Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
A study published in Critical Care identified eight different trauma phenotypes associated with lower in-hospital mortality when treated with tranexamic acid. The researchers used a machine learning model to analyze data from over 50,000 patients and found subgroups of patients who received no benefit from treatment.
Researchers at University of Missouri are developing software that allows drones to fly independently, perceiving and interacting with their environment while achieving specific goals. This technology has the potential to assist in mapping and monitoring applications, such as 3D or 4D advanced imagery for disaster response.
Researchers developed a new method to predict thermoelectric materials using AI, avoiding trial-and-error and overfitting. The approach achieved remarkable accuracy in predicting newly available materials, providing guidance for experiments.
Researchers at Klick Labs developed an algorithm to detect deepfakes with 80% accuracy by analyzing speech pause patterns, offering a solution to the growing problem of AI-generated content. The study's findings suggest that vocal biomarkers can distinguish between real and fake voices, providing a novel approach to flagging deepfakes.
A team of researchers has developed a machine learning interatomic potential that predicts molecular energies and forces acting on atoms, reducing computational time and expense. This breakthrough enables scientists to study complex chemistry systems with greater accuracy and speed.
Researchers developed a novel machine learning-based depth estimation technique for satellite-derived bathymetry, improving accuracy in coastal regions with unique characteristics. The model demonstrated generalizability and potential for enhancements through incorporation of additional seabed spatial data.
Researchers from Kobe University developed an AI image recognition algorithm that can predict mouse behavior based on brain functional imaging data, achieving 95% accuracy. The model identified critical cortical regions for behavioral classification and demonstrated near real-time speeds.
A new AI tool can predict which breast cancer patients are at risk of chronic arm swelling after surgery and radiotherapy. The tool, developed by international researchers, uses machine learning algorithms to analyze patient data and provides easily understandable explanations for doctors and patients.
Researchers developed machine learning models that can recognize emotions in voice recordings as short as 1.5 seconds with high accuracy comparable to humans. The study used three ML models and achieved an accuracy of over 90%, with potential applications in therapy, interpersonal communication technology and more.
Artificial intelligence has been developed to spot COVID-19 features in lung ultrasound images, combining computer-generated images with real scans to identify signs of disease. The tool holds potential for developing wearables that track illnesses like congestive heart failure and monitor fluid buildup in patients' lungs.
A new study uses machine learning to classify fossils of extinct pollen with high accuracy, leveraging morphological features and phylogenetic data. The model successfully placed nearly all specimens within Podocarpus based on their shape and form.
The Rensselaer Polytechnic Institute researcher is working with the Tachyon Project to create surrogate machine learning models that can simulate and analyze particle physics data in real-time. This project aims to improve scientific discovery and workflow performance for scientists at Fermilab and ALCF.
Researchers used hyperspectral imaging and machine learning to classify rapeseed maturity, achieving high accuracy rates. The study identified key wavelengths and preprocessing methods that improved model performance, offering a non-destructive solution for uniform seed maturity.
A new approach to nutrient level detection in rubber leaves uses semi-supervised learning with unlabelled hyperspectral data, outperforming traditional supervised methods. The study balances class imbalance using resampling techniques, enhancing classification accuracy and reliability.
Researchers at Rice University have developed a custom-built miniaturized chemical vapor deposition (CVD) system that can observe and record the growth of 2D MoS2 crystals in real-time. Through advanced image processing and machine learning algorithms, they were able to extract valuable insights into the growth processes of these mater...
Researchers developed chronological age prediction models by analyzing gene expression changes in the prefrontal cortex, identifying genes associated with aging and potential mechanisms. The models showed high correlation with age and demonstrated female and male-specific differences.
Researchers have discovered new genetic mechanisms related to spinocerebellar ataxia type 37, a rare neurological disorder that affects balance and movement. The study employed advanced techniques such as CRISPR/Cas9 gene editing and machine learning to uncover the disease's underlying causes.
A review published in Intelligent Computing outlines the strengths of automatic approaches to designing metaheuristics, which can lead to more successful outcomes and reduce redundant, metaphor-based algorithms. The authors encourage research that relies on automatic design, utilizing modular software frameworks and configuration tools.
Scientists use machine learning algorithms to model atomic masses of nuclide chart, complementing research on nuclear structure and astrophysical processes. The approach enables physics-based extrapolations and provides information on 'missing physics'.
Elfi Kraka and colleagues develop a novel machine learning interatomic potential called ANI-1xnr that accurately simulates atomic-level interactions in various environments. The model has the potential to aid in understanding planetary care, drug interactions and exploring cosmic materials.
A study by Washington State University found that exposure to multiple air pollutants was associated with asthma symptoms among elementary school children. The researchers identified 25 combinations of air pollutants linked to asthma, with one group from a lower-income neighborhood experiencing higher exposure levels.
Researchers at ETH Zurich taught ANYmal, a quadrupedal robot, to perform parkour and navigate rubble using machine learning. The robot uses its camera and artificial neural network to determine obstacles and perform movements likely to succeed based on previous training.
A new method using multi-target regression and hyperspectral imaging enhances crop nutritional analysis, predicting multiple element concentrations with improved accuracy. The approach considers inter-element relationships, outperforming traditional single-target regression methods.
A new study led by Worcester Polytechnic Institute aims to determine whether AI can help doctors predict which patients will benefit from mindfulness-based stress reduction in managing chronic lower back pain. The research uses machine learning and physiological data from fitness sensors to detect patterns that may not be apparent to d...
A new study using a combination of traditional and machine learning techniques found a previously unknown theropod species in the famous Kem Kem beds of Morocco. The analysis confirmed the presence of Noasauridae, a rare group of small theropods with long necks.
Researchers are developing a new framework to integrate renewable energy sources with the power grid using machine learning. The goal is to ensure efficient and stable operations while maximizing wind and solar power usage.
Researchers at Bar-Ilan University discovered that each filter recognizes small clusters of images, with sharpened recognition as layers progress. This breakthrough can improve AI performance by reducing latency and memory usage while maintaining accuracy.
Researchers at the University of California - San Diego developed a mathematical formula that reveals how neural networks learn to detect relevant patterns in data. The Average Gradient Outer Product (AGOP) formula helps interpret which features the network is using to make predictions, improving the accuracy and reliability of AI syst...
Researchers developed an AI framework that combines dimension reduction techniques with a new clustering algorithm to quickly identify groups of viral genomes at risk. This enables proactive response measures like tailored vaccine development, potentially eliminating emerging variants before they spread.
TaskMatrix.AI uses APIs to connect general-purpose foundation models with specialized models for specific tasks. The tool can perform digital and physical tasks, provide interpretable responses, and learn continuously.
Researchers develop framework to assess relative value of rules and data in AI models, improving efficiency and accuracy in scientific problems. The framework optimizes model training by tweaking the influence of different rules, filtering out redundant ones, and identifying synergistic relationships between rules.
New research combines images with computer-enabled analysis to tackle biological questions globally. Imageomics aims to improve image classification and analysis using machine learning and computer vision, enabling faster scientific discoveries.
Researchers at Carnegie Mellon University have created a new machine learning model that can simulate reactive processes in diverse organic materials and conditions. The model, called ANI-1xnr, performs simulations with significantly less computing power and time than traditional quantum mechanics models.
Researchers at XPANCEO and Nobel laureate Konstantin S. Novoselov unveil new properties of rhenium diselenide and rhenium disulfide, enabling novel light-matter interaction. This breakthrough has huge potential for integrated photonics, healthcare and AR applications.
Researchers developed a machine learning model to assess the quality of health news stories, outperforming laypeople in evaluating their accuracy. The model used expert criteria to classify articles as 'satisfactory' or 'not satisfactory', highlighting the need for multiple criteria in evaluating news.
A study using machine learning classifies galaxy mergers and finds that mergers are not strongly associated with black-hole growth. Cold gas at the center of the host galaxy is necessary for rapid growth, suggesting a more complex relationship between galaxy evolution and supermassive black holes.
The new AI model uses a visual map to explain each diagnosis, helping doctors follow its line of reasoning and check for accuracy. The tool aims to catch diseases in their earliest stages, making it easier on doctors and patients alike.
A new machine learning approach using Alu elements in blood plasma has improved the detection of cancer early by reaching 98.9% specificity. The test can catch 41% more cancer cases than existing biomarkers and is expected to complement other cancer tests.
Researchers develop AI model to predict combinations of gene perturbations that can transform cell type or restore diseased cells. The study's findings have potential applications in regrowing injured tissues and transforming cancer cells back into normal cells.
Researchers discovered similar genetic elements underlying vocal learning in humans, bats, whales, and seals. AI-powered analysis identified 50 gene regulatory elements associated with vocalization and autism in multiple mammalian species.
A new American Heart Association scientific statement highlights the value of AI technology in improving cardiovascular disease patient outcomes. Despite its potential, AI applications face gaps and challenges, including robust clinical validation and addressing biases.
Researchers at RMIT University have developed a reprogrammable light-based processor that could enable efficient quantum computations. The device, which uses photons to carry information, reduces 'light losses', a critical factor in maintaining computation accuracy.
Researchers argue that current automated toxicity detection methods can improve but require human intervention to review decisions. Companies can take steps to improve working conditions and platform cultures that prioritize kindness and respect.
An international team of scientists developed AI technology to analyze limited data on rare diseases. The method uses multi-layer networks to explore relationships between genes in patients, revealing genetic causes and severity. This breakthrough opens new avenues for treating rare diseases, including myasthenic-congenital syndromes.
A new AI model developed by MIT researchers breaks down complex warehouse navigation into smaller chunks, identifying optimal areas for decongesting robots. The technique improves efficiency by nearly four times, opening up potential applications in other complex planning tasks.
A groundbreaking technology recognizes human emotions in real time, combining verbal and non-verbal expression data for accurate emotional information extraction. The system features a personalized skin-integrated facial interface that enables self-powered, flexible, and transparent emotion recognition.
Researchers have developed a method to decode mouse neural activity, enabling accurate determination of location and direction within an open environment. This breakthrough could inform the design of intelligent machines that navigate autonomously without GPS or satellite guidance.
The USC team created a low-cost, customizable learning kit for students to build their own 'robot friend' using the Blossom robot. The three-part module provides hands-on experience and instruction on various AI aspects, including robotics, machine learning, and software engineering.
A new process using artificial intelligence (AI) predicts carbon cycles in agroecosystems, surpassing traditional models in accuracy and speed. This breakthrough enables fair and accurate compensation for farmers, fostering trust in carbon markets and promoting sustainable practices.
A new machine learning model predicts how ingested drugs interact with transport proteins in the body, identifying previously unknown interactions and potential dangers.
A new emulator model improves auroral current system simulations, enabling faster and more efficient space weather forecasts. The Surrogate Model for REPPU Auroral Ionosphere version 2 (SMRAI2) is a million times faster than physics-based simulations and incorporates seasonal effects.
Researchers at UCSF used clinical data and a precision medicine approach to identify early risk factors for Alzheimer’s disease, predicting its onset with 72% accuracy. High cholesterol, osteoporosis, and erectile dysfunction were found to be predictive factors in both men and women.
Researchers at the University of Manchester have developed new methods to simulate blood flow, enabling faster and more accurate modeling of vascular diseases. These advancements have the potential to transform medical treatment and device innovation, providing real-time insights during surgical procedures and improving patient outcomes.
Researchers developed a novel machine learning-based approach to analyze diffuse reflectance spectroscopy data, achieving higher accuracies and speeds than existing methods. The 'wavelength-independent regressor' model overcomes use-error limitations by incorporating diverse datasets, making it suitable for clinical settings.
A team of scientists proposed a general deep learning framework based on DQN algorithm to efficiently design wavelength-selective thermal emitters (WS-TEs) with excellent performance for different applications. The framework autonomously selects materials and optimizes structural parameters for optimal emissivity spectra.
Scientists at Tokyo University of Science used deep learning to predict single-molecule magnets from a pool of 20,000 metal complexes, identifying 70% accuracy in distinguishing between SMMs and non-SMMs.
Researchers developed a multipronged strategy to identify transporters used by different drugs, revealing potential interactions between commonly prescribed antibiotics and blood thinners. The approach has the potential to improve patient treatment and predict potential toxicities.
Researchers are using machine learning tools to understand biological traits from images, enabling new discoveries about life on Earth. Imageomics is analyzing the relationship between observable phenotypes and genome, leading to a better understanding of direct connections.
Researchers found that widely used machine learning tools produce biased results for immunotherapy research, as they rely heavily on datasets from higher-income communities. This can lead to ineffective treatments for lower-income populations. The study highlights the need for accurate and unbiased data in machine learning models.
A recent study found that a commercial machine learning tool was moderately successful in predicting hospitalization-related kidney injury, but struggled to identify high-risk patients. The tool performed better for low-risk patients and Stage 1 HA-AKI cases.