Research using machine learning to solve scientific problems. Covers AI quality control, sea-surface temperature estimation, label-free cell analysis, and cyber defense.
Researchers have developed PrISMa, a novel platform that seamlessly connects materials science, process design, techno-economics, and life-cycle assessment to identify effective and sustainable carbon capture solutions. The platform has been tested on over 60 real-world case studies, providing valuable insights for stakeholders.
The study uses machine learning framework to estimate global rooftop area growth, projecting a 20-52% increase by 2050. Rooftop solar power holds significant potential for emerging economies, driving sustainable development and prosperity.
Researchers used machine learning to create highly detailed maps of individual trees, providing valuable information for conservation efforts and ecological projects. The algorithm achieved high accuracy in classifying common tree species, with strengths shown in areas with open space and lower species diversity.
A new model, PAG-STAN, predicts short-term origin-destination demand in urban rail transit systems with remarkable precision. The model improves interpretability and enhances training efficiency through a masked physics-guided loss function.
UCF researchers George Atia and Yue Wang received a $1.2 million DARPA grant to develop AI-based technologies that can help autonomous systems adapt to unknown variables and overcome simulation-to-real gap issues.
Researchers at Weill Cornell Medicine used machine learning to define three subtypes of Parkinson’s disease, each with distinct driver genes and molecular mechanisms. These subtypes may suggest customized treatment strategies for patients, potentially targeting specific drugs such as metformin to slow down progression.
Scientists have developed a new technique that leverages X-ray photon correlation spectroscopy, artificial intelligence, and machine learning to create unique 'fingerprints' of materials. These fingerprints can be analyzed by neural networks to yield new information about material behavior under stress and relaxation.
A new AI technique uses physics-informed machine learning to create high-fidelity atmospheric transmission profiles without prior knowledge, enabling accurate remote sensing data correction. This approach enhances remote sensing capabilities for tasks like target detection, requiring limited data and computational resources.
Researchers identified six clinical subtypes in older adults starting long-term care in Japan, including cardiac disease, respiratory disease/cancer, and insulin-dependent diabetes, which incur higher mortality risks and worsen care needs. These findings can inform optimal interventions for each subtype and influence healthcare policy.
Researchers at Max Planck Institute propose a new method for implementing neural networks with optical systems, which could lead to faster and more energy-efficient alternatives. The approach allows for parallel computations in high speeds limited by the speed of light, and can be applied to various physically different systems.
The Grid Event Signature Library provides an online collection of anonymized datasets containing waveforms, enabling utilities and research institutions to understand the increasingly complex grid. Machine learning can be trained to recognize waveforms that provide early warnings of equipment malfunction, preventing blackouts and damage.
A UCLA-led team created a machine-learning model that can accurately predict short-term CRRT survival, providing a data-driven tool for clinical decision-making. The study aims to improve patient outcomes and resource use by serving as a basis for future clinical trials.
A Danish study of over 100,000 children used machine learning to identify at-risk kids earlier, potentially improving child maltreatment detection and social worker decision-making. The findings suggest that predictive risk models could enhance outcomes for these vulnerable children.
Researchers developed a machine learning approach to identify potential subtypes in diseases, significantly enhancing disease classification and treatment strategies. The model uncovered 515 previously unannotated disease subtypes.
Researchers at University of Cambridge developed machine-learning tool to identify drug-resistant Salmonella Typhimurium bacteria from microscopy images. The algorithm correctly predicted resistance or susceptibility without culturing the bacteria, reducing diagnosis time from days to hours.
Researchers create an analog system that can learn complex tasks like XOR relationships and nonlinear regression, using local learning rules without centralized processor. The system is fast, low-power, and scalable, offering a unique opportunity for studying emergent learning.
A study published in JMIR Medical Informatics found that machine learning can accurately classify patients into differing levels of opioid use disorder (OUD) risk, demonstrating substantial agreement with clinicians' reviews. The research suggests that this technology can enhance personalized and safer care for patients early in opioid...
The college will develop a Knowledge-enhanced Antidepressant Recommendation Dialogue System (KARDS) to engage users in a conversation to identify the appropriate antidepressant medication. The AI chatbot aims to address medication needs of Black and African Americans with depression, a gap in current management.
A new machine learning-based method uses 3D structure of protein backbone with large language models to predict molecular changes that lead to better antibody drugs. The approach resulted in a 25-fold improvement against a virus, outperforming traditional methods that rely on generating huge amounts of data about protein sequences.
Researchers at UAB have developed a method to assess cardiac dynamics in fruit flies using deep learning and high-speed video microscopy. The study uses this approach to analyze the effects of aging and dilated cardiomyopathy on heart function, with potential applications for human cardiovascular research.
Scientists have developed a machine learning program that can identify blobs of plasma in outer space known as plasmoids. The program will analyze data from NASA's Magnetospheric Multiscale (MMS) mission to better understand magnetic reconnection and its effects on the electrical grid.
Researchers developed a one-dimensional convolutional neural network (1D CNN) to compensate for errors due to sample location variations. The model achieved high accuracy, reducing mean absolute error to 0.695% and mean squared error to 0.876%.
The Mount Sinai researchers have developed a comprehensive epidemiological dataset for youth diabetes and prediabetes research, derived from NHANES data collected from 1999 to 2018. The newly launched POND portal aims to facilitate an understanding of factors that may influence youth diabetes risk.
Researchers are working on improving the quality of high frequency wireless networks. Dr. Murat Yuksel is hoping to realize the dream of unimpeded communication at distances near and far. He is developing a smart wireless network system using machine learning, which can fine-tune the networks' efficacy.
Researchers used AI to analyze over 16,000 butterfly images, finding both males and females contribute to diversity among species. The study resolves a century-old debate between Charles Darwin and Alfred Russel Wallace on the role of natural selection in female evolution.
Researchers found that AI models that analyze medical images can predict patient demographics with high accuracy but struggle to diagnose patients from diverse backgrounds. The models use demographic shortcuts, leading to incorrect results for women, Black people, and other groups.
Researchers used AI to map slush on Antarctic ice shelves and found that 57% of all meltwater is held in slush, with a significant impact on ice shelf stability and sea level rise. This discovery could lead to more accurate predictions of ice sheet melting and collapse.
A recent study by Prof. Martin Bichler suggests that dividing Germany into several price zones may not reduce total power costs as expected. In contrast, nodal pricing shows promise in reducing costs by up to 9% due to efficient resource allocation and reduced re-dispatch measures.
Researchers developed a machine learning estimator to classify charge states in quantum dots, enabling automatic tuning of qubits. The estimator achieved high accuracy with visualizations revealing decision-making patterns, paving the way for scaling up quantum computers.
Researchers developed an AI model that accurately predicts metal yield strength by combining physical theory with machine learning. The model outperforms traditional methods, which often rely on extensive experimentation.
Researchers from Florida Atlantic University developed a novel approach using wearable sensors and machine learning to assess balance. The method achieved high accuracy and strong correlation with ground truth balance scores, suggesting it is effective and reliable in estimating balance.
A Chinese research team introduced a novel two-stage framework using stacked transformers for multimodal sentiment analysis, improving the analysis of emotions expressed through modality combinations. The framework was tested on three open datasets and performed better than or as well as benchmark models.
Researchers at Boston University developed an AI model that analyzes speech patterns to predict the likelihood of Alzheimer's disease in patients with mild cognitive impairment. The model achieved an accuracy rate of 78.5% and could potentially revolutionize dementia screening, making it more accessible and efficient.
Researchers developed machine learning models predicting upper secondary education dropout from kindergarten age, using a 13-year longitudinal dataset. The study marks an advancement in early automatic classification, potentially leading to transformative changes in educational systems and policies.
Researchers at FSU used machine learning to analyze patterns in dried salt solution drops, identifying the chemical composition of different salts with accuracy. The tool has potential applications in lab safety testing, rapid screening for suspected drugs and low-cost blood analysis.
University of Texas at Dallas researchers develop AI model that can automatically reroute electricity in milliseconds to prevent power outages. The system uses machine learning to map complex relationships between entities in a power distribution network, enabling faster response times than human-controlled processes.
A new model developed by Flatiron Institute researchers proposes that individual neurons exert more control over their surroundings, which could be replicated in artificial neural networks. This updated model treats neurons as tiny 'controllers' and may lead to better AI performance and efficiency.
A new system named SQUID, a computational tool created by Cold Spring Harbor Laboratory scientists, helps interpret how AI models analyze the genome. It reduces background noise and leads to more accurate predictions about genetic mutations.
Scientists at UVA and Toyota Research Institute create language representations of driving behavior to enable robots to associate words with environmental interactions. This allows cars to provide guidance and adjust speed in challenging situations, improving safety and usability.
Researchers at Bar-Ilan University have discovered a new scaling law that governs how artificial neural networks handle an increasing number of categories for identification. This law reveals how the identification error rate increases with the number of required recognizable objects, impacting AI latency and efficiency.
Researchers at Cold Spring Harbor Laboratory designed a new way for AI algorithms to move and process data more efficiently, inspired by the human brain. This design allows individual AI neurons to receive feedback and adjust on the fly, processing data in real-time.
Research suggests that large-language models could play a role in managing the energy grid, particularly in emergency response, crew assignments, and wildfire preparedness. However, significant challenges remain, including data availability, safety guardrails, and reliability, which must be addressed to ensure safe deployment.
PSICHIC uses sequence data and AI to decode protein-molecule interactions with state-of-the-art accuracy, eliminating costly processes like 3D structures. The tool effectively screens new drug candidates and performs selectivity profiling, offering a more efficient and reliable approach to drug discovery.
A new machine-learning model using serum fusion-gene levels predicts HCC with an accuracy of 83-91%, significantly improving upon current biomarkers like serum alpha-fetal protein. This breakthrough tool may help identify patients at risk and monitor cancer recurrence, leading to improved survival rates.
A new study on learning has provided insights into the balance between habitual and goal-directed behaviors, with implications for AI development. The research suggests that a balance between these two types of behavior is necessary for efficient and adaptable decision-making in AI systems.
Researchers used machine learning to integrate high-throughput transcriptomic, proteomic, metabolomic, and lipidomic profiles to identify four distinct molecular profiles of Alzheimer's Disease. These profiles were associated with varying levels of cognitive function and neuropathological features.
Researchers at UC San Diego School of Medicine are developing an AI model to predict opioid addiction in high-risk patients. The model uses generative artificial intelligence to analyze genomic, social determinants of health, clinical, procedural, and demographic data to identify patients at greatest risk.
A recent study published in Critical Care Medicine found that real-time machine learning alerts significantly improved patient outcomes by predicting clinical deterioration. The study showed that patients who received AI-generated alerts were 43% more likely to have their care escalated and had a lower risk of death.
The article explores how AI can be applied to the electric power and energy industry, demonstrating its potential as a valuable technology for asset management. Machine learning techniques are showcased as a solution to improve efficient and sustainable energy networks.
A recent study published in The Lancet Oncology found that an AI system can detect prostate cancer nearly seven percent more significantly than a group of radiologists using MRI scans. Additionally, the AI identifies suspicious areas less often, potentially reducing unnecessary biopsies by half.
A new machine-learning system can automatically produce detailed maps from satellite data to show locations of likely beetle-killed spruce trees in Alaska. This helps forestry and wildfire managers make critical decisions as the beetle infestation spreads, affecting approximately 2 million acres across Southcentral Alaska.
A new computer vision technique developed by MIT engineers significantly speeds up the characterization of newly synthesized electronic materials. The technique automatically analyzes images of printed semiconducting samples and quickly estimates two key electronic properties: band gap and stability.
Researchers at the Complexity Science Hub analyzed friendships and listening habits to find social networks are a crucial predictor of song popularity. The study showed that individuals with strong influence and large friend circles accelerate a song's popularity, making social connections a key factor in music trends.
Using fMRI, researchers analyzed brain activity while participants experienced sustained pain and pleasure induced by capsaicin and chocolate fluids. The study identified common brain regions activated by both experiences and developed predictive models to capture affective intensity and valence information.
A trash-sorting robot has been developed that can recognize and classify objects using tactile information and machine learning algorithms. The robot achieved a classification accuracy of 98.85% in recognizing diverse garbage objects not encountered previously.
A team of researchers at Penn has developed an artificial intelligence tool that can mine the vast and largely unexplored biological data from over 10 million molecules to discover new candidates for antibiotics. The deep learning approach identified thousands of candidates in just a few hours, with many showing preclinical potential.
A new study reveals that ChatGPT's automated content moderation filters can flag nearly 20% of its own generated scripts for content violations, including half of PG-rated shows. The research raises questions about the efficacy of using AI as a tool in scriptwriting and its potential impact on artistic expression.
Researchers developed HypOp, a framework using unsupervised learning and hypergraph neural networks to solve combinatorial optimization problems significantly faster. The framework can also tackle certain problems that prior methods cannot effectively solve.
Researchers confirmed that elephant calls contained a name-like component identifying the intended recipient through machine learning analysis. Elephants responded affirmatively to calls addressed to them and less so to those meant for others, suggesting an ability to learn and use arbitrary vocal labels like humans.
Researchers at Osaka Metropolitan University developed a machine learning-based deicer that offers higher performance while minimizing environmental harm. The new mixture of propylene glycol and sodium formate solution shows improved ice penetration capacity, reducing the need for substance use.