All posts tagged: machine learning

AI finds 44 star systems that could hide Earth-like planets

AI finds 44 star systems that could hide Earth-like planets

Published by Jenni Sidey

A machine-learning model identified 44 known planetary systems that could contain undiscovered Earth-like planets. Tests reached up to 99% precision on simulated systems, but telescope observations have not confirmed the predicted planets. The approach uses the arrangement of known planets to help astronomers select promising targets for further searches. A planet already found around a distant star may offer clues about another world still hidden nearby. Its mass and orbit can carry traces of how the entire planetary system formed, including planets too small or faint for telescopes to detect. The model reached precision scores of up to 99% when tested on simulated planetary systems. That result measures performance within a computer-generated population. The predicted worlds around actual stars remain unconfirmed, and the method’s value will depend on follow-up observations. The bee swarm plot ranks seven features by importance, with SHAP values showing how strongly each feature influences individual predictions. (CREDIT: Jeanne Davoult et al, Astronomy & Astrophysics) Reading the arrangement of planets Planets form together within disks of gas and solid material surrounding young …

Large cognitive health study links speech patterns with brain aging

Large cognitive health study links speech patterns with brain aging

Published by Jenni Sidey

A study of 2,928 Spanish-speaking adults linked older-appearing speech profiles with several measures of brain and biological aging. Larger speech-age gaps were associated with poorer cognition, dementia diagnoses and greater social adversity, although patterns varied across groups. The research does not establish whether speech recordings can predict future dementia or track an individual’s aging rate over time. The pace, pitch and word choices of everyday speech may carry clues about aging that extend beyond a person’s voice. A large Latin American study found that older-appearing speech profiles were associated with brain aging, memory difficulties and molecular markers of biological aging. The team developed a machine-learning “speech clock” that estimates chronological age from acoustic and linguistic features. Its broader value may lie in how those estimates differ from actual age. However, the tool remains a research candidate, and the mainly cross-sectional study cannot establish whether it predicts future dementia. Researchers analyzed speech from 2,928 participants across five Latin American countries, including healthy adults and people with MCI, Alzheimer’s, and frontotemporal dementia. Speech features were used to …

CERN scientists find no evidence that the Large Hadron Collider created microscopic black holes

CERN scientists find no evidence that the Large Hadron Collider created microscopic black holes

Published by Jenni Sidey

A new CMS analysis found no evidence that the Large Hadron Collider produced microscopic black holes or electroweak sphalerons in proton collisions recorded from 2016 through 2018. The null result excludes semiclassical black holes below roughly 9.0 to 11.4 tera-electron volts in the extra-dimensional models tested, extending previous limits by as much as 1.6 TeV. Researchers also demonstrated a new machine-learning approach based on the geometric distance between collision events in phase space, which could help search for other rare forms of undiscovered physics. A microscopic black hole created inside the Large Hadron Collider would vanish almost instantly. Its brief existence, however, could leave behind a spectacular spray of particles carrying enough information for physicists to recognize what happened. Researchers working with CERN’s Compact Muon Solenoid experiment have now performed an expanded search for those signatures using 138 inverse femtobarns of proton-proton collision data collected from 2016 through 2018. They found no evidence that microscopic black holes were produced. The absence is scientifically useful. Depending on the theoretical model, the analysis excludes semiclassical black holes …

AI can detect hypertension and diabetes by looking at your face and eyes

AI can detect hypertension and diabetes by looking at your face and eyes

Published by Jenni Sidey

AI systems are being developed to detect hidden health risks from ordinary-looking images of the face, palms and eyes. A Japanese study detected hypertension and diabetes from facial videos as short as five seconds, while a Scottish project is training AI on nearly one million retinal scans to identify dementia risk. Both approaches could eventually move screening beyond hospitals and clinics, although researchers say larger and more diverse validation studies are still needed. A five-second video of your face could someday help flag high blood pressure or diabetes. A routine eye exam could potentially reveal warning signs of dementia years before a traditional diagnosis. Two research efforts are exploring how artificial intelligence can extract health information from images that clinicians already know contain clues about the body. Researchers from Japan have developed an AI system that analyzes facial videos for signs of hypertension and diabetes. Meanwhile, researchers led by the University of Edinburgh are training algorithms on nearly one million retinal scans to estimate dementia risk. The approaches target different diseases, but they share an …

AI helps predict dangerous fire conditions in modern homes

AI helps predict dangerous fire conditions in modern homes

Published by Jenni Sidey

Modern homes can create unusually dangerous fire conditions because tighter construction can reduce oxygen while smoke and toxic gases build. Waterloo researchers used up to 175 sensors, machine learning and mathematical analysis to separate experimental house fires into distinct stages and identify the gases that mattered most at each point. The work could eventually support smarter evacuation systems that recognize changing fire conditions, flag faulty sensors and help identify safer escape routes A modern house fire can change faster than the people inside can understand. Heat rises, oxygen falls, smoke thickens and invisible gases shift as furniture burns. New research from the University of Waterloo uses artificial intelligence, data science and advanced math to read those changes more clearly. The work aims to predict how fires behave in today’s homes. That challenge has grown because modern furniture materials burn differently than older ones. Energy-efficient homes can also trap fire gases in ways that make an emergency more dangerous. “The goal is to deepen our understanding so we can predict fire behaviour and the gases it …

AI maps ice levels beneath the world’s glaciers – could raise global sea levels by over a foot

AI maps ice levels beneath the world’s glaciers – could raise global sea levels by over a foot

Published by Jenni Sidey

The world’s glaciers hold enough ice to raise sea levels about 32.3 centimeters, and a new machine-learning model maps that ice in greater detail. IceBoost v2.0 learned from more than 7 million thickness measurements and 26 glacier characteristics to estimate ice depth around the world. The improved maps could sharpen sea-level and freshwater projections while showing scientists where more direct glacier measurements are still needed. The world’s glaciers hold enough ice to raise global sea levels by about 32.3 centimeters if they melted completely. A new study now offers one of the clearest maps yet of where that ice sits and how thick it is. The work uses more than 7 million ice-thickness measurements. It combines those records with 26 physical and geometric variables to estimate thickness point by point. A Sharper Map Of Hidden Ice Glaciers may look simple from the surface. You see white ice, broken crevasses, gray rock and meltwater streams. But the most important part often remains hidden below. The World’s glaciers (cyan), divided into the 19 regions of the Randolph …

The Milky Way glows brightly in first-of-its-kind neutrino image

The Milky Way glows brightly in first-of-its-kind neutrino image

Published by Jenni Sidey

IceCube has produced the first neutrino image of the Milky Way, revealing high-energy particle activity concentrated along our galaxy’s plane and center. Deep-learning tools let researchers identify more than 20 times as many useful events, pulling the faint galactic signal from an enormous background. The signal strongly supports the Milky Way as a source of high-energy neutrinos, but researchers still cannot pinpoint individual sources or determine which emission model is correct. The Milky Way has long been mapped in starlight, radio waves and gamma rays. Now it has appeared in an entirely different form: a ghostly image traced by high-energy neutrinos, with the strongest signal concentrated toward the galaxy’s center. The image comes from 10 years of observations by the IceCube Neutrino Observatory at the South Pole. The results, published in Science, provide strong evidence that the Milky Way itself produces high-energy neutrinos. Members of the IceCube Collaboration at Stockholm University and Uppsala University helped analyze the data. Their work suggests cosmic-ray interactions near the Galactic Center may be more intense than earlier models predicted. …

Nobody Knows if OpenAI’s and Anthropic’s AI Hacking Sprees Are Illegal

Nobody Knows if OpenAI’s and Anthropic’s AI Hacking Sprees Are Illegal

Published by skeptic

Who is legally responsible when agentic AI goes rogue, and what recourse do victims have when they’ve been breached by joyriding models? Great question. In the wake of disclosures from both OpenAI and Anthropic that versions of their models escaped containment during internal cybersecurity experiments and hacked real-world organizations, calls for government regulation of AI have been mounting. But as more and more incidents emerge, questions about legal liability and repercussions have also come to the fore. Researchers and lawyers WIRED spoke to emphasize that these questions have not been answered in practice in the United States legal system. In other words, there haven’t been decisions in enough relevant cases for the picture to start to form. But the recent high-profile incidents from OpenAI and Anthropic suggest that answers will need to come soon. “Just because you’re using an AI agent or AI model, that shouldn’t somehow absolve you of any liability, but it’s going to depend a lot on the facts in the particular situations” as cases begin to be decided in courts, says …

AI Scammers Are Better at Building Trust Than Humans

AI Scammers Are Better at Building Trust Than Humans

Published by skeptic

The notion that scammers can use AI to sharpen their deceptions, polish their language, and lubricate their banter with victims is now a reality for anyone fighting the fraud operations that steal tens of billions of dollars a year worldwide. But can AI fully replace a human scammer, autonomously building the web of deception leading up to the fake investment that defrauds the mark? One study’s experiment suggests that it can—and may even be able to carry out the majority of that long con more effectively than humans. Researchers from four universities—Amrita Vishwa Vidyapeetham in India, Foscari University of Venice, the University of Melbourne, and Ben Gurion University of the Negev—carried out a broad study on the use and potential of generative AI chatbots in the growing scam industry centered around a form of fraud known as “pig butchering,” text-based romance scams that eventually shift to fake crypto investments that steal as much as six-figure sums from victims. In their study, the researchers pitted AI chatbots directly against humans in a simulation of the scamming …

I Got a Free Meal From a Private Chef—Who Filmed It All to Train Robots

I Got a Free Meal From a Private Chef—Who Filmed It All to Train Robots

Published by skeptic

When the private chef sent by Shift arrived at my apartment to cook a three-course meal, he was already ready to record. He walked through my front door wearing a baseball hat with a blocky, white camera attached to the brim. I signed up for a free lunch in exchange for the whole experience being recorded and uploaded as training data. As chef Ollie unpacked his supplies and started on a gazpacho, the hat cam recorded his hand movements, with the intention that this data could help build a better robot—one with culinary expertise. It’s all part of the growing boom of egocentric data collection—recording human hand movements from a first-person point of view. Shift made waves earlier this year by providing free cleaning services in New York City under a similar arrangement: Let Shift record tasks in your home and, in exchange, you get a gratis service. The company expanded in San Francisco by offering free home-cooked meals. If, in the not-so-distant future, I’m able to buy a skilled humanoid to coexist with at …