Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
A recent study employs machine learning to guide the design of novel materials for CO2 capture, identifying elemental composition and textural properties as key factors. The research team's findings suggest prioritizing adsorption parameters and surface area optimization for high CO2 adsorption efficiency.
Researchers develop crowd-assisted deep learning system to analyze disasters, integrating human intelligence with AI models for better results. The project aims to improve AI's interpretability and accuracy in disaster assessment applications.
The Ariel Machine Learning Data Challenge has announced its winners, ML Analytics and TU Dortmund University, who developed highly accurate solutions for observing exoplanet atmospheres despite instrument noise. The winning teams will receive a cash prize of €500 and present their research at various conferences.
A team of NTU Singapore scientists developed a predictive computer model called NSGA-II, which proposed strategies to reduce COVID-19 infections and deaths by an average of 72%. The model recommended timely and country-specific advice on interventions such as home quarantines and social distancing measures.
Researchers developed an AI tool using data from hospitals across four continents to predict oxygen needs of hospital Covid patients anywhere in the world. The study achieved high-quality predictions with a sensitivity of 95% and specificity of over 88%, demonstrating the transformative power of federated learning in healthcare.
Researchers developed a new, accurate method to detect North Atlantic Right Whale up-calls using Multimodal Deep Learning algorithms. The technology outperformed conventional methods in detecting up-calls, non-up-calls, and false alarms.
Researchers developed a machine learning model that identified promising compounds for treating yellow fever. The team synthesized five of the most active molecules and found one with a half-maximal effective concentration of 3.2 uM, suggesting a potential new antiviral drug.
The NYU-led Learning the Earth with Artificial Intelligence and Physics (LEAP) center will combine artificial intelligence and climate modeling to better predict climate change impacts. The center aims to provide more accurate climate predictions by analyzing satellite images, large-scale observational data, and developing new algorithms.
Researchers developed a machine learning model that produced fewer decision-making errors than human behavior analysts. The AI system showed improved consistency and predictability in treatment decisions, with potential applications in diagnosing and treating autism, ADHD, anxiety, and depression.
Researchers developed a robust, deep neural network model to analyze automobile traffic impacts of construction zones. The model estimates hourly traffic volumes without adjustment factors, helping transportation agencies plan for efficient work zone operations.
A team of researchers from the Beckman Institute for Advanced Science and Technology has developed a fast, accurate, and cost-effective COVID-19 test. Using label-free microscopic imaging combined with artificial intelligence, they can detect and classify SARS-CoV-2 in under one minute.
The RIT workshop series on sustainable computing aims to create computers with environmental consciousness from raw materials to recycling. Keynote speakers will discuss trends in computing and its environmental footprint.
The article explores the potential of AI in improving climate modeling by proposing a hybrid approach combining classical and machine learning methods. Researchers found that AI can help improve forecasts, but limitations include computational resource requirements and uncertainties in predictions.
A new platform called LEARNER allows researchers to share private patient data from multiple institutions while protecting privacy. This enables single institutions to access advanced predictive tools and improve patient outcomes.
A new electronic 'nose' has been developed to detect when a lung transplant is beginning to fail, with 86% accuracy. The device uses machine learning algorithms to analyze exhaled breath patterns and identify lung diseases, offering new hope for patients diagnosed with chronic allograft dysfunction.
A team of researchers at Osaka University created a custom dataset to train an AI algorithm to digitally remove unwanted objects from building façade images. The algorithm achieved high accuracy in inpainting occluded regions with digital inpainting.
Researchers at The Hebrew University of Jerusalem have developed a new deep learning artificial infrastructure inspired by individual neurons. Their approach uses complex mathematical modeling to replicate the brain's electrical processes and create more intelligent AI systems.
A new algorithm, Phe2vec, accurately identified patients with certain diseases, outperforming traditional methods in classifying diagnoses. The study suggests that this automation will facilitate further research in clinical informatics.
A new machine learning technology can rapidly screen for genetic syndromes in children, considering facial variability related to sex, age, racial and ethnic background. This innovation has the potential to reduce health inequality in under-resourced areas worldwide.
Researchers at Technical University of Munich have developed a new machine learning algorithm that can analyze complex markets and their equilibrium strategies. This breakthrough has potential applications in auction theory, wireless spectrum auctions, and more.
A new unsupervised machine learning algorithm, B-SOiD, developed by Carnegie Mellon University researchers makes studying animal behavior more accurate and efficient. The algorithm identifies patterns in an animal's body position to discover behaviors, removing human error and bias.
Using complex-valued layers can improve performance against adversarial attacks without sacrificing efficiency. This technique, combined with gradient regularization, allows neural networks to resist small perturbations and maintain accuracy.
A team of scientists from Osaka University developed a machine learning method for classifying the type of building and its primary façade color using deep learning models applied to street-level images. This work may assist in fostering neighborhood cohesion and support urban renewal by providing tailored street-view datasets.
Researchers developed a model to predict the risk of death and unplanned cardiac hospitalization among patients awaiting cardiac surgery. The tool, based on an ethnically diverse group of 62,375 patients, identifies factors such as gender, urban residency, and cardiac symptoms that increase the risk of adverse events.
A multi-site, multi-disorder resting-state magnetic resonance image database was compiled to harmonize data from patients with various diseases, measured at 14 sites. The dataset comprises over 2,400 samples and includes 'traveling-subject' data to minimize inter-site differences.
Researchers offer standards for machine learning studies in life sciences to promote reproducibility and advance knowledge. The bronze standard requires data, code, and models to be publicly available, while the silver and gold standards add more information to facilitate duplication of training processes.
The team used machine learning technique generative adversarial networks to digitally remove clouds from aerial images, generating accurate datasets of building image masks. This work may help automate computer vision jobs critical to civil engineering, enabling the detection of buildings in areas without labeled training data.
Researchers have developed an approach that predicts accurate structures computationally, overcoming the problem of determining molecular shapes. The algorithm succeeds even when learning from only a few known structures, making it applicable to difficult-to-determine molecules.
Scientists have created a system dubbed "NanoporeTERs" allowing cells to express themselves in a whole new light. These new reporter proteins can detect multiple protein expression levels and shed new light on biological systems, enabling deeper analysis than before.
A recent study has uncovered the evolutionary forces at play in the aging of the blood system and identified individuals at increased risk of blood cancer. The research provides a robust indicator for classifying patients with ARCH mutations, allowing for more frequent screening and early treatment.
A new Yale University study reveals that social media platforms like Twitter amplify expressions of moral outrage over time, encouraging users to express more outrage with increased likes and shares. This finding has significant implications for leaders and policymakers who use these platforms.
A machine learning analysis has identified 50 genes strongly associated with neurological aging in both Drosophila fruit flies and humans. The study suggests that fruit flies could be used as a model organism to further investigate aging-related processes.
Researchers have created an AI software that uses Minecraft to test its ability to plan for future events and solve complex tasks. The software, developed by Penn State researchers, aims to advance artificial intelligence in areas such as robotics, logistics management, and drone flight.
A new machine learning tool developed by Stanford researchers can accurately predict extreme precipitation events in the Midwest, accounting for over half of all major US flood disasters. The approach uses atmospheric circulation patterns to identify factors responsible for recent increases in Midwest extreme precipitation.
Scientists at CiTIUS have developed a new fast support vector classifier (FSVC) that significantly improves data classification using Machine Learning techniques. The FSVC is much faster and operates with less memory than traditional approaches, making it suitable for large-scale classification problems.
Researchers at C-Crete Technologies have developed a method that utilizes deep learning to quickly predict and design novel hybrid organic-inorganic materials, offering improved materials design for various industries. By feeding quantum mechanics calculations to layered machine learning based on artificial neural networks, they can un...
Researchers from Nara Institute of Science and Technology developed a machine learning program that accurately predicts the location of proteins related to actin in cells. The program achieved a high degree of similarity with actual images, showing promise for future applications in cell analysis and artificial cell staining.
A new study by AKASA finds that its Read, Attend and Code machine learning model exceeds the performance of current state-of-the-art models for automatic coding of inpatient clinical notes. The technology surpasses human coders in terms of both efficiency and accuracy.
A team of computer scientists has developed an assembly selection process that balances representation and fairness in citizens' assemblies. By using a machine learning-based algorithm, the researchers ensure that all volunteers have an equal chance of being chosen, regardless of demographic quotas or education level.
A new study from Washington University in St. Louis shows that guided by sparsity, silicon neurons learn to pick the most energy-efficient perturbations and wave patterns, enabling an emergent phenomenon of efficient communication between neurons. This research has significant implications for designing neuromorphic AI systems.
The University of Washington is leading a new NSF institute focused on using artificial intelligence to understand dynamic systems, which describe chaotic situations where conditions are constantly shifting and hard to predict. The institute aims to integrate fundamental AI theory with applications in critical technological areas.
Researchers at NYU Tandon School of Engineering develop a holistic view for machine learning in healthcare, incorporating data about communities and environments. They introduce a novel approach to understanding fairness relationships using causal inference, synthesizing a means to assess effects of sensitive macro attributes.
Researchers at Texas A&M University have developed a method to cool steam turbines using phase change materials, potentially reducing fresh water usage. By leveraging machine learning techniques, they created a system that can predict when and how much of the PCM will melt and freeze, maximizing cooling power and capacity.
A study from McGill University and France uses AI to identify factors predicting suicidal behavior in students, finding self-esteem as a major predictor. Approximately 17% of students exhibited suicidal behaviors, highlighting the need for large-scale screening tools.
Researchers developed a new framework using deep learning techniques to create 3D visualizations from X-ray data. This method is hundreds of times faster than traditional methods, enabling scientists to analyze large amounts of 3D data more efficiently.
Researchers applied supervised machine learning to lidar data from fishery surveys, automating the identification process in regions with a strong possibility of harboring fish. The technique decreased manual inspection in datasets from Yellowstone Lake and Gulf of Mexico studies by 61.14% and 26.8%, respectively.
Researchers predict two nicotine biomarkers, NMR and TNE, in smokers of multiple ethnicities using machine learning approaches. These models can estimate nicotine biomarker levels from DNA data or existing genomic information.
A new machine learning-based model predicts ICU patients' mortality risk based on characteristics such as demographics, comorbidities, and APACHE II score. The model overcomes traditional approaches' weak points, offering a better alternative for personalized medical predictions.
Researchers have created an AI-powered method to automate the identification of promising lunar landing and exploration areas. The technique uses machine learning and deep learning frameworks to accurately detect craters and rilles with precision rates as high as 83.7%, outperforming existing state-of-the-art methods.
A team of researchers developed a method to study how parents adjust their language to match their child's speech development. They found that caregivers have an incredibly precise knowledge of their child's language and use this information to fine-tune the linguistic input they provide.
The 'SynRap' project aims to accelerate the production of large amounts of synthetic data by a factor of one thousand using machine learning algorithms. The project will assess the quality of generated data sets in high energy density physics and high energy physics research areas.
Researchers developed resource-efficient federated learning to train analytic models on local data, enabling coalition partners to learn similar tasks without sharing sensitive data. The new technology provides cutting-edge capability over adversaries and is crucial for defense applications.
A new AI-based tool has been developed to predict genomic subtypes of pancreatic cancer using histology slides, offering a potential solution for patient molecular stratification. The tool, trained and validated on machine learning models, can be used in clinical practice worldwide.
Researchers aim to identify when individuals are falling out of the flow state by detecting physiological cues with off-the-shelf sensors. AI will be used to introduce interactive stimuli that nudge subjects back into a higher cognitive state.
Researchers from KIT and universities in Göttingen and Toronto develop machine learning methods to simulate material behavior, achieving high accuracy and speed. Hybrid methods combining machine learning and molecular mechanics are also suggested to accelerate simulations of large biomolecules.
Lehigh University engineers use Frontera supercomputer to simulate photovoltaic fabrication and train AI to optimize energy production. Their 'physics-informed machine learning' approach reduces time required to reach optimal process by 40%.
The Shadow Figment technology uses AI-powered deception to keep attackers engaged in a pretend world, rewarding them with false signals of success while defenders learn about the attackers' methods. This creates a distraction that allows defenders to take action and protect real systems.
Researchers developed P-Flash, an AI-powered tool predicting flashover in burning buildings. It uses temperature data from heat detectors and shows promise in anticipating simulated flashovers, identifying unmodeled physical phenomena that can improve forecasting in real fires.
Researchers from UTSA, UCF, AFRL, and SRI International have developed a new method that improves how artificial intelligence learns to see. By adding noise to multiple layers of a neural network, the team creates more robust representations of images recognized by AI, leading to better explanations for AI decisions.
According to a review article in Science Robotics, researchers are making progress in learned robot manipulation, which enables robots to adapt to changing stimuli. The authors propose nine promising areas for future exploration, including representation learning, modular design, and task/skill customization.