Artificial intelligence uses less energy by mimicking the human brain

 Artificial Intelligence (AI) can perform complex calculations and analyze data faster than any human, but to do so requires enormous amounts of energy. The human brain is also an incredibly powerful computer, yet it consumes very little energy.

As technology companies increasingly expand, a new approach to AI's "thinking," developed by researchers including Texas A&M University engineers, mimics the human brain and has the potential to revolutionize the AI industry.

Dr. Suin Yi, assistant professor of electrical and computer engineering at Texas A&M's College of Engineering, is on a team of researchers that developed "Super-Turing AI," which operates more like the human brain. This new AI integrates certain processes instead of separating them and then migrating huge amounts of data like current systems do.

The Energy Crisis In AI

Today's AI systems, including large language models such as OpenAI and ChatGPT, require immense computing power and are housed in expansive data centers that consume vast amounts of electricity.

"These data centers are consuming power in gigawatts, whereas our brain consumes 20 watts," Suin explained. "That's 1 billion watts compared to just 20. Data centers that are consuming this energy are not sustainable with current computing methods. So while AI's abilities are remarkable, the hardware and power generation needed to sustain it is still needed."

The substantial energy demands not only escalate operational costs but also raise environmental concerns, given the carbon footprint associated with large-scale data centers. As AI becomes more integrated, addressing its sustainability becomes increasingly critical.

Emulating The Brain

Yi and team believe the key to solving this problem lies in nature -- specifically, the human brain's neural processes.

In the brain, the functions of learning and memory are not separated, they are integrated. Learning and memory rely on connections between neurons, called "synapses," where signals are transmitted. Learning strengthens or weakens synaptic connections through a process called "synaptic plasticity," forming new circuits and altering existing ones to store and retrieve information.

By contrast, in current computing systems, training (how the AI is taught) and memory (data storage) happen in two separate places within the computer hardware. Super-Turing AI is revolutionary because it bridges this efficiency gap, so the computer doesn't have to migrate enormous amounts of data from one part of its hardware to another.

"Traditional AI models rely heavily on backpropagation -- a method used to adjust neural networks during training," Yi said. "While effective, backpropagation is not biologically plausible and is computationally intensive.

"What we did in that paper is troubleshoot the biological implausibility present in prevailing machine learning algorithms," he said. "Our team explores mechanisms like Hebbian learning and spike-timing-dependent plasticity -- processes that help neurons strengthen connections in a way that mimics how real brains learn."

Hebbian learning principles are often summarized as "cells that fire together, wire together." This approach aligns more closely with how neurons in the brain strengthen their connections based on activity patterns. By integrating such biologically inspired mechanisms, the team aims to develop AI systems that require less computational power without compromising performance.

In a test, a circuit using these components helped a drone navigate a complex environment -- without prior training -- learning and adapting on the fly. This approach was faster, more efficient and used less energy than traditional AI.

Why This Matters For The Future Of AI

This research could be a game-changer for the AI industry. Companies are racing to build larger and more powerful AI models, but their ability to scale is limited by hardware and energy constraints. In some cases, new AI applications require building entire new data centers, further increasing environmental and economic costs.

Yi emphasizes that innovation in hardware is just as crucial as advancements in AI systems themselves. "Many people say AI is just a software thing, but without computing hardware, AI cannot exist," he said.

Looking Ahead: Sustainable AI Development

Super-Turing AI represents a pivotal step toward sustainable AI development. By reimagining AI architectures to mirror the efficiency of the human brain, the industry can address both economic and environmental challenges.

Yi and his team hope that their research will lead to a new generation of AI that is both smarter and more efficient.

"Modern AI like ChatGPT is awesome, but it's too expensive. We're going to make sustainable AI," Yi said. "Super-Turing AI could reshape how AI is built and used, ensuring that as it continues to advance, it does so in a way that benefits both people and the planet."

sources-science daily

Making sturdy, semi-transparent wood with cheap, natural materials

 Can you imagine a smartphone with a wooden touchscreen? Or a house with wooden windows? Probably not -- unless you've heard of transparent wood. Made by modifying wood's natural structure, this material has been proposed as a sturdy, eco-friendly plastic alternative. But wood's biodegradability is often sacrificed in the process. Researchers are hoping to change that by creating transparent woods from almost entirely natural materials and making them electrically conductive.

The researchers will present their results at the spring meeting of the American Chemical Society (ACS).

"In the modern day, plastic is everywhere, including our devices that we carry around. And it's a problem when we reach the end of that device's life. It's not biodegradable," explains Bharat Baruah, a professor of chemistry at Kennesaw State University and the presenter of this research. "So, I asked, what if we can create something that's natural and biodegradable instead?"

Baruah became interested in transparent woods thanks to his outside-of-work pursuits -- namely, his woodworking hobby. But he realized that the transparent woods reported by other scientists used materials such as epoxies, a form of plastic, for strength. To find natural materials that would keep wood sturdy and stable over time, he again turned to his personal experiences.

Growing up in the state of Assam in northeastern India, Baruah encountered buildings that had been standing for centuries -- long before the modern-day version of cement was invented. Instead, ancient masons created cement by mixing sand with sticky rice and egg whites. Baruah hypothesized that those same materials might be perfect for incorporating strength and stability into his transparent woods.

Wood has three components: cellulose, hemicellulose and lignin. To make it transparent, the lignin and hemicellulose are removed, leaving behind a porous, paper-like network of cellulose. Then, a colorless material fills in those pores, also restoring some rigidity.

Joined by Ridham Raval, an undergraduate student at the university, Baruah transformed pieces of balsa wood into natural, semi-transparent woods by pulling out the lignin and hemicellulose using a vacuum chamber and chemicals, including sodium sulfite (a delignifying agent), sodium hydroxide (a version of lye) and diluted bleach. Then, the pores were refilled by soaking them in an egg white and rice extract mixture, along with a curing agent called diethylenetriamine to keep the material see-through. The researchers say that these reagents, when used in small amounts, such as in this experiment, pose little threat to the environment.

In the end, the team was left with semi-transparent slices of wood that were durable and flexible.

The team next investigated some potential applications for their engineered woods, including as a replacement for glass windows. Again, Baruah tapped into his woodworker skills and renovated a birdhouse into a tiny, one-windowed, insulated home. To test the modernized abode's energy efficiency, he put the birdhouse under a heat lamp and placed a temperature gauge inside. The temperature inside the house was between 9 to 11 degrees Fahrenheit (5 to 6 degrees Celsius) cooler when transparent wood was used than when glass was, suggesting that this new material could serve as an energy-efficient alternative to glass in windows.

To further expand the transparent wood's potential applications, the team also incorporated silver nanowires into certain samples. This addition allowed the wood to conduct electricity, which could be useful for wearable sensors or coatings for solar cells. Silver nanowires aren't biodegradable, but the team hopes to conduct further experiments using other conductive materials like graphene to maintain their fully natural transparent woods.

Though additional research is needed to boost the transparency of the woods, Baruah is happy that this initial step used natural and inexpensive materials. "I want to send a message to my undergraduate students that you can do interesting research without spending thousands of dollars," he concludes.

The research was funded by Kennesaw State University and Purafil Inc., an air filtration manufacturer.

sources-science daily

How cells respond to stress is more nuanced than previously believed

 The body's cells respond to stress -- toxins, mutations, starvation or other assaults -- by pausing normal functions to focus on conserving energy, repairing damaged components and boosting defenses.

If the stress is manageable, cells resume normal activity; if not, they self-destruct.

Scientists have believed for decades this response happens as a linear chain of events: sensors in the cell "sound an alarm" and modify a key protein, which then changes a second protein that slows or shuts down the cell's normal function.

But in a new study published today in the journal Nature, researchers at Case Western Reserve University have discovered a cell's response is more nuanced and compartmentalized -- not fixed or rigid, as previously thought.

The groundbreaking research suggests this adaptive response to stress -- which the researchers call "split-integrated stress response" or s-ISR -- could potentially be exploited to kill cancer cells and more effectively treat neurodegenerative diseases.

Maria Hatzoglou, professor of the Department of Genetics and Genome Sciences at the Case Western Reserve School of Medicine and the study's principal investigator, found for the first time a cell's response to stress can be fine-tuned depending its nature, intensity and duration. This flexibility provides novel insights into how cells in organisms -- from yeast to humans -- adapt to their environment.

"This study represents a new way of thinking about cellular stress," Hatzoglou said. "ISR is not a one-size-fits-all system like we used to think. Instead, it can change and adjust depending on the type, strength and length of the stress the cell is experiencing."

The study

The study used mouse models of Vanishing White Matter Disease, which causes progressive degeneration of the brain's white matter in children, leading to neurological problems like motor difficulties, seizures and cognitive decline.

Hatzoglou's research revealed that cells carrying the gene causing the disease had mutations in the key protein normally responsible for shutting down operations in the cell under stress. Somehow, the brain cells adapt and mostly function normally but are exceptionally vulnerable, self-destructing even under mild stress.

The research team, which included colleagues at Case Western Reserve, McGill University and Karolinska Institute, determined how the cells reacted explains why patients show significant decline in cognitive and motor abilities after relatively minor stressors like fever or mild head trauma.

Other late-onset neurodegenerative diseases like multiple sclerosis and amyotrophic lateral sclerosis (better known as ALS) may share a similar mechanism, the researchers said. Diseased brain cells adapt to preserve functions under normal conditions, but modest stressors accelerate decline.

Understanding this adaptation to stress could lead to new targets for cancer chemotherapy, Hatzoglou said, because cancer cells respond to stressors like chemotherapy in one of two ways: either self-destruct or mutate to preserve their function, becoming resistant to the treatment.

sources-science daily

Melting ice, more rain drive Southern Ocean cooling

 Global climate models predict that the ocean around Antarctica should be warming, but in reality, those waters have cooled over most of the past four decades.

The discrepancy between model results and observed cooling, Stanford University scientists have now found, comes down mainly to missing meltwater and underestimated rainfall.

"We found that the Southern Ocean cooling trend is actually a response to global warming, which accelerates ice sheet melting and local precipitation," said Earle Wilson, an assistant professor of Earth system science in the Stanford Doerr School of Sustainability and senior author of the March 27 study in Geophysical Research Letters.

As rising temperatures melt Antarctica's ice sheet and cause more precipitation, the Southern Ocean's upper layer is growing less salty -- and thus, less dense. This creates a lid that limits the exchange of cool surface waters with warmer waters below. "The fresher you make that surface layer, the harder it is to mix warm water up," Wilson explained.

But this freshening is not fully represented in state-of-the-art climate models -- a flaw that scientists have long recognized as a major source of uncertainty in projections of future sea level rise. "The impact of glacial meltwater on ocean circulation is completely missing from most climate models," Wilson said.

Reconciling global discrepancies

The mismatch between observed and simulated sea surface temperatures around Antarctica is part of a larger challenge for scientists and governments seeking to prepare for climate impacts. Global climate models generally do not accurately simulate the cooling observed over the past 40 years in the Southern Ocean and the eastern Pacific around the equator or the intensity of the warming observed in the Indian and western Pacific Oceans. There is also a discrepancy between simulations and the observed frequency of La Niña weather conditions, defined by the eastern Pacific being colder than average.

Warming events in the Southern Ocean over roughly the past eight years have somewhat diminished the 40-year-long cooling trend. But if sea surface temperature trends around the globe continue to resemble patterns that have emerged in recent decades, rather than shifting toward the patterns predicted in simulations, it would change scientists' expectations for some near-term impacts from climate change. "Our results may help reconcile these global discrepancies," Wilson said.

Oceans globally have absorbed more than a quarter of the carbon dioxide emitted by human activities and more than 90% of the excess heat trapped in our climate system by greenhouse gases. "The Southern Ocean is one of the primary places that happens," said lead study author Zachary Kaufman, a postdoctoral scholar in Earth system science.

As a result, the Southern Ocean has an outsized influence on global sea level rise, ocean heat uptake, and carbon sequestration. Its surface temperatures affect El Niño and La Niña weather patterns, which influence rainfall as far away as carlifornia

sources-science daily

How did the large brain evolve?

 Two specific genes that evolve exclusively in humans jointly influence the development of the cerebrum. Researchers from the German Primate Center -- Leibniz Institute for Primate Research and the Max Planck Institute of Molecular Cell Biology and Genetics have discovered this in a recently published study. They have thus provided evidence that these genes contribute together to the evolutionary enlargement of the brain.

The results of the study show that the two genes act in a finely tuned interplay: one ensures that the progenitor cells of the brain multiply more, while the other causes these cells to transform into a different type of progenitor cell -- the cells that later form the nerve cells of the brain.

In the course of evolution, this interplay has led to the human brain being unique in its size and complexity.

The newly gained insights not only provide a deeper understanding of the evolutionary development of our brain but could also help to better comprehend how certain developmental disorders or diseases of the brain arise.

'Our findings deepen the fundamental understanding of brain development and provide new insights into the evolutionary origins of our large brain.

In the long term, they could contribute to the development of therapeutic approaches for malformations of the brain,' says Nesil Eşiyok, first author of the study.

Various methods were combined for the study: In addition to animal experiments with mice, alternative methods such as chimpanzee brain organoids were also used.

'The remarkable feature of our study is that the results from animal experiments and alternative methods complement each other well and mutually confirm their findings.

. This not only emphasizes the high significance of our results, but could also help to reduce the need for animal experiments in the future by further developing, refining and confirming alternative methods,' explains Michael Heide, the study's lead researcher.

The German Primate Center (DPZ) -- Leibniz Institute for Primate Research conducts biological and biomedical research on and with primates in the fields of infection research, neuroscience and primate biology. The DPZ also maintains five field stations in the tropics and is a reference and service center for all aspects of primate research. The DPZ is one of the 96 research and infrastructure facilities of the Leibniz Association.

sources-science daily

How movement affects the way the brain processes sound and sight

 A research team at the Institute for Basic Science (IBS) has uncovered a fundamental principle of how the brain prioritizes vision and hearing differently depending on whether we are still or in motion. The study, led by Dr. LEE Seung-Hee, Associate Director of the IBS Center for Synaptic Brain Dysfunctions and Associate Professor at KAIST, provides new insights into how movement alters the brain's sensory decision-making process.

In daily life, we constantly process visual (sight) and auditory (sound) information to navigate the world. For instance, when watching a movie, our brain seamlessly integrates images and sounds to create a complete experience. However, when moving -- such as when walking on a busy street -- our brain may prioritize visual information over sound.

Until now, it was unclear how the brain decides which sense to prioritize in different situations. This is particularly relevant for individuals with sensory processing disorders such as autism or schizophrenia, where the brain may struggle to integrate sensory information correctly. Understanding how the brain naturally shifts between sensory inputs could lead to better treatments for these conditions.

To investigate this phenomenon, the research team conducted behavioral experiments on mice while tracking real-time brain activity using miniature microscopes and optogenetics (a method that uses light to control neurons). The mice were trained to respond to both visual and auditory cues while running on a treadmill.

The researchers found that when the mice were stationary, their brains relied more on sound to make decisions. However, when they were moving, their brains shifted to prioritize vision.

Further analysis revealed the specific brain circuits responsible for this switch:

- The posterior parietal cortex (PPC), a key region for decision-making, played a central role in prioritizing sensory information. When the PPC was turned off, mice could no longer make decisions based on visual cues and relied more on sound instead.

- The secondary motor cortex (M2) acted as a "sensory gatekeeper." When the mice were moving, M2 sent inhibitory signals to the auditory cortex, blocking auditory signals from reaching the PPC. This effectively made vision the dominant sense during motion.

- Despite this suppression, the auditory cortex continued processing sounds, meaning the brain was adjusting how it integrates sensory information rather than completely ignoring sound.

This study demonstrates that the brain does not treat all sensory inputs equally at all times -- instead, it dynamically adjusts depending on movement and environmental needs. When stationary, sound is more useful for detecting nearby events, while vision takes priority during movement because it is more reliable for navigation.

This discovery could have important implications for understanding and treating sensory processing disorders, where the brain may struggle to properly filter and prioritize sensory inputs.

Dr. LEE Seung-Hee, the study's lead researcher, emphasized the significance of the findings, stating, "Our study reveals how the brain flexibly shifts between vision and hearing based on behavior. This natural adaptability is crucial for survival, and understanding it could help us develop better treatments for individuals with sensory integration difficulties."

sources-science daily

Neurons in brain that regulate energy levels and body temperature

 Scientists at Pennington Biomedical Research Center have gained greater clarity in the brain regions and neurons that control metabolism, body temperature and energy use. Featured in the February edition of the journal Metabolism, Dr. Heike Münzberg-Gruening and a team of researchers discovered which chemicals influence the signals that control how much energy the body uses. In "Leptin Receptor Neurons in the dorsomedial hypothalamus require distinct neuronal subsets for thermogenesis and weight loss," researchers laid out the pathways, chemicals, neurons and brain regions that are activated.

In previous research, Dr. Münzberg-Gruening and her team identified that leptin receptors, or Lepr, control the metabolic effects of Leptin. These receptors are neurons in the dorsomedial hypothalamus, or DMH, which is a nucleus located in the hypothalamus at the base of the brain. In their latest study, researchers found that these Lepr neurons communicate using two different chemical signals: glutamate, which excites neurons, or GABA, which calms neurons.

The study revealed that the neurons that send signals to the brain region called the raphe pallidus -- a region that controls metabolism -- only use glutamate to send their signals. The neurons that signal to another brain region -- the arcuate nucleus, which regulate body weight, satiety and metabolism -- use only GABA. These neurons also have special receptors that allow them to respond to the new weight-loss medications known as GLP-1 receptor agonists.

"This discovery sheds light on the fundamental neuronal interplay that influences how much energy the body uses and how the body adjusts to changes in temperature levels or food availability," said Dr. Münzberg, professor in Pennington Biomedical's Central Leptin Signaling Lab. "This research expands our knowledge of the circuitry of thermoregulation and emphasizes the unique capability of leptin signaling in the DMH to promote beneficial metabolic effects. It also clarifies leptin signaling's role in the stability of body weight and energy usage. We believe these neurons manage the body's ability to adapt to a variety of environmental changes like ambient temperature or food scarcity by integrating signals like leptin and the gut hormone glucagon-like-peptide-1 or GLP-1."

In addition to identifying which neurons from the DMH affect certain regions and their functions, researchers further found that some Lepr neurons are muted by leptin, while others were activated specifically when certain indirect signals were blocked. This suggests that the DMH is part of a larger neuronal network and leptin increases the muting effect of outside connections with the DMH, but when these outside connections are blocked, leptin is able to un-mute DMH neurons. Such networks may be relevant to integrate and, if necessary, override environmental and humoral signals to allow proper adaptation of body energy balance.

This study further clarifies that Lepr neurons are a unique selection of DMH neurons that promote metabolic benefits. They might also explain the paradox that robust weight loss with GLP-1-based medications is able to override the slow metabolism usually associated with weight loss, but this paradox needs to be further tested in future studies.

"There are still so many fundamental processes in our bodies and brains that remain a mystery to us, and that's exactly what drives our researchers at Pennington Biomedical -- to explore these unknowns, make new discoveries and deepen our understanding of metabolism," said Dr. John Kirwan, Executive Director of Pennington Biomedical Research Center. "I want to congratulate Dr. Münzberg and her team on this exciting discovery. It really showcases the incredible work that is happening in our Pre-Clinical Basic Science labs, and I can't wait to see what they'll discover next."

sources-science daily