faculty of Aerospace Engineering, we work on tiny autonomous drones which can be used for applications ranging from monitoring crop in greenhouses to keeping track of stock in warehouses. The advantages of tiny drones are that they are very safe and can navigate in narrow environments like in between ranges of tomato plants. Moreover, they can be very cheap, so that they can be deployed in swarms. This is useful for more quickly covering an area, as we have shown in exploration and gas source localization settings.”

“The current work is a great step in this direction. However, the realization of these applications will depend on further scaling down the neuromorphic hardware and expanding the capabilities towards more complex tasks such as navigation.”



Source link

Continue Reading
Click to comment

Leave a Reply

TOP SCEINCE

Orchids support seedlings through ‘parental nurture’ via shared underground fungal networks

Published

on

By

Orchids support seedlings through ‘parental nurture’ via shared underground fungal networks


The Common Spotted Orchid (Dactylorhiza fuchsii) is found all over the UK. These orchids produce tiny seeds that can be carried anywhere by the wind, yet they often appear in clumps with small seedlings growing near mature plants. This phenomenon has puzzled ecologists since Darwin’s time, with the exact reason remaining a mystery.

A new study, led by researchers from the University of Sheffield in collaboration with The University of Manchester, provides the first evidence that early stage orchid seedlings germinate and thrive near to adult plants due to a kind of parental nurture using underground fungal networks.

Scientists investigated the idea that fungal networks, known as mycorrhizal networks, act as a direct pathway for established orchid plants to share recently produced sugars with developing seedlings.

Professor Katie Field, co-author of the study and Professor of Plant-Soil Processes at the University of Sheffield’s School of Biosciences, said: “Our results support the idea that some orchids engage in a form of ‘parental nurture’ with their seedlings. By supplying early stage seedlings with essential nutrients via shared fungal connections, the parent orchids give the seedlings an advantage over neighbouring plants that are competing for the same resources.

“This finding is exciting because why these orchids are often found in clumps, despite their seeds being wind dispersed, has been a puzzle for hundreds of years.”

The study focused on the Common Spotted Orchid and its fungal partner, Ceratobasidium cornigerum. Researchers created a system where mature, green orchids were connected to developing, chlorophyll-free seedlings through a fungal network grown on agar.

The mature plants were then exposed to a special form of carbon dioxide that could be tracked within the system.

Here’s how it worked:

  • Green orchid plants were connected to developing seedlings through the fungal network.
  • The green plants were then exposed to a special form of carbon dioxide that could be tracked within the system.
  • After a period of time, the researchers analysed both the seedlings and the fungal network to see where the labelled carbon ended up.

The results were clear, the seedlings were accumulating the labelled carbon, indicating they were being supported by the adult plants. By tracking the movement of carbon, the study showed that the mature orchids were indeed sharing their recently produced sugars with the seedlings through the fungal network.

Sir David Read, Emeritus Professor of Plant Sciences from the University of Sheffield and lead author of the study, said: “Whereas the seeds of most plants, for example legumes (peas, beans) and grasses (rice, corn, wheat) are fully provisioned with food reserves by their parent plants, the so-called dust seeds of orchids receive insufficient reserves from the parents to develop on their own.

“They are instead produced in their millions by each individual parent orchid plant from which they are dispersed, by wind, to the surrounding environment. Even Charles Darwin was puzzled by this strategy, suggesting that while it should enable the seeds of an individual orchid plant to be so widely distributed that within a few years it could colonise the whole world. He observed that their failure to do so ‘could not be understood at this time’.

“What is now revealed is that the belowground development of these essentially reserve-free seeds can be supported by photosynthetically produced sugars that are transported to them from mature plants growing above ground through a shared mycelium of symbiotic fungi.”

Results of the study, published in New Phytologist, show the amount of carbon transferred seemed to depend on the environment. When the fungus had access to a richer food source (oatmeal agar), less carbon was transferred to the seedling. This suggests that the strength of the demand from the seedlings may influence the flow of nutrients through the network.

This research has important implications for understanding orchid ecology and conservation efforts. By recognising the importance of fungal connections, scientists can develop better strategies for protecting these unique and often threatened plants.

Next steps are to research this theory in the natural habitat that the orchids are found in and to look at whether this applies to other species.



Source link

Continue Reading

TOP SCEINCE

Bird flu: Diverse range of vaccines platforms ‘crucial’ for enhancing human pandemic preparedness

Published

on

By

Bird flu: Diverse range of vaccines platforms ‘crucial’ for enhancing human pandemic preparedness


Vaccination remains the most effective strategy for avian influenza prevention and control in humans, despite varying vaccine efficacy across strains.

That’s according to the authors of a new review which delves into existing research into bird flu vaccines for humans.

Published in the peer-reviewed journal Human Vaccines & Immunotherapeutics, the results of the paper are particularly timely following news last week (Wednesday 22nd May) that the bird flu strain H5N1 had once again, for a second time, jumped from cattle in America to a human — prompting fears of subsequent human-to-human infection, with possible critical consequences.

Instances of the avian influenza were first recognized in US cattle in March. Since then, this strain has mainly spread from cow-to-cow and scientists have discovered very high levels of virus in raw milk (pasteurized milk is safe, having shown viral RNA but not infectious virus). To-date two people, however, are known to have contracted the bird flu virus. Both patients — US farmers — only reported eye symptoms and with treatment they made a full recovery.

Following tests on the first human instance, it was seen that the strain had mutated to be better adapted to mammalian cells, but as long as that human didn’t pass it onto another person it likely stopped the spread at that point. With the second case, the CDC has released a statement to say it has been monitoring influenza surveillance systems intently, especially in impacted states. “There has been no sign of unusual influenza activity in people, including in syndromic surveillance,” they report.

The concern now, though, is that if H5N1 continues to be given the environment in which to mutate (such as in close quarter cattle farms) — and this continues long enough — it has the potential to find a combination that will easily spread to humans.

The results of this new research, carried out by a team at the University of Georgia, USA, suggests vaccines still remain our “primary defense” against potential spread of avian influenzas such as the H5N1 and others assessed.

“The H5N1, H7N9, and H9N2 subtypes of avian influenza virus pose a dual threat, not only causing significant economic losses to the global poultry industry but also presenting a pressing public health concern due to documented spillover events and human cases,” explains lead author Flavio Cargnin Faccin, who alongside his mentor Dr. Daniel Perez of the University of Georgia, USA, analyzed the current landscape of research into human vaccines for these bird flus.

“This deep delve into the landscape of avian influenza vaccines for humans shows vaccination remains the primary defense against the spread of these viruses.”

The team examined studies of vaccines tested in mice, ferrets, non-human primates, and clinical trials of bird flu vaccines in humans, and assessed both established platforms and promising new directions.

The review carried out suggests inactivated vaccines are a safe and affordable option that primarily activate humoral immunity — the part of our immune system that produces antibodies.

Live attenuated influenza vaccines (LAIVs) are known to induce a wider immune response than inactivated vaccines, activating not only antibody production but also mucosal and cellular defenses. In this review, the authors suggest this broader response may offer greater protection, though, the authors suggest further research is needed to fully understand and harness its potential benefits for both human and agricultural applications.

The review also examined alternatives, such as virus-like particle (VLP) vaccines and messenger RNA (mRNA) vaccines, that have emerged more recently. Although VLP vaccines for bird flu have limited clinical trial data in humans, results from studies in mice and ferrets showed promise, the authors found. mRNA vaccines against H5N1 and H7N9 bird flu subtypes also generated a rapid and strong immune response in mice and ferrets, and, while data in humans is scarce, results from a phase 1 study of an H7N9 mRNA vaccine in healthy humans were “encouraging.”

Overall, the team suggests “exploring and employing a diverse range of vaccine platforms,” will be “crucial for enhancing pandemic preparedness and mitigating the threat of avian influenza viruses.”



Source link

Continue Reading

TOP SCEINCE

When should you neuter or spay your dog?

Published

on

By

When should you neuter or spay your dog?


Researchers at the University of California, Davis, have updated their guidelines on when to neuter 40 popular dog varieties by breed and sex. Their recent paper in Frontiers in Veterinary Science adds five breeds to a line of research that began in 2013 with a study that suggested that early neutering of golden retrievers puts them at increased risk of joint diseases and certain cancers.

That initial study set off a flurry of debate about the best age to neuter other popular breeds. Professors Lynette and Benjamin Hart of the School of Veterinary Medicine, the study’s lead authors, set out to add more breed studies by examining more than a decade of data from thousands of dogs treated at the UC Davis veterinary hospital. Their goal was to provide owners with more information to make the best decision for their animals.

They specifically looked at the correlation between neutering or spaying a dog before 1 year of age and a dog’s risk of developing certain cancers. These include cancers of the lymph nodes, bones, blood vessels or mast cell tumors for some breeds; and joint disorders such as hip or elbow dysplasia, or cranial cruciate ligament tears. Joint disorders and cancers are of particular interest because neutering removes male and female sex hormones that play key roles in important body processes such as closure of bone growth plates.

For the most recent study, they focused on German short/wirehaired pointer, mastiff, Newfoundland, Rhodesian ridgeback and Siberian husky. Data was collected from the UC Davis veterinary hospital’s records that included more than 200 cases for each of these five breeds weighing more than 20 kg (or 44 pounds), spanning January 2000 through December 2020.

The Harts said their updated guidelines emphasize the importance of personalized decisions regarding the neutering of dogs, considering the dog’s breed, sex and context. A table representing guidelines reflecting the research findings for all 40 breeds that have been studied, including the five new breeds, can be found here.

Health risks different among breeds

“It’s always complicated to consider an alternate paradigm,” said Professor Lynette Hart. “This is a shift from a long-standing model of early spay/neuter practices in the U.S. and much of Europe to neuter by 6 months of age, but important to consider as we see the connections between gonadal hormone withdrawal from early spay/neuter and potential health concerns.”

The study found major differences among these breeds for developing joint disorders and cancers when neutered early. Male and female pointer breeds had elevated joint disorders and increased cancers; male mastiff breeds had increased cranial cruciate ligament tears and lymphoma; female Newfoundland breeds had heightened risks for joint disorders; female Ridgeback breeds had heightened risks for mast cell tumors with very early neutering; and Siberian huskies showed no significant effects on joint disorders or cancers.

“We’re invested in making contributions to people’s relationship with their animals,” said Benjamin Hart, distinguished professor emeritus. “This guidance provides information and options for veterinarians to give pet owners, who should have the final decision-making role for the health and well-being of their animal.”

Their combined research studies will soon be available with others in the open access journal, Frontiers of Veterinary Science, as a free e-book, Effective Options Regarding Spay or Neuter of Dogs.

Other researchers on this UC Davis study include: Abigail Thigpen, Maya Lee, Miya Babchuk, Jenna Lee, Megan Ho, Sara Clarkson and Juliann Chou with the School of Veterinary Medicine; and Neil Willits with the Department of Statistics.

The research received a small amount of funding from the Center for Companion Animal Health, but was primarily conducted by the above authors as volunteers.



Source link

Continue Reading

Trending

Copyright © 2017 Zox News Theme. Theme by MVP Themes, powered by WordPress.