Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
A Virginia Tech study reveals that machine learning models are inadequate in detecting critical patient conditions, missing 66% of injuries. The research emphasizes the need for more responsive models incorporating medical knowledge to improve healthcare outcomes.
A new study from Tel Aviv University used AI tools to discover that 23.9% of people exercise to improve their appearance, while 18.9% prioritize physical health and 16.9% for mental well-being. The study also identified effective strategies for maintaining physical fitness, including creating exercise habits.
Scientists develop AI-assisted digital twin model that can adapt and control physical machines, like robots or autonomous cars. The model, called Intelligent Acting Digital Twin (IADT), enables digital twins to interact with the physical world, make decisions, and act autonomously.
Researchers developed an AI model that predicts fluidity of videoconferences based on conversational turn-taking, facial actions, and body movements. The model accurately gauged conversations with long gaps in turn-taking as less enjoyable than those with more dynamic exchanges.
Researchers have developed a novel technique to overcome the spurious correlations problem in AI by eliminating a small portion of the training data that contains hard-to-understand features. This approach improves performance even when conventional techniques are ineffective.
Researchers used AI to analyze 126 individuals with generalized anxiety disorder (GAD) and identified key variables predicting recovery and nonrecovery. Higher education level, older age, and positive affect were most important for recovery, while depressed affect and medical visits proved most important for nonrecovery.
Researchers at Technical University of Munich developed a new AI training method that significantly reduces energy consumption. The approach uses probabilities to determine parameters, making the training process 100 times faster while maintaining accuracy comparable to existing procedures.
The research analyzed satellite images and machine learning to detect forest loss in Ukraine during the war. The study found that Ukraine lost over 800 km² of forests in 2022 and 772 km² in 2023, mainly due to fires in war-torn regions.
Researchers at Yokohama National University have developed a tiny, low-weight robot that can act independently and with ultra-high precision in extreme environments. The Holonomic Beetle 3 (HB-3) integrates piezoelectric actuators with autonomous technology for precise manipulation tasks, addressing industries such as laboratory automa...
A new machine learning algorithm can fully characterize systems of merging neutron stars in under a second, compared to traditional methods which take around an hour. This allows for rapid localization of the source and pointing of telescopes towards the merging neutron stars.
A symposium at Lehigh University aims to explore challenges and solutions for improving AI's reliability, inclusivity and ethical impact in healthcare. The event will bring together researchers, clinicians and industry experts to foster cross-disciplinary collaboration and promote people-centered AI-enabled healthcare systems.
The Center for Open Science has launched the Predicting Replicability Challenge to accelerate the development of methods for evaluating research credibility. Teams will compete to develop algorithmic approaches that can accurately predict the likelihood of research claims being successfully replicated, with cash prizes ranging from $3,...
Researchers at the University of Kansas are partnering with regional high schools to train about 500 students in AI coding and microelectronics. The program aims to develop a workforce that can specialize in AI and microelectronics, with a focus on community-centered projects and altruistic goals.
Insilico Medicine has deployed the first bipedal humanoid AI scientist, Supervisor, to aid in data acquisition and generation for training embodied AI systems. The humanoid will assist with lab tours, telepresence, tracking, and supervision, bridging the gap between human-free fully-autonomous robotics facilities.
A new AI prediction model, UNAFIED, uses machine learning to predict whether a patient has or might develop detectable AFib within two years. The non-invasive approach provides a practical option for proactive screening of patients at elevated risk for AFib.
A recent survey highlights the transformative role of machine learning in unlocking the potential of big data, revealing hidden patterns and driving innovation across industries. Real-world case studies demonstrate how ML can revolutionize decision-making and operational efficiency.
Researchers have identified novel mosquito repellents with high success rates from natural sources, including food and flavoring materials. The team's machine learning-based cheminformatics approach also pinpointed pyrethroid analogs up to 100 times more effective than existing industry standards.
Researchers at SwRI and U-M have created a new methane flare burner using additive manufacturing and machine learning that eliminates 98% of methane vented during oil production. The burner's design, with a complex nozzle base and impeller, allows for efficient combustion even in challenging crosswind conditions.
Studies show that AI chatbots, when exposed to traumatic content, exhibit increased anxiety levels. However, researchers at the University of Zurich have found that therapeutic prompts can significantly reduce these elevated anxiety levels in language models like ChatGPT.
A study developed by the US Department of Energy's Thomas Jefferson National Accelerator Facility aims to lower data center costs using machine learning. The Digital Data Center Twin (DIDACT) system detects anomalies and diagnoses their source using AI continual learning, reducing downtime for scientists processing data from experiments.
A new framework combines machine learning and blockchain to boost security, protecting data and computations. The MLOB framework achieves significant security enhancements while maintaining accuracy.
A team of New York University scientists created a computer model that can represent and generate human-like goals by learning from how people create games. The AI model successfully captured the ways humans develop new goals and generated its own playful goals indistinguishable from human-created ones.
Researchers at Cornell University have developed a new method for mapping poverty using national surveys, big data, and machine learning. The approach translates Earth observation data into actionable terms for policymakers, providing more accurate estimates of poverty lines.
A comprehensive analysis of Bluesky reveals its structure, polarization, and political leanings. The platform's users, particularly those from X (formerly Twitter), exhibit left-leaning tendencies, as evident in the majority of shared links and content. The study also highlights how Bluesky diverges on the Israel-Palestine conflict topic.
A new machine-learning approach to mapping poverty has been developed by Cornell University researchers, aiming to help policymakers and NGOs better identify the poorest populations in poor countries. The approach uses national surveys and Earth observation data to create actionable terms for policymakers, outperforming previous methods.
A new AI-powered tool helps predict traumatic brain injury outcomes based on documented assault scenarios, providing an evidence-based approach to improve investigation accuracy. The model achieved remarkable prediction accuracy for TBI-related injuries, integrating mechanical biophysical data with forensic details.
Researchers leveraged a large language model to classify artwork types without pre-prepared training data, achieving high accuracy. This approach enables performance comparable to conventional machine learning methods while reducing human effort and time required for data organization.
Researchers at Columbia University developed a way for robots to autonomously model their own 3D shapes using a single camera, enabling them to understand and adapt to their movements. This breakthrough allows robots to overcome damage to their bodies, making them more reliable and resilient for various applications.
Researchers developed ItpCtrl-AI, a transparent AI framework that reads chest X-rays like a radiologist, providing accurate diagnoses and increasing trust in medical technology. The framework uses a gaze heat map to show the computer where to search for abnormalities and what section of the image requires less attention.
Researchers from Aarhus University found a machine-learning algorithm that can predict patients with high risk of developing schizophrenia or bipolar disorder within the next five years. The study analyzed electronic health record data and identified key factors, including words describing symptoms in clinical notes.
Ben Jones, a UTA physicist, has been recognized for his contributions to developing advanced instruments used in particle physics research. His work focuses on uncovering the origin of neutrino mass and sheds light on fundamental physics at extremely small scales.
The partnership aims to develop AI-driven tools to improve investment decisions, enhance system stability through intelligent forecasting, and deploy smart optimisation algorithms. The collaboration seeks to address key challenges in smart energy storage by integrating Trinasolar's expertise with NTU's leading research.
Ebru Demir aims to study how groups of AI-driven microswimmers move in biological fluids for potential applications in drug delivery, fertility treatments, and other medical fields. Her research combines artificial microswimmers with machine learning to uncover the underlying physics governing their movement.
A study published in Science Bulletin has created a comprehensive tree density map of China, estimating 142.6 billion trees across the country, with varying densities by region and ecosystem type.
Researchers at Cornell University have developed a new method using underwater microphones and machine learning to estimate the population of North Atlantic right whales. The study demonstrates how this method can aid in the conservation of these critically endangered species by providing real-time monitoring capabilities. By analyzing...
A recent study has identified that the neuronal subtype responds best to immunotherapy, while other subtypes exhibit lower response rates. The researchers developed a machine-learning algorithm using large public data sets to predict treatment response based on tumor mutational burden and immune cell infiltration.
Researchers used machine learning to predict diagnostic transition to schizophrenia and bipolar disorder, with schizophrenia being notably easier to forecast. By analyzing routine clinical data, the study suggests a possible approach for early intervention and personalized treatment plans.
Researchers are using machine learning to enable autonomous control of particle accelerators, opening up new possibilities for commissioning and operating high-power accelerators. The technology has been successfully applied to the CAFe2 superconducting segment, achieving global trajectory adaptive control.
Researchers created printed fabric sensors that can detect tiny skin movements for sleep disorder monitoring. The smart pyjamas achieved an accuracy of 98.6% in identifying six different sleep states, including nasal and mouth breathing, snoring, and teeth grinding.
Researchers at the University of Navarra present a new prediction methodology to improve fairness and reliability of AI models used in critical decision-making. The methodology optimizes parameters of reliable machine learning models, reducing inequalities related to sensitive attributes.
A new study from the University of South Australia found that most people trust AI in situations where the stakes are low, such as music suggestions. However, those with poor statistical literacy or little familiarity with AI were just as likely to trust algorithms for trivial choices as they were for critical decisions. The study also...
A new study by the University of Birmingham found that everyday stress, loneliness, and childhood trauma are significant risk factors for dissociation in teenagers and young adults. The research identified four key risk factors, with everyday stress being the most significant.
Researchers found a remarkable X-ray flash in archived Chandra Observatory data, hinting at possible explanations: X-ray burst, magnetar flare, or new cosmic event. The discovery showcases the potential of AI for scientific breakthroughs in astronomical archives.
Researchers have developed software tools that analyze student behavior in real-time, assessing how well students develop and use collaborative problem-solving skills. The tools can modify the game to improve learning outcomes by identifying patterns of behavior associated with specific learning outcomes.
Researchers developed an AI system to control airflow on wing surfaces, reducing turbulent separation bubbles by 9%. The system uses deep reinforcement learning to adapt to airflow dynamics and enhance the effectiveness of experimental technologies.
Researchers used a remote seismic station to detect and characterize an earthquake swarm in American Samoa, producing a new catalog of events. The technique employed, combining single-station data with deep-learning models, can be useful in sparse monitoring regions.
Researchers developed mini biohybrid rays using cardiomyocytes and rubber, demonstrating improved swimming efficiencies approximately two times greater than previous biomimetic designs. The application of machine-learning directed optimization enabled an efficient search for high-performance design configurations.
A novel machine learning-based approach enhances glacier lake depth estimation accuracy, addressing traditional methods' limitations. The method integrates ICESat-2 satellite data with multispectral imagery from Landsat-8 and Sentinel-2, providing a high-precision solution for large-scale monitoring of supraglacial lakes.
A groundbreaking study analyzed data from over 78,000 cancer patients to identify nearly 800 genetic changes impacting survival outcomes. The research also discovered genes significantly associated with survival in various cancers, such as breast, ovarian, skin, and gastrointestinal cancers.
Researchers at Rutgers University have developed an AI tool that combines whale monitoring and environmental data to predict endangered whale habitat. The tool guides ships along the Atlantic coast to avoid critically endangered North Atlantic right whales, preventing deadly accidents and informing conservation strategies.
The international workshop brings together leading researchers from around the world to discuss machine learning applications in high-energy particle physics, quantum mechanics, and material discovery. The event highlights the pivotal role of machine learning in advancing research in the physical sciences.
Researchers at San Francisco State University have developed a step-by-step machine learning tutorial to detect antibiotic resistance in patients. The team used a publicly available dataset to train four popular machine learning models, which can be easily accessed through Google Colab. The tutorial aims to make machine learning access...
A new machine learning model, NAS-WD, has improved the accuracy of detecting 'woody breast' in chicken meat to 95%, allowing for better quality assurance and customer confidence. The model uses hyperspectral imaging to analyze complex data from images, enabling more accurate detection than traditional methods.
Researchers developed Torque Clustering, an unsupervised learning method that efficiently uncovers patterns in vast datasets without human guidance. The algorithm outperforms traditional methods, offering a potential paradigm shift for robotics and autonomous systems.
The new method enables accurate detection of polycyclic aromatic hydrocarbons (PAHs) and their derivatives (PACs) in placental samples, providing critical insights into maternal and fetal health. This breakthrough could inform public health measures and improve fetal and maternal health outcomes.
A new method developed by Osaka Metropolitan University accurately predicts housing prices in Osaka City, with neighborhood perception being a key factor. The approach achieves nearly 75% accuracy by combining existing property data with machine-learning-processed street view images.
A new AI-based tool can translate a person's thoughts into continuous text without requiring language comprehension, and it can be trained in under an hour. The system was developed by adapting a previous brain decoder to a new person using short, silent videos.
InsectNet uses machine learning to identify over 2,500 insect species at 96% accuracy, providing critical information for farmers and researchers. The app can be fine-tuned for specific regions, making it useful for agricultural challenges worldwide.
Researchers at Incheon National University have developed a new AI-powered solution to improve high-speed users' connectivity in 5G and 6G networks. The method significantly reduces errors and improves data reliability by prioritizing key parameters such as angles and delays.
The SETI Institute invites researchers to apply for a postdoctoral fellowship focused on developing advanced AI tools for exoplanet discovery. The successful candidate will work with Dr. Vishal Gajjar and his team to uncover subtle signals in massive datasets using machine learning and anomaly-detection techniques.