Tag: Science

  • Environmental Sensing Is Here, Tracking Everything from Forest Fires to Threatened Species

    Environmental Sensing Is Here, Tracking Everything from Forest Fires to Threatened Species

    [ad_1]

    You are in a lush forest. Sunlight filters through the bright green canopy, casting dappled shadows on the ground. Towering trees rise over delicate ferns, wildflowers, and colorful mushrooms. A deer slinks behind a shrub. But there are subtle signs of human intervention: small electronic devices gathering vital data on potential threats such as drought or pests and transmitting them miles away.

    Although technology has long been used to study animals and plants in forests, it’s evolving rapidly—becoming smaller, smarter, and more interconnected. Some devices are so small they can be placed on a single leaf. “For developers of such devices, the forest presents a completely new challenge,” says Ulrike Wallrabe, a professor of microsystem technology. Here, tech must withstand ever-changing conditions, from fluctuating temperatures to rain and snowfall, and even curious creatures, such as mice.

    Once up and running, the new generation of smart devices will offer unprecedented insights into the forest. “Drones already monitor large areas of forest, but they cannot explain why one tree is thriving and another is struggling,” says Wallrabe. “We need to understand what’s happening on a small scale and over time.”

    Wallrabe and her fellow researchers are working on a range of devices that will be deployed from the ground to the treetops, transmitting data from Germany’s Black Forest to their labs at the University of Freiburg. At the same time, scientists and engineers at other companies are also focused on making their devices unobtrusive and, ultimately, self-sustaining.

    Silvanet Wildfire Sensor

    Time is of the essence when fighting forest fires. Sensors attached to trunks “smell” tell-tale gases like hydrogen and carbon monoxide, and alert firefighters within the first hour—before satellites or cameras can spot open flames. German startup Dryad Networks has built AI into its solar-powered sensors to ensure that they can distinguish between real fires and, say, passing diesel trucks.

    Treevia

    Digital dendrometers relieve foresters of tedious work. As trees grow, the elastic band wrapped around their trunk stretches and transmits data directly to a computer. The lightweight device from Brazilian startup Treevia can even be attached to saplings. It also contains a humidity and heat sensor, providing insights into climatic impacts on reforested areas.

    The Guardian

    What does it take to catch illegal loggers or poachers? A smartphone is a good start. Rainforest Connection’s recycled, solar-powered smartphone listens for the sound of chain saws or gunshots within a 1-mile radius. The recordings are transmitted to the cloud for analysis and alert local authorities in near real time. This device also provides insights into the distribution and calling behavior of animals.

    BiodivX Drone

    As animals move through trees, they shed DNA through feces, skin, and hair. This innovative drone collects what is known as environmental DNA (eDNA) from leaves and branches—with particles sticking to its adhesive strips. Scientists from Switzerland programmed the drone so it can navigate autonomously through dense forests and hover steadily around branches to take samples.

    Leaf Sensor

    Wallrabe and her team at the University of Freiburg have developed a glass capsule that measures gas exchange between a leaf and its surroundings. It can detect specific chemicals that trees emit under stress, for example, in the event of a drought, infestation, or disease. The capsule is transparent so that sunlight can reach the leaf without impairing its function.

    Plant-e

    When sunlight is limited, most devices are powered by batteries. Plant-e, a Dutch company spun out of Wageningen University, makes use of a natural process: Plants produce organic material through photosynthesis; some they use for growth, the rest ends up into the soil. Bacteria break down this material and release electrons that Plant-e uses to power its sensors.

    Seed-dropping drones

    To scale up and accelerate tree planting efforts, several companies, including UK-based Dendra Systems, have developed cutting-edge drones. These drones, loaded with an array of seeds, hover over target areas and release their precious cargo. This is particularly helpful in remote areas that are difficult for humans to access. By recording the exact location of drop sites, foresters can monitor growth and health of the newly planted trees.

    This article appears in the January/February 2025 issue of WIRED UK magazine.

    [ad_2]

    Source link

  • Why Soccer Players Are Training in the Dark

    Why Soccer Players Are Training in the Dark

    [ad_1]

    “What’s interesting is that they haven’t just changed the light levels, they’ve changed the specific quality and waves of that light,” says Niall Macfarlane, a professor in physiology and sports science at the University of Glasgow who wasn’t involved in the research. Experimenting with different types of low light—blue, red, green—has far more scope for generating improvements than just using dimmed light on its own, he believes.

    Macfarlane does, however, harbor some reservations about how smoothly the research’s positive results would be translated into improvements in performance in a real-life soccer match. “I think it is far better suited to something like baseball or cricket where you can replicate the game scenario more closely,” he says. Unlike in the highly variable environment of a soccer game, with batting sports “you know where the ball is going to come from, you know roughly the speed it’s going to come at, so [through the training] you can refine the tracking of that ball and the consistency of the contact.”

    It is little surprise then that when I arrived earlier this afternoon, O’Connor was in the process of booking flights to Arizona. He is currently in contact with more than half of the teams in Major League Baseball.

    But there’s still more work to do. Further research is needed to see how long the benefits of the light training last, as well as to confirm that there was no performance bias at work in the recent study—it was obvious to the participants which group was receiving the Okkulo intervention and which was not, which may have subconsciously influenced the participants’ performance in testing. Future tests could account for this by involving some lighting adjustment for the control group as well, to avoid indicating to the participants which group they are in.

    Outside of elite sports, Okkulo is in the early stages of collaborating with several universities to delve deeper into the potential medical benefits of the technology, leading O’Connor to imagine a future in which hospitals and clinics have ambient light rooms to help patients to recover from severe brain injuries or paralysis. He hopes one day to move Okkulo into the home use and commercial gym sectors, following in the footsteps of companies such as Peloton.

    But for now, it is my turn. Redemption awaits. The room returns to normal light levels. I stand on my mark and wait to receive the ball. My systems are finely tuned, my heart rate is slightly raised, and my senses are on high alert. It’s time for me to flourish.

    Clunk—pause—whoosh!

    I control the ball. It’s perfect. Deft and soundless, falling under my spell like a lost duckling returning gratefully to the embrace of its nest. It belongs here. I experience a few seconds of immense pride before O’Connor reveals that we’ve been working with a ball-speed setting recommended for 10-year-olds. The technology may be exemplary, but even science cannot perform miracles.

    [ad_2]

    Source link

  • AI learns language of atom arrangements in solids

    AI learns language of atom arrangements in solids

    [ad_1]

    crystal
    Credit: CC0 Public Domain

    A new artificial intelligence model that can predict how atoms arrange themselves in crystal structures could lead to faster discovery of new materials for everything from solar panels to computer chips.

    The technology, called CrystaLLM, was developed by researchers at the University of Reading and University College London. It works similarly to AI chatbots, by learning the “language” of crystals by studying millions of existing crystal structures.

    Published 6 December in Nature Communications, the new system will be distributed to the scientific community to aid the discovery of new materials.

    Dr. Luis Antunes, who led the research while completing his Ph.D. at the University of Reading, said, “Predicting crystal structures is like solving a complex, multidimensional puzzle where the pieces are hidden. Crystal structure prediction requires massive computing power to test countless possible arrangements of atoms.

    “CrystaLLM offers a breakthrough by studying millions of known crystal structures to understand patterns and predict new ones, much like an expert puzzle solver who recognizes winning patterns rather than trying every possible move.”

    Predicting structures for unfamiliar materials

    The current process for figuring out how atoms will arrange themselves into crystals relies on time-consuming computer simulations of the physical interactions between the atoms. CrystaLLM works in a simpler way. Instead of using complex physics calculations, it learns by reading millions of crystal structure descriptions contained in Crystallographic Information Files—the standard format for representing crystal structures.

    CrystaLLM treats these crystal descriptions just like text. As it reads each description, it predicts what comes next, gradually learning patterns about how crystals are structured. The system was never taught any physics or chemistry rules, but instead figured them out on its own. It learned things like how atoms arrange themselves and how their size affects the crystal’s shape, just from reading these descriptions.

    When tested, CrystaLLM could successfully generate realistic crystal structures, even for materials it had never seen before.

    The research team has created a free website where researchers can use CrystaLLM to generate crystal structures. The integration of this model within crystal structure prediction workflows could speed up the development of new materials for technologies like better batteries, more efficient solar cells, and faster computer chips.

    More information:
    Crystal structure generation with autoregressive large language modelling, Nature Communications (2024). DOI: 10.1038/s41467-024-54639-7

    Provided by
    University of Reading


    Citation:
    Speaking crystal: AI learns language of atom arrangements in solids (2024, December 6)
    retrieved 6 December 2024
    from https://phys.org/news/2024-12-crystal-ai-language-atom-solids.html

    This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
    part may be reproduced without the written permission. The content is provided for information purposes only.



    [ad_2]

    Source link

  • Biodegradable polymers show promise for nutrient encapsulation and replacing microbeads in beauty products

    Biodegradable polymers show promise for nutrient encapsulation and replacing microbeads in beauty products

    [ad_1]

    beauty products
    Credit: Pixabay/CC0 Public Domain

    Microplastics are an environmental hazard found nearly everywhere on Earth, released by the breakdown of tires, clothing, and plastic packaging. Another significant source of microplastics is tiny beads that are added to some cleansers, cosmetics, and other beauty products.

    In an effort to cut off some of these microplastics at their source, MIT researchers have developed a class of biodegradable materials that could replace the plastic beads now used in beauty products. These polymers break down into harmless sugars and amino acids.

    “One way to mitigate the microplastics problem is to figure out how to clean up existing pollution. But it’s equally important to look ahead and focus on creating materials that won’t generate microplastics in the first place,” says Ana Jaklenec, a principal investigator at MIT’s Koch Institute for Integrative Cancer Research.

    These particles could also find other applications. In the new study, Jaklenec and her colleagues showed that the particles could be used to encapsulate nutrients such as vitamin A. Fortifying foods with encapsulated vitamin A and other nutrients could help some of the 2 billion people around the world who suffer from nutrient deficiencies.

    Jaklenec and Robert Langer, an MIT Institute Professor and member of the Koch Institute, are the senior authors of the paper, which appears in Nature Chemical Engineering. The paper’s lead author is Linzixuan (Rhoda) Zhang, an MIT graduate student in chemical engineering.

    Biodegradable plastics

    In 2019, Jaklenec, Langer, and others reported a polymer material that they showed could be used to encapsulate vitamin A and other essential nutrients. They also found that people who consumed bread made from flour fortified with encapsulated iron showed increased iron levels.

    However, since then, the European Union has classified this polymer, known as BMC, as a microplastic and included it in a ban that went into effect in 2023. As a result, the Bill and Melinda Gates Foundation asked the MIT team if they could design an alternative that would be more environmentally friendly.

    The researchers, led by Zhang, turned to a type of polymer that Langer’s lab had previously developed, known as poly(beta-amino esters). These polymers, which have shown promise as vehicles for gene delivery and other medical applications, are biodegradable and break down into sugars and amino acids.

    By changing the composition of the material’s building blocks, researchers can tune properties such as hydrophobicity (ability to repel water), mechanical strength, and pH sensitivity. After creating five different candidate materials, the MIT team tested them and identified one that appeared to have the optimal composition for microplastic applications, including the ability to dissolve when exposed to acidic environments such as the stomach.

    The researchers showed that they could use these particles to encapsulate vitamin A, as well as vitamin D, vitamin E, vitamin C, zinc and iron. Many of these nutrients are susceptible to heat and light degradation, but when encased in the particles, the researchers found that the nutrients could withstand exposure to boiling water for two hours.

    They also showed that even after being stored for six months at high temperature and high humidity, more than half of the encapsulated vitamins were undamaged.

    To demonstrate their potential for fortifying food, the researchers incorporated the particles into bouillon cubes, which are commonly consumed in many African countries. They found that when incorporated into bouillon, the nutrients remained intact after being boiled for two hours.

    “Bouillon is a staple ingredient in sub-Saharan Africa, and offers a significant opportunity to improve the nutritional status of many millions of people in those regions,” Jaklenec says.

    In this study, the researchers also tested the particles’ safety by exposing them to cultured human intestinal cells and measuring their effects on the cells. At the doses that would be used for food fortification, they found no damage to the cells.

    Discover the latest in science, tech, and space with over 100,000 subscribers who rely on Phys.org for daily insights.
    Sign up for our free newsletter and get updates on breakthroughs,
    innovations, and research that matter—daily or weekly.

    Better cleansing

    To explore the particles’ ability to replace the microbeads that are often added to cleansers, the researchers mixed the particles with soap foam. This mixture, they found, could remove permanent marker and waterproof eyeliner from skin much more effectively than soap alone.

    Soap mixed with the new microplastic was also more effective than a cleanser that includes polyethylene microbeads, the researchers found. They also discovered that the new biodegradable particles did a better job of absorbing potentially toxic elements such as heavy metals.

    “We wanted to use this as a first step to demonstrate how it’s possible to develop a new class of materials, to expand from existing material categories, and then to apply it to different applications,” Zhang says.

    The researchers are working on further testing the microbeads as a cleanser and potentially other applications, and they plan to run a small human trial later this year. They are also gathering safety data that could be used to apply for GRAS (generally regarded as safe) classification from the U.S. Food and Drug Administration and are planning a clinical trial of foods fortified with the particles.

    The researchers hope their work could help to significantly reduce the amount of microplastic released into the environment from health and beauty products.

    “This is just one small part of the broader microplastics issue, but as a society we’re beginning to acknowledge the seriousness of the problem. This work offers a step forward in addressing it,” Jaklenec says.

    “Polymers are incredibly useful and essential in countless applications in our daily lives, but they come with downsides. This is an example of how we can reduce some of those negative aspects.”

    More information:
    Degradable poly(β-amino ester) microparticles for cleansing products and food fortification, Nature Chemical Engineering (2024). DOI: 10.1038/s44286-024-00151-0

    Provided by
    Massachusetts Institute of Technology


    This story is republished courtesy of MIT News (web.mit.edu/newsoffice/), a popular site that covers news about MIT research, innovation and teaching.

    Citation:
    Biodegradable polymers show promise for nutrient encapsulation and replacing microbeads in beauty products (2024, December 6)
    retrieved 6 December 2024
    from https://phys.org/news/2024-12-biodegradable-polymers-nutrient-encapsulation-microbeads.html

    This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no
    part may be reproduced without the written permission. The content is provided for information purposes only.



    [ad_2]

    Source link

  • Nobel prizewinner who transformed cancer therapies with research into brain hormones

    Nobel prizewinner who transformed cancer therapies with research into brain hormones

    [ad_1]

    Andrew Schally on the phone at the Veterans Administration Hospital in New Orleans USA shared the Nobel Prize for Medicine.

    Credit: Keystone/Zuma Press/Alamy

    Andrew V. Schally’s most enduring legacy lies in his groundbreaking discovery of brain hormones that regulate the pituitary gland. Located below the hypothalamus, the gland produces, stores and releases several hormones. It also controls the function of other glands. Schally’s identification of hypothalamic hormones upended our understanding of the endocrine system and the treatment of hormone-dependent conditions. His work on these peptides — small proteins that can serve as messengers in the body — earned him the Nobel Prize in Physiology or Medicine in 1977. He shared the honour with Rosalyn Yalow and Roger Guillemin for their independent contributions to the field. Schally has died, aged 97.

    Born in 1926 in Wilno, Poland, (now Vilnius in Lithuania), Schally’s early years were marked by hardship, surviving Nazi-occupied eastern Europe and spending part of the Second World War in Romania. In 1945, he journeyed through Italy and France to settle in the United Kingdom. After receiving a secondary-school diploma in Scotland, he went to London, where he studied chemistry. His endocrinology journey began in 1949, when he joined the National Institute for Medical Research in London as a research assistant. There, he worked with Charles Harington, the director of the institute and a pioneering biochemist who had analysed the chemical constitution of the thyroid hormone thyroxine.

    In 1952, Schally’s passion for mammalian physiology led him to McGill University in Montreal, Canada, where he obtained a bachelor’s degree in 1955 and a PhD in 1957. He joined the ranks of those studying the long- and short-range effects of hormones, as well as the factors that affected their release, laying the foundation for his lifelong dedication to understanding the intricate interface between brain function and endocrine activity.

    In 1957, Schally joined Baylor College of Medicine in Houston, Texas. There, he collaborated with Roger Guillemin on research on the hypothalamus, and particularly its role in self-regulating processes (homeostasis). Although their relationship later became contentious, the affiliation drove both investigators to achieve remarkable scientific breakthroughs. Schally was determined to identify the structure of thyrotropin-releasing factor (TRF), which is secreted by the hypothalamus to regulate the release of thyrotropin from the anterior pituitary gland. Thyrotropin stimulates the thyroid gland to produce thyroid hormones, which regulate nearly every aspect of the body’s metabolic activity.

    Isolating the delicate TRF molecule in the brain while maintaining its active conformation posed huge challenges, casting doubt on the initial findings. But Schally stood firm in his observations on TRF and his hypotheses about how other hypothalamic hormones can regulate the anterior pituitary’s function. A pivotal moment came in 1961, during Schally’s visit to Uppsala University in Sweden to see biochemist Jerker Porath. There, he gained valuable experience in the use of the filtration gel Sephadex and column electrophoresis, which would prove crucial in isolating the TRF.

    A year later, Schally established a research group at Tulane University in New Orleans, Louisiana, and the New Orleans VA Medical Center. With unwavering faith and patience, he meticulously followed the demanding steps of the isolation process, overcoming the immense challenge of extracting increasingly pure materials from a crude hypothalamic extract. He yielded 800 micrograms of peptide from 160,000 pig hypothalami and provided conclusive evidence that the brain controls hormonal secretions in the body. In 1969, Schally and Guillemin independently isolated TRF and identified its structure. Their studies provided experimental confirmation of the relationship predicted by the UK physiologist Geoffrey Harris in the 1940s.

    In 1971, Schally doubled his efforts and successfully identified the luteinizing hormone-releasing hormone (LH-RH), the peptide that regulates the release of luteinizing hormone from the anterior pituitary gland, triggering ovulation and stimulating the secretion of progesterone and oestrogen from the ovaries.

    His development and innovative use of LH-RH agonists and antagonists — agents that mimic or block LH-RH — provided less-invasive and more-effective alternatives to conventional treatments for cancers that require hormones to grow, such as prostate and breast tumours.

    In 2005, Schally relocated his laboratory from New Orleans to the Miami VA Medical Center and the University of Miami in Florida. His research turned to growth hormone-releasing hormone (GH-RH) agonists and antagonists, which had shown potential in treating various cancers and cardiovascular and inflammatory diseases.

    Schally left a lasting impact on medical science and those around him. Lab staff and colleagues enjoyed his famous ‘Schally’s special gin and tonic’ at his residence. He was an avid swimmer in the Atlantic Ocean even in his later years, notwithstanding various lifeguards’ objections to his swimming far out from the shore at Miami Beach.

    After the COVID-19 pandemic, Schally continued to work from his home in Florida, reviewing and writing manuscripts and savouring conversations about lab results. Often, discussions would bring up old manuscripts and references, and he would always know exactly where each paper was in his cabinet. This mastery of both the scientific and historical progress of the field of hypothalamic hormones and peptides made him an effective research director and mentor.

    Schally was dedicated to advancing health care, particularly for veterans. He devoted 62 years to conducting research at the US Department of Veterans Affairs until his final years. His dedication continues to inspire investigators who build on his pioneering work.

    Competing Interests

    The authors declare no competing interests.

    [ad_2]

    Source link

  • Wuhan lab samples hold no close relatives to virus behind COVID

    Wuhan lab samples hold no close relatives to virus behind COVID

    [ad_1]

    Shi Zhengli wearing protective clothing in a biosafety lab.

    Chinese virologist Shi Zhengli has presented evidence that her lab has not worked with close relatives of SARS-CoV-2.Credit: Johannes Eisele/AFP via Getty

    After years of rumours that the virus that causes COVID-19 escaped from a laboratory in China, the virologist at the centre of the claims has presented data on dozens of new coronaviruses collected from bats in southern China. At a conference in Japan this week, Shi Zhengli, a specialist on bat coronaviruses, reported that none of those viruses stored in her freezers are the most recent ancestors of the virus SARS-CoV-2.

    Shi was leading coronavirus research at the Wuhan Institute of Virology (WIV), a high-level biosafety laboratory, when the first cases of COVID-19 were reported in that city. Soon afterwards, theories emerged that the virus had leaked — either by accident or deliberately — from the WIV.

    Shi has consistently said that SARS-CoV-2 was never seen or studied in her lab. But some commentators have continued to ask whether one of the many bat coronaviruses her team collected in southern China over decades was closely related to it. Shi promised to sequence the genomes of the coronaviruses and release the data.

    The latest analysis, which has not been peer reviewed, includes data from the whole genomes of 56 new betacoronaviruses, the broad group to which SARS-CoV-2 belongs, as well as some partial sequences. All the viruses were collected between 2004 and 2021.

    “We didn’t find any new sequences which are more closely related to SARS-CoV-1 and SARS-CoV-2,” said Shi, in a pre-recorded presentation at the conference, Preparing for the Next Pandemic: Evolution, Pathogenesis and Virology of Coronaviruses, in Awaji, Japan, on 4 December. Earlier this year, Shi moved from the WIV to the Guangzhou Laboratory, a newly established national research institute for infectious diseases.

    The results support her assertion that the WIV lab did not have any bat-derived sequences from viruses that were more closely related to SARS-CoV-2 than were any already described in scientific papers, says Jonathan Pekar, an evolutionary biologist at the University of Edinburgh, UK. “This just validates what she was saying: that she did not have anything extremely closely related, as we’ve seen in the years since,” he says.

    The closest known viruses to SARS-CoV-2 were found in bats in Laos and Yunnan, southern China — but years, if not decades, have passed since they shared their closesta common ancestor with the virus that causes COVID-19. “She’s basically found a lot of what we expect,” says Leo Poon, a virologist at the University of Hong Kong.

    Longtime collaboration

    For decades, Shi collaborated with Peter Daszak, president of the EcoHealth Alliance, a New York City-based non-profit organization, to survey bats in southern China for coronaviruses and study their risk to humans. The work was funded by the US National Institutes of Health and the US Agency for International Development, but in May this year, the government suspended federal funding to EcoHealth because it had not provided adequate oversight of research activities at the WIV. Those activities included modifying a coronavirus linked to severe acute respiratory syndrome (SARS), to study the potential origins of this type of virus in bats.

    [ad_2]

    Source link

  • Ancient stacks of dishes tell tale of society’s dissolution

    Ancient stacks of dishes tell tale of society’s dissolution

    [ad_1]

    Nature, Published online: 06 December 2024; doi:10.1038/d41586-024-03977-z

    Artefacts from a Mesopotamian archaeological site suggest that people in the region founded and later rejected an early form of the organized state.

    [ad_2]

    Source link

  • More-powerful AI is coming. Academia and industry must oversee it — together

    More-powerful AI is coming. Academia and industry must oversee it — together

    [ad_1]

    Close-up portrait of Sam Altman with a screen reading OpenAI behind him

    Sam Altman of OpenAI, the company that developed ChatGPT, says that machines will create superintelligence. Research is needed to verify these and other claims.Credit: Jason Redmond/AFP/Getty

    “It is possible that we will have superintelligence in a few thousand days (!); it may take longer, but I’m confident we’ll get there.”

    So wrote Sam Altman, the chief executive of OpenAI, a technology firm based in San Francisco, California, on 23 September. This was less than two weeks after the company behind ChatGPT released o1, its most advanced large language model (LLM) yet. Once confined to the realms of science fiction, the rise of LLMs in the past few years has renewed the relevance of the question of when we might create artificial general intelligence (AGI). Although it lacks a precise definition, AGI broadly refers to an AI system capable of human-level reasoning, generalization, planning and autonomy.

    Policymakers everywhere have questions about AGI, including what its benefits and risks will be. These are not easy questions to answer, especially given that much of the work is happening in the private sector, in which studies are not always published openly. But what is clear is that AI companies are laser-focused on giving their systems the whole range of cognitive abilities enjoyed by humans. Companies developing AI models have a strong incentive to maintain the idea that AGI is nigh, to attract interest and therefore investment.

    There was a consensus among researchers who spoke to Nature for a News Feature published this week (see Nature 636, 22–25; 2024) that large language models (LLMs), such as o1, Google’s Gemini and Claude, made by Anthropic, based in San Francisco, have not yet delivered AGI. And, informed by insights from neuroscience, many say that there are good reasons to think that LLMs never will, and that another technology will be needed for AI to achieve human-level intelligence.

    Despite their breadth of ability — from generating computer code to summarizing academic articles and answering mathematics questions — there are fundamental limitations in the way the most powerful LLMs work, which involves essentially devouring a mass of data and using that to predict the next ‘token’ in a series. That generates plausible answers to a problem, rather than actually solving it.

    François Chollet, a software engineer formerly at Google, based in Mountain View, California, and Subbarao Kambhampati, a computer scientist at Arizona State University in Tempe, have tested o1’s performance on tasks that require abstract reasoning and planning, and found that it comes up short of AGI. If AGI is to happen, some researchers think that AI systems would need coherent ‘world models’, or representations of their surroundings that they can use to test hypotheses, reason, plan and generalize knowledge learnt in one domain to potentially limitless other situations.

    This is where ideas from neuroscience and cognitive science could propel the next breakthroughs. Yoshua Bengio’s team at the University of Montreal, Canada, for example, is exploring alternative AI architectures that would better support the building of coherent world models and the ability to reason using such models.

    Some researchers argue that the next breakthroughs in AI might come not from the biggest systems, but from smaller, more-energy-efficient AI. Smarter systems in the future could also require less data to train if they had the ability to decide which aspects of their environment to sample, rather than simply ingesting everything they are fed, says Karl Friston, a theoretical neuroscientist at University College London.

    Such work demonstrates that researchers from an array of fields need to be involved in AI development. This will be necessary to verify what systems are truly capable of, ensure that they live up to the technology companies’ claims and identify the breakthroughs needed for development. However, right now, accessing the leading AI systems can be difficult for researchers who don’t work at the companies that can afford the vast quantity of graphical processing units (GPUs) needed to train the systems (A. Khandelwal et al. Preprint at arXiv https://doi.org/nt67; 2024).

    To give a sense of the scale of activity, in 2021, US government agencies (excluding the Department of Defense) allocated US$1.5 billion to AI research and development, and the European Commission spends around €1 billion (US$1.05 billion) annually. By contrast, companies worldwide spent more than $340 billion on AI research in 2021 (N. Ahmed et al. Science 379, 884–886; 2023). There are ways that governments could fund AI research on a bigger scale, for example by pooling resources. The Confederation of Laboratories for Artificial Intelligence Research in Europe, a non-profit organization based in The Hague, the Netherlands, has suggested building a ‘CERN for AI’ that can attract the same level of talent as AI companies do and so create a cutting-edge research environment.

    It is hard to predict when AGI might arrive — estimates range from a few years from now to a decade or longer. But more huge advances in AI will certainly happen, and many of them will probably come from industry, considering the scale of investment. To ensure that these advances are beneficial, the research from technology companies needs to be verified using the best current understanding of what constitutes human intelligence, according to neuroscience, cognitive science, social science and other relevant fields. This publicly funded research needs to have a key role in AGI’s development.

    Humanity needs to bring all knowledge to bear so that applications of AI research are robust and its risks are mitigated as much as possible. Governments, companies, research funders and researchers need to recognize their complementary strengths. If they do not, then insights that could help to improve AI will be missed — and the resulting systems risk being unpredictable and therefore unsafe.

    [ad_2]

    Source link

  • Loose ends

    Loose ends

    [ad_1]

    Nature, Published online: 06 December 2024; doi:10.1038/d41586-024-04022-9

    Course correction.

    [ad_2]

    Source link

  • A science mega-programme is taking shape in the EU: what it means for researchers

    A science mega-programme is taking shape in the EU: what it means for researchers

    [ad_1]

    Nature, Published online: 06 December 2024; doi:10.1038/d41586-024-04017-6

    A new research commissioner will help to build the European Union’s next massive science scheme — with a renewed focus on innovation.

    [ad_2]

    Source link