The researchers have developed an AI algorithm called M3GNet that can predict the structure and dynamic properties of any material. The algorithm was used to create a database of over 31 million yet-to-be-synthesized materials with predicted properties, facilitating the discovery of new technological materials.
A Rutgers researcher has created a machine learning model that can estimate arsenic contamination in private wells without sampling the water. The model identifies geological bedrock type and soil type as primary contributors to higher arsenic concentrations, highlighting the need for targeted well testing programs.
A Cornell-led collaboration used machine learning to predict Alzheimer's progression in cognitively normal and mildly impaired individuals. The modeling showed that MRI scans are most informative for asymptomatic cases, while PET scans are more effective for those with mild cognitive impairment.
A new study using artificial intelligence has found that a simple eye test can accurately predict the risk of heart disease. The researchers developed an algorithm that can analyze retinal images to assess cardiovascular health, providing a non-invasive alternative to traditional risk scores.
Researchers developed a new machine-learning framework that enables cooperative or competitive AI agents to consider the future behaviors of all agents, not just their teammates or competitors. This framework, FURTHER, uses two modules: an inference module and a reinforcement learning module, to enable agents to adapt their behaviors a...
A research team developed an optical chip that can train machine learning hardware, improving AI performance and reducing energy consumption. This innovation uses photonic tensor cores and electronic-photonic application-specific integrated circuits to speed up the training step in machine learning systems.
The University of Manchester's Centre for Robotics and AI will focus on interdisciplinary research, exploring applications of robotics in extreme environments and the intersection of AI and humanity. The centre aims to develop robots that can work safely in hostile zones, such as nuclear decommissioning sites.
Researchers developed an AI model to analyze spatial and temporal gait parameters, identifying key features for diagnosing Parkinson's. The model achieved high diagnostic accuracy, reducing the probability of error in clinical assessments.
Researchers from the University of Johannesburg deployed Few Shot Learning (FSL) for NIALM, a non-intrusive appliance load monitoring system. FSL requires only 7 test images to recognize appliances with 97.83% accuracy, making it faster and more cost-effective than traditional Machine Learning.
Researchers demonstrate that tetraplegic users can operate mind-controlled wheelchairs in a cluttered environment after training, with improvements in accuracy and brain activity patterns. The study highlights the importance of long-term training and neuroplastic reorganization for successful brain-machine interface control.
Researchers from Kessler Foundation are enrolling participants in a national trial testing a breakthrough device for improving recovery after stroke. The EMAGINE Stroke Recovery Trial pairs therapeutic exercise with electromagnetic stimulation to enhance brain and spinal cord function.
This study employs machine learning to analyze existing experimental results and predict the device performance of metal halide perovskite solar cells. The authors applied shapley additive explanations (SHAP) analysis to understand the correlations between fabrication processes, composition, and device performance.
Researchers used weather radar to track bird movements and found peak roosting stages shifting earlier due to warmer temperatures. This shift may lead to a shortened pre-migratory season, impacting birds' survival during migration.
Researchers at Oak Ridge National Laboratory have discovered genetic markers for autism, developed recyclable composites to drive the net-zero goal, and created a tool for real-time building evaluation. Additionally, they have made significant progress in growing hydrogen-storage crystals using a novel nano-reactor material.
A study by University of Minnesota researchers found that personalized federated learning may offer an opportunity to develop both internal and externally validated algorithms. This technique enables multiple parties to train AI models collaboratively without exchanging or centralizing data sets, protecting sensitive medical informatio...
Researchers used machine learning to analyze high-frequency oscillations in patients' brains during deep sleep, distinguishing between epileptogenic and non-epileptogenic regions. The study achieved an 85% accuracy rate, suggesting a promising method for predicting seizure outcomes.
Researchers developed an algorithm using amino acid sequences of proteins called T cell receptors to predict patient response to treatment, providing insights into the biology behind an effective response. The algorithm, DeepTCR, identified patterns that are predictive of patient response as accurately as known biomarkers.
Researchers modelled relationship between plant diversity and environmental conditions, capturing how diversity varies along environmental gradients. The models predict highest concentrations of plant diversity in environmentally heterogeneous tropical areas like Central America and the Amazonia.
Researchers developed a method to learn complex Boolean systems, enabling faster and more accurate diagnoses of urinary diseases, cardiac conditions, and financial risks. The technique uses optimal causation entropy to narrow down correct solutions and turn complex diagnostic processes into decision trees.
A new AI-based chemical sensor can accurately detect specific gases in the air by analyzing temperature changes in a microbeam resonator. The device uses machine learning to differentiate between gases with varying thermal conductivities, achieving 100% accuracy in identifying helium, argon, and CO2.
A new study published in The Journal of Nuclear Medicine found that a PET/MRI machine learning model can reliably distinguish between patients with and without lymph node metastases. This breakthrough technology has the potential to eliminate sentinel lymph node biopsy, a common procedure for breast cancer treatment.
Researchers induced brain-like sleep in an artificial spiking neural network, enabling it to retain previously learned information. This breakthrough helps the model avoid catastrophic forgetting and improve its overall performance over time.
SourcePLOS·JournalPLOS Computational Biology·TypeComputational simulation/modeling·DateNov 10, 2022
Researchers created a new set of standards, called FAIR, to manage AI models, making them findable, accessible, interoperable and reusable. This standardization enables cross-pollination across teams and reduces duplication of effort, ultimately facilitating scientific discovery.
Researchers used AI to automate the process of analyzing X-ray snapshots of materials, accelerating the technique by ten times on its own and 100 times with improved hardware. The new method can extract information from a range of previously inaccessible materials, including high-temperature superconductors and quantum spin liquids.
The University of Tsukuba researchers developed a machine-learning model to predict undertriage in phone-based triage systems. The model identified risk factors such as age, sex, and comorbid conditions, which can be used to update protocols and improve patient outcomes.
Researchers developed an intelligent compaction technology that integrates into a road roller, assessing real-time the quality of road base compaction. This improves road construction, reducing potholes and maintenance costs, leading to safer and more resilient roads.
Researchers at CABBI used unmanned aerial vehicles with machine learning methods to select the best candidate genotypes in miscanthus breeding programs. The new method leverages high-resolution aerial imagery and three-dimensional neural networks to estimate crop traits such as flowering time, height, and biomass production.
A new nationwide study found that 34.7% of Canadians experienced severe loneliness in January 2021, with migrants and those experiencing job insecurity being most at risk.
Researchers at University of Jyväskylä used machine learning to predict ACL injuries in elite athletes but found a low overall accuracy rate. The study analyzed the largest data set ever collected and provided valuable insights into the challenges of predicting injuries in individual athletes.
Researchers used machine learning to track turbulent structures in fusion reactors, gaining detailed information on their behavior and heat flows. The approach enables more accurate engineering requirements for reactor walls and could lead to improved energy efficiency.
A team of researchers is developing a new method to evaluate the viability of donor kidneys using optical coherence tomography (OCT) scanning technology. OCT scans provide a more complete view of the organ's overall functional capacity, allowing for better matching with transplant patients. The goal is to increase kidney availability a...
In an experiment pitting human expertise against artificial intelligence, researchers found that the computer program correctly predicted six out of nine proteins capable of self-assembly. This breakthrough suggests machine learning can complement human intuition in biotechnology research.
A large study of over 5,800 tween children found that growing up in a socioeconomically disadvantaged household can have lasting effects on brain development, with different patterns of connections between brain regions observed. Parental education emerged as the most significant factor associated with variations in brain connections.
The US Department of Energy's Oak Ridge National Laboratory has developed a massive geographic dataset, USA Structures, using deep learning to forecast potential damage and accelerate emergency response. The dataset provides critical information on building outlines and attributes, enabling FEMA to prioritize response efforts.
Researchers at MIT have developed a machine-learning model that captures how sounds propagate through spaces, allowing for accurate visual renderings of rooms. This technique has potential applications in virtual and augmented reality, as well as improving AI agents' understanding of their environment.
Researchers at KAUST developed an inverse mixture-design approach using machine learning to create high-performance transport fuels. The model accurately predicted fuel properties and identified suitable blends, offering a promising solution to reduce greenhouse gas emissions.
Researchers from the University of Tokyo's Interactive Intelligent Systems Laboratory developed a new system called LookHere that incorporates natural hand gestures into the teaching process. This approach eliminates extraneous details and provides better input data for machines to create models, resulting in improved efficiency and ac...
A team from Massachusetts General Hospital developed an AI-based method to predict which patients with early-stage melanoma are most likely to experience a recurrence. Machine learning algorithms extracted predictive signals from clinicopathologic features, including tumor thickness and rate of cancer cell division.
Researchers at WVU are developing software for robots to learn and adapt in real-time, inspired by the neural networks of electric fish. The goal is to enable robots to navigate different terrains autonomously without human supervision.
A new study reveals significant gender bias in Reddit discussions about female politicians, with more frequent use of given names and language related to their body or family. This perpetuates stereotypes and can affect voter perception, incentivizing female politicians to conform to online expectations.
Researchers used AI to predict compounds that could neutralize reactive oxygen species causing baldness. They successfully regenerated hair on mice using microneedle patches, providing a promising new treatment for androgenic alopecia.
Researchers at MIT have developed a new approach to identify topological materials using machine learning and X-ray absorption spectroscopy. The method is over 90% accurate in identifying known topological materials and can predict properties of unknown compounds.
Researchers at the University of Illinois developed an AI-powered system that uses a molecule-making machine to find optimal reaction conditions for synthesizing chemicals. The system doubled the average yield of a challenging class of reactions, paving the way for faster innovation and automation in biomedical and materials research.
Researchers at Penn State aim to create a minimally invasive system using machine learning and multimodal biosensing platforms to detect Alzheimer's disease in its early stages. The system has the potential to improve detection accuracy and reduce false positives.
Researchers at FAU aim to empower amputees to maximize their individual potential for controlling the full dexterity of artificial hands using a novel bimodal skin sensor and machine learning algorithms. The project will develop customized prosthetic sockets and training programs to overcome limitations with current sensing technology.
A new project aims to create more race-inclusive AI in medicine by developing a distributed, inclusive data collection and learning framework that relies on smartphone apps. The framework uses federated learning, which allows models to be trained on device data while protecting user privacy.
A data visualization tool, TrafficVis, assists law enforcement in identifying patterns in online escort advertisements that indicate illegal activity. The tool analyzes millions of ads to highlight common phrasing or duplication among them, helping analysts direct investigations and identify human traffickers and victims.
A novel algorithm uses near-infrared spectroscopy to estimate intracranial pressure (ICP) based on hemoglobin levels. The research validates the accuracy of this method using invasive ICP data.
Researchers developed a machine learning-based diagnostic method that combines genomic sequencing and analysis of patients' immune response for remarkable accuracy. The approach identifies and predicts sepsis cases with high accuracy, potentially exceeding current diagnostic capabilities.
Researchers developed an AI model that uses daily step counts to predict the likelihood of unplanned hospitalization during cancer therapy, offering personalized care and potential prevention of treatment interruptions. The model was tested on data from wearable consumer devices and showed strong predictive performance.
The platform enhances neurosurgical procedures by delivering precise treatment and diagnosis in deep brain tissue. Researchers successfully implanted the catheter in live sheep without damage or infection, paving the way for potential human trials within four years.
Researchers have developed a deep learning algorithm that can accurately assess the stage of head and neck cancer using standard CT scans, outperforming expert radiologists. The algorithm demonstrated superior accuracy in measuring the extent of cancer spread, especially for patients with high-risk disease.
Researchers have developed a novel method for antibiotic resistance testing that can analyze bacterial cells in real-time, allowing for faster identification of susceptible and resistant bacteria. This breakthrough technology has the potential to transform microbial screening in clinical and research labs.
Researchers at MIT have developed a new method that uses optics to accelerate machine-learning computations on low-power devices. By encoding model components onto light waves, data can be transmitted rapidly and computations performed quickly, leading to over a hundredfold improvement in energy efficiency.
Researchers developed an AI-based model that combines artificial intelligence and weather forecast models to predict extreme wildfire danger with high accuracy. The new method can produce forecasts of extreme fire danger out to one week at finer scales (4km x 4km resolution), increasing its utility for fire suppression and management.
Researchers use machine learning to generate beta-barrel proteins that can detect metal ions in water, a potential solution for identifying pollutants like lead in waterways. The 'deep-fake' protein is designed to be ideal for binding specific metals and creating conductance differences.
Researchers found that content overlap between critic and user reviews increases movie demand, particularly for movies with mediocre review ratings. The study suggests that producers and marketers can leverage this by engaging with both professional critics and general consumers to find commonalities in their reviews.
FathomNet aggregates images from multiple sources to create a publicly available, expertly curated underwater image training database. The database uses artificial intelligence and machine learning to alleviate the bottleneck for analyzing underwater imagery, accelerating important research around ocean health.
A new AI app improved the accuracy of COVID-19 lateral flow test results, increasing sensitivity from 92% to 97.6%. The app was tested at UK Health Security Agency assisted test centers and found promising results in reducing false negatives.
Omnipose, a deep learning software, can identify various types of tiny objects in micrographs with high precision, including bacteria of all shapes and sizes. It overcomes limitations of previous approaches by handling object overlap and detecting cell intoxication, making it a game-changer for biological image analysis.