Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
Researchers from the UK and Australia are combining forces to tackle pressing global challenges through eight collaborative projects focusing on AI, sustainability, space, and cybersecurity. The partnership aims to develop innovative technologies that work for people and the planet, not just profit.
Researchers from Waseda University used machine learning to enhance the performance of photomechanical crystals, achieving up to 3.7 times greater force output compared to previously reported values. This breakthrough has significant implications for remote-controlled actuators, medical devices, and energy-efficient systems.
Researchers have created a breakthrough photonic chip that can train nonlinear neural networks using light, accelerating AI training while reducing energy use. The chip uses a special semiconductor material to reshape how light behaves, enabling reconfigurable systems with wide mathematical function expression.
A novel method generates binary class labels in highly imbalanced datasets, addressing the challenge of labeling severely imbalanced data. The approach outperformed traditional methods, providing a more efficient way to identify fraud while minimizing false positives.
Researchers from East China Normal University developed an AI-driven system to analyze classroom videos, revealing teacher-centered instruction prevails in primary and secondary schools. The study also found that older students engage less in critical discussions and more in structured questions.
Dr. Latifur Khan, a renowned computer science professor, has been elected as an AAAS Fellow for his pioneering work in machine learning and big-data analytics. He developed innovative solutions to adapt machine learning models to cybersecurity risks and created an AI-driven tool to analyze political conflict and violence.
A novel machine learning-driven approach uses deep-learning operator-surrogate models to monitor critical degradation indicators in nuclear power facilities. This technique provides real-time predictions and addresses limitations of physical sensors or classical modeling predictions.
Researchers developed an AI model that creates synthetic fibrosis patterns to aid in treating atrial fibrillation. The system accurately mimics real heart scarring and enables clinicians to test different treatment approaches on digital models before performing procedures.
A new study published in Newton uses artificial intelligence to identify complex quantum phases in materials, significantly speeding up research into quantum materials. The breakthrough applies machine-learning techniques to detect clear spectral signals, allowing for a fast and accurate snapshot of phase transitions.
The researchers propose a novel defense algorithm, Wavelet-Based Adversarial Training (WBAD), to protect medical digital twins. The two-stage defense mechanism achieves 98% accuracy in breast cancer prediction, even under adversarial attacks, providing a comprehensive and effective defense against cyberattacks.
The Deep Computational Text Analyser (DECOTA) is an open-access AI tool that transforms open-ended survey responses into clear themes in minutes, not months. Developed by the University of Bath, DECOTA delivers insights around 380 times faster and over 1,900 times cheaper than human analysis.
Fiddler crabs use vibrational signals to communicate during courtship, and a new study reveals that signal features are influenced by the male's claw size. Researchers found that males with larger claws produce higher-energy signals, allowing females to assess their quality from afar.
A University of Ottawa-led research team has deciphered the message that serotonin conveys to the brain, discovering a 'prospective code for value' that explains why neurons are activated by rewards and punishments. This finding has implications across multiple fields, including neuroscience, psychology, and psychiatry.
Researchers utilized machine learning models to identify key surface attributes modulating immune response, paving the way for improved implant materials. The study revealed pivotal factors regulating cytokine secretion and offered insights into designing alloys with optimized immunoregulatory functions.
A new study published in Annals of Internal Medicine found that AI diagnostic and treatment recommendations were more accurate than those made by physicians in a virtual urgent care setting. In 77% of cases, AI recommendations were rated as optimal, compared to 67% for physicians' decisions.
A recent study from UTSA researchers reveals that large language models (LLMs) can pose a serious threat to programmers who use them to help write code. The study found that up to 97% of software developers incorporate generative AI into their workflow, and 30% of code written today is AI-generated.
The collaboration aims to advance research on brain health with a focus on Alzheimer's disease. Initial projects will use CAS Content Collection and advanced technologies, including AI models and quantum computing to build and train disease-specific models.
A new artificial intelligence-based method detects genetic markers of antibiotic resistance in bacteria, potentially leading to faster and more effective treatments. The method, called Group Association Model, uses machine learning to identify key mutations linked to drug resistance, reducing false positives and misdiagnoses.
Researchers warn of misunderstandings in handling AI models, highlighting conditions for confidence in predictions. Explainability methods are crucial to understand algorithmic decisions, but interpreting results requires caution due to AI limitations.
The conference gathered international researchers to discuss AI's role in drug discovery and development, including generative AI strategies for designing chemical compounds. The speakers emphasized the significance of personalized medicine, where therapies will be tailored to each patient's unique molecular profile.
University of Queensland scientists used machine learning to predict when bacteria evolved oxygen use, finding some bacteria used it before photosynthesis. The study suggests aerobic metabolism occurred before oxygenic photosynthesis, dating back around 3.2 billion years.
Researchers develop AI model to predict lightning-induced wildfires with over 90% accuracy, integrating data from satellites, weather systems, and environmental factors. The model has the potential to transform emergency response and disaster management worldwide, saving lives and preserving ecosystems.
A new framework developed by MIT researchers allows large language models (LLMs) to break down complex planning problems into manageable parts and find optimal solutions using software optimization tools. The framework achieves an 85% success rate on nine complex challenges, outperforming the best baseline.
A UC Irvine study found that machine learning algorithms using electronic health records can predict two-year dementia risk among American Indian/Alaska Native adults. The researchers identified novel predictors of all-cause dementia, including health service utilization.
Rachid Guerraoui has made groundbreaking contributions to distributed computing, shaping the landscape of concurrent environments. He also promotes computer science education in Africa, fostering academic excellence through initiatives like the Excellence in Africa program.
Cleveland Clinic researchers successfully tested quantum computing's ability to simulate proton affinity, a fundamental chemical process critical to life. The study used machine learning applications on quantum hardware, achieving higher accuracy than classical computing in predicting proton affinity.
The CONCERTO Project aims to advance the understanding and modelling of the terrestrial carbon cycle, reducing uncertainty in climate predictions. By integrating Earth observation data and innovative models, the project will contribute to improved climate policy and global efforts towards carbon neutrality.
A new study developed an AI model that can predict whether bacteria will become antibiotic-resistant by analyzing their genetic data. The model shows that antibiotic resistance is more easily transmitted between genetically similar bacteria and mainly occurs in wastewater treatment plants and inside the human body.
A new machine learning-based method analyzes corporate reporting for responsibility and innovation, providing insights into company actions. Essi Nousiainen's research also explores blockchain-related trends in corporate reports.
Researchers developed new AI models, InstaNovo and InstaNovo+, to vastly improve accuracy and discovery in protein science. These models excel in tasks such as de novo peptide sequencing, identifying microorganisms, and discovering novel peptides, with implications for personalized medicine, cancer immunology, and beyond.
Researchers at NCSA and UICOMP have created an improved, automated screening method for anxiety and depression disorders using machine learning and acoustic analysis of verbal fluency data. The new tool offers promise for overcoming barriers to diagnosis and treatment.
Juan Gamella's mini-labs provide a flexible test environment for new AI algorithms, allowing researchers to test their performance beyond simulated data. The mini-labs help identify issues early on, enabling targeted improvements to underlying mathematical assumptions and algorithms.
Göttingen research team develops infomorphic neurons that learn independently and self-organize among neighboring neurons. This allows the smallest unit in the network to control its own learning, enabling novel machine learning approaches and a deeper understanding of brain function.
Crop mapping uses satellite imagery to create accurate crop type maps in various regions. The research team trained machines to recognize crops from satellite images, achieving high accuracy rates. However, models pre-trained on general image datasets performed better than those pre-trained on satellite images.
A new machine learning model improves risk assessment for patients with myelofibrosis seeking a transplant, identifying high-risk patients with a 40% chance of dying within a year. The open-access model provides a practical tool for clinicians to enhance shared decision-making with their patients.
A machine learning algorithm has been developed to diagnose coeliac disease with high accuracy, outperforming human pathologists in over 97 cases. The AI tool has the potential to speed up diagnosis and reduce delays in receiving an accurate diagnosis for patients suffering from this autoimmune disease.
Professor Yousung Jung's team uses LLMs to accurately predict and explain material synthesizability, overcoming limitations of existing methods. This technology is expected to accelerate material design and reduce development time for the semiconductor and secondary battery industries.
The study used AI analysis on over 4,200 clinical reports to identify key factors in autism diagnosis. Criteria related to repetitive movements and special interests were strongly linked to an autism diagnosis, while socialization criteria were not highly specific.
A study published in Biological Psychiatry identified the Shisa7 gene as a key driver of heroin addiction. The research team used machine learning to analyze brain tissue from human opioid users and found that modulating this gene's expression influenced heroin-seeking behavior and cognitive flexibility.
Researchers used large language models to analyze healthcare records of over 1,000 children with suspected autism, finding that current criteria prioritize socialization skills and not enough on interests and natural behaviors. The study suggests revising the criteria to focus more on repetitive behaviors and special interests.
Scientists have identified COVID-derived molecular mimics that trigger autoimmunity, with some associated with type 1 diabetes and multiple sclerosis. The study suggests people with specific genetics may be at higher risk of COVID-induced autoimmunity.
Researchers evaluate traditional precipitation phase partitioning methods and machine learning models, revealing near-freezing temperatures create inherent limitations in distinguishing between rain and snow. Accurate identification is critical for weather forecasting, hydrologic modeling, and climate research.
A new study from Duke University reveals that dopamine levels in the brains of baby birds increase with every effort they make while practicing their songs, regardless of success. This discovery has implications for understanding human learning and neurological disorders.
A recent study refines Siberia's land cover data using machine learning techniques, revealing a high-precision map that enhances climatic predictions. The new dataset improves assessments of carbon flux and ecosystem changes, providing essential insights for climate scientists.
A new approach to AI developed by Texas A&M University engineers mimics the human brain's neural processes, integrating learning and memory in a single system. This 'Super-Turing AI' has the potential to revolutionize the industry by reducing energy consumption and environmental impact.
A study using machine learning models found that people with IBD are at risk for premature death when developing other chronic health conditions earlier in life. Chronic conditions like arthritis, hypertension, and mood disorders were common causes of death among those with IBD.
A new real-time surface PM10 retrieval framework uses interpretable automated machine learning to provide accurate data across China. The framework demonstrates robust generalization and stability, outperforming previous studies in cross-validation and rolling iterative validation experiments.
Researchers at University of Toronto develop a new framework to optimize laser Directed Energy Deposition (AIDED) for higher quality and more reliable metal parts. The AIDED framework uses machine learning to predict optimal process parameters and enhance the accuracy and robustness of finished products.
A University of Cincinnati study found that machine learning models can aid clinicians in treating patients with spreading depolarizations (SDs), a condition that can cause significant brain damage. The algorithm was able to identify SD events with high sensitivity and specificity, detecting many events not identified by human scoring.
Researchers developed an automated MRI processing and machine learning software to diagnose Parkinson’s disease, reducing diagnostic time by up to 96%. The software uses diffusion-weighted MRI and a noninvasive biomarker technique to identify neurodegeneration in the brain, providing more precise diagnoses.
This special issue explores AI's applications across various subspecialties, including coronary interventions, structural heart disease, and cardiovascular imaging. It highlights the importance of responsible AI integration and addressing bias in decision support systems.
The collaboration aims to accelerate the development and commercialization of inait's innovative AI technology, using its unique digital brain AI platform. It will focus on joint product development, go-to-market strategies, and co-selling initiatives, initially targeting the finance and robotics sectors.
Researchers discovered plastic-degrading enzymes in landfills worldwide, suggesting a promising method for plastic recycling. The study identified 31,989 possible enzymes and predicted protein functions using machine learning and tertiary structure modeling.
Researchers developed a simple, inexpensive tool using robotics and artificial intelligence to analyze dried salt solutions from images. The method increases the accuracy of chemical analysis in scenarios where large samples are difficult to obtain, making it valuable for space exploration, law enforcement, and hospital use.
Researchers at Saarland University are developing leaner, customized AI models and techniques like knowledge distillation to reduce energy consumption. These smaller models enable small and medium-sized businesses to access powerful AI technology without a large technical infrastructure.
A machine learning model developed by a team of researchers from Washington University in St. Louis can accurately predict which adolescents with HIV are at risk of nonadherence to antiretroviral therapy. The model incorporates socio-behavioral and economic factors, including economic stability, education, and family structure.
Scientists identify the origin of magnetic moment enhancement in an iridium-doped iron-cobalt alloy through high-throughput X-ray measurements. The study reveals that Ir addition leads to increased electron localization and spin-orbit coupling, resulting in enhanced magnetic moments.
Researchers developed a portable system using AI to spot cognitive impairment by measuring subtle differences in motor function. The device accurately identified 83% of participants with mild cognitive impairment (MCI), offering potential for early intervention and improved outcomes.
A new study challenges the long-held belief that fear is primarily communicated through facial expressions, suggesting instead that situational context plays a critical role in fear recognition. The research involved analyzing real-life fear reactions in videos and found that facial expressions alone fail to reliably signal fear.
Astronomers have created a detailed 3D map of dust in the Milky Way galaxy, providing new insights into the effects of dust on celestial observations. The map reveals unexpected properties of interstellar dust clouds, including a steepening extinction curve in areas of intermediate density.