Articles tagged with Computer Modeling
Researchers developed a new benchmark for health care using reinforcement learning, which shows promise in managing chronic or psychiatric diseases. However, current methods are data-hungry and fail to perform accurately when tested on real-world data.
A new tool developed by Penn State researchers uses computer vision and artificial intelligence to analyze placenta images, detecting abnormalities and risks such as neonatal sepsis. The PlacentaCLIP+ model has the potential to transform neonatal and maternal care in low- and high-resource settings.
Researchers from Florida Atlantic University developed an AI system that accurately detects and tracks American Sign Language alphabet gestures using computer vision. The model achieved a detection process with high accuracy, recognizing complex hand gestures with an F1 score of 99% and mAP of 98%.
The Mcity Test Facility's first open-source digital twin enables researchers to test autonomous algorithms in a virtual environment. The digital twin introduces real-world data and simulated safety-critical events, accelerating the development of connected and automated vehicle software.
A new AI tool called PlacentaVision can rapidly evaluate placentas at birth to detect abnormalities associated with infection and neonatal sepsis. This could lead to improved health outcomes for both mothers and babies, particularly in areas with limited medical resources.
Scientists have developed a new technique to study faults, which can improve earthquake forecasts by determining the origins and directions of past rupture events. By analyzing curved scratches left on the fault plane, researchers can pinpoint where earthquakes start and spread, providing valuable insights for modeling future scenarios.
Researchers at Graz University of Technology developed a new database to improve speech recognition of Austrian German using speech data from 38 speakers. They found that traditional HMM-based systems are more robust for short sentences and dialectal language, while transformer-based models excel with longer sentences and context.
A new study estimates that India will see over 62.4 million cases of tuberculosis in the next two decades, posing a significant burden on the country's health and economy. Scaling up existing treatment regimens could generate at least $28 billion in GDP gains, highlighting the need for increased investment in TB control measures.
A team of researchers has developed a novel technique to steal artificial intelligence (AI) models by monitoring electromagnetic signals. The method allows attackers to recreate the high-level features of an AI model with 99.91% accuracy, potentially undermining intellectual property rights and exposing sensitive data.
A research team led by USC aims to create comprehensive maps of retinal nerve connections to understand and combat retinitis pigmentosa, a progressive eye disease affecting 2 million people globally.
A new technique identifies and removes specific points in a training dataset that contribute most to a model's failures on minority subgroups. This approach maintains the overall accuracy of the model while improving its performance regarding underrepresented groups.
A new AI model developed at the University of Georgia predicts nearby traffic movements and incorporates innovative features for planning safe vehicle movements. This approach helps reduce crashes and near-misses by consolidating two steps: predicting surrounding traffic movements and planning a self-driving car's motion.
A new system, EXPLINGO, enables AI models to generate readable narratives explaining their predictions, helping users make better decisions. The system, developed by MIT researchers, uses large language models to transform complex explanations into plain language and automatically evaluate their quality.
A new research paper published in Oncotarget introduces an innovative AI tool combining CT scans and body composition data to predict severe liver problems in primary sclerosing cholangitis (PSC) patients. The model achieved impressive results, correctly identifying at-risk patients with 97% accuracy.
Researchers at Tokyo University of Science have developed a new method called black-box forgetting, which enables selective removal of unnecessary information from large pre-trained AI models. This approach enhances model efficiency and improves privacy by reducing computational resources and information leakage.
Researchers develop innovative hybrid control strategy to improve product yields in biosynthetic processes. The new approach combines model-based optimization with in-cell feedback control, outperforming traditional methods and promising reduced costs and environmental impact.
Evo, a generative AI model, uses patterns in microbial genomes to write new genetic code, expanding the length of sequences models can process and improving resolution. Researchers use Evo to understand microbial and viral genomes, fashion new proteins, and reprogram microbes for remarkable tasks.
Dartmouth researchers propose a new way of thinking about masking and social distancing rules using game theory, considering them as two distinct actions. The study finds that people respond differently to these measures based on their perception of the disease's severity, trending towards masking or no protective action over time.
Researchers observed microfiber plastics tumbling, rolling, and getting stuck in soil particles, revealing slow movement and potential trapping. The study improves understanding of exposure risks and possible health impacts of widespread plastic pollutants.
Researchers propose FTI-SLAM to maintain system performance while addressing critical privacy and communication concerns. It leverages federated learning, reducing data transmission and improving generalisation capabilities.
The Polymathic AI team has released two massive datasets for training artificial intelligence models to find and exploit transferable knowledge between seemingly disparate fields. The datasets include data from dozens of sources, covering astrophysics, biology, acoustics, chemistry, fluid dynamics, and more.
Scientists at MIT developed a fully integrated photonic processor that can perform all key computations of a deep neural network optically on the chip. The device completed machine-learning classification tasks in under half a nanosecond while achieving over 92% accuracy, similar to traditional hardware.
A new study reveals that children's awareness of famous computer scientists is dominated by white men, with few women and people from diverse backgrounds represented. The researchers argue that this lack of diversity limits the relevance and usefulness of computer science for underrepresented communities.
Experts from UTHealth Houston and Baylor College of Medicine developed a pragmatic approach to monitor and manage AI systems in healthcare organizations. The guidance emphasizes the need for robust governance systems, testing processes, and transparency with patients to ensure safe AI adoption.
The National Center for Supercomputing Applications (NCSA) has received the Readers' Choice Award: Best HPC Collaboration and Editors' Choice: Best Use of HPC in Physical Sciences. This is the 14th consecutive year NCSA has been honored with an HPCwire award.
A novel no-code prototype, Auto-DSM, generates design structure matrices (DSMs) using large language models. It improves productivity and accuracy compared to traditional methods, with speeds of up to four minutes per DSM generation.
Researchers at MIT have introduced a new algorithm that strategically selects the best tasks for training an AI agent, resulting in improved performance and reduced training costs. The technique outperforms existing methods by five to 50 times, making it more efficient and effective.
The winning team presented an exascale climate emulator that addresses the growing computational and storage requirements of high-resolution Earth System Models. This innovation enables more advanced climate modeling capabilities, holding significant potential for advancing climate research and policy-making.
A team of researchers led by Jeff Zacks used computer models to analyze over 25 hours of video footage of people performing everyday tasks. They found that the models were most accurate when responding to uncertainty, suggesting a more complex role for prediction errors in human cognition. Additionally, another researcher, Maverick Smi...
A novel deep learning model called Co-Plane Attention Across MRI Sequences (CoPAS) has been developed to assist with classifying 12 common types of knee abnormalities. The model achieved high classification accuracy comparable to that of radiologists, improving diagnostic efficiency and reducing errors.
Researchers developed an AI-powered model called FastGlioma that can detect residual tumor tissue with high accuracy in 10 seconds. The technology outperformed conventional methods, reducing the risk of missed tumors by nearly 75%. This innovation could change the field of neurosurgery and minimize reliance on radiographic imaging.
Researchers developed a novel machine-learning model to identify and measure causal interactions that vary over time. The Temporal Autoencoders for Causal Inference (TACI) model performed well on synthetic and real-world datasets, detecting changes in strength and direction of causal relationships.
Researchers at Lehigh University are using advanced algorithms and cross-domain data to help cities predict human movement patterns, enabling better planning and preparedness for events and emergencies. The model will account for variations in data streams from different sources, such as cell towers, GPS, and financial transactions.
A new crowdsourcing system, FireLoc, uses a network of low-cost mobile phones to detect wildfires minutes—even seconds—after they ignite. The system prioritizes privacy and accurately maps wilderness fires to within 180 feet of their origin.
Researchers developed an optimized model for defibrillation that reduces energy consumption by 1,000 times, minimizing cardiac tissue damage and pain. The new approach exploits the sensitivity of fibrillation to small changes in electrical field profiles.
Despite their ability to provide turn-by-turn driving directions in New York City with near-perfect accuracy, generative AI models do not form a coherent internal map of the city. The researchers developed new metrics to test a transformer's world model and found that only one generated a coherent world model for Othello moves.
A team of MIT engineers has developed a new computational method for analyzing complex biological systems, including the immune system's response to tuberculosis vaccination. The approach uses probabilistic graphical networks to identify key interactions and mechanisms, shedding light on how vaccines induce immunity.
A computer modeling study found that glacial isostatic adjustment caused downward movements in the eastern US, while upward movements occurred in eastern Canada, contributing to relative sea-level rise. The research will help generate maps for aquifer management and inform decisions on sea-level rise impacts.
A team of MIT engineers developed an algorithm to identify causal links in complex systems, taking data from various sources and analyzing interactions between variables. The method generates a causality map linking variables with likely cause-and-effect relationships, including synergistic and redundant links.
A new study reveals that ocean eddies off the leeward side of Hawaiian Islands supply nutrients to both sides of the island chain, stimulating blooms of phytoplankton and boosting biological productivity. This mechanism may also impact fisheries near Hawaii and other nutrient-poor regions.
Researchers developed a new index combining human comfort and social vulnerability with heat island mitigation strategies. Trees were found to provide the best relief from heat in vulnerable areas, while cool roofs and green roofs were preferred in less vulnerable neighborhoods.
A computer model uses camera pose estimation and phototourism data to reconstruct the 3D scene of Antarctic penguin colonies, allowing for their identification and tracking. This innovative approach enables researchers to study and monitor penguin populations more effectively.
Researchers developed a method to assess which patients with metastatic triple-negative breast cancer can benefit from immunotherapy. The tool uses predictive biomarkers and machine learning-based approaches to identify patients who are likely to respond to treatment.
MIT researchers developed a versatile technique that combines diverse data from various sources into a shared language for generative AI models. This approach outperformed traditional techniques by 20% in simulation and real-world experiments.
Researchers at Wayne State University are developing new AI-powered methods to design and develop new drugs, including carbohydrate-based treatments for cancer. The study aims to improve the accuracy of simulations and machine learning models to predict the behavior of complex biological molecules.
A new study used the 2023 Texas-Louisiana heat wave as a test case to establish processes for evaluating extreme weather events. The researchers found that the heat wave was directly related to climate change and will see even hotter heat waves in the future.
Researchers discovered that Titan's icy surface is warmed by an insulating layer of methane clathrate ice, which relaxes the impact craters' shape. This finding helps explain Titan's unique hydrological cycle and climate.
A team of University of Houston engineers developed an AI tool to predict and control pandemic spread by analyzing international air travel. The analysis found that reducing flights in Western Europe can lead to fewer global COVID-19 cases, making it a key strategy for controlling the pandemic.
Researchers found that restoring coastal marshes can significantly help protect coastlines at a reasonable cost. A study by MIT graduate student Ernie I. H. Lee and professor Heidi Nepf shows that enhancing salt marshes in front of protective seawalls can reduce construction costs while still providing adequate protection from storms.
A new tool, SymGen, enables users to verify AI model responses more quickly and easily by displaying data citations. This speeds up the manual validation process by 20 percent, making it easier for users to spot errors in LLMs deployed in various real-world situations.
A team of researchers developed a wearable camera system that uses artificial intelligence to detect potential medication delivery errors. The AI achieved high sensitivity and specificity in identifying vial-swap errors, making it a critical safeguard in operating rooms, intensive-care units, and emergency-medicine settings.
Researchers found that verified users with entrenched opinions can trigger echo chambers, while centrist ideologues may actually facilitate consensus. The study used a computational model to simulate how people post and receive messages on social media platforms.
A new study suggests that Betelgeuse's pulsing is due to an orbiting companion star known as the 'Betelbuddy'. The star acts like a snowplow, pushing light-blocking dust out of the way and making Betelgeuse appear brighter. Researchers used computer simulations to confirm this hypothesis, ruling out other possible causes.
A team of scientists and experts led by PNNL has developed a cloud computing approach to democratize access to emerging resources. They demonstrated that cloud computing can provide an agile complement to high-performance computing facilities, enabling complex chemistry workflows to be completed in days instead of months. The initiativ...
A UVA professor has developed a new computational algorithm to find tightly connected clusters, or triangle-dense subgraphs, within large networks. This breakthrough can help uncover suspicious activity in fraud detection and identify community dynamics on social media with greater precision.
Researchers at UVA have developed an AI-driven system that optimizes manufacturing processes, improving speed and quality while reducing waste. The system uses Multi-Agent Reinforcement Learning to coordinate tasks in real-time, leading to faster production and reduced downtime across various industries.
Researchers at UVA's School of Engineering and Applied Science have developed an AI-driven intelligent video analyzer capable of detecting human actions with unprecedented precision and intelligence. The system, called SMAST, promises to transform industries such as surveillance, healthcare, and autonomous driving.
Researchers at Chung-Ang University developed a novel GAN model, PMF-GAN, to address stability and efficiency issues. The model utilizes kernel functions and histogram transformations to improve the generator's ability to produce diverse outputs, reducing mode collapse and gradient vanishing.
A new project, 'Crowd-Assisted Human-AI Teaming with Explanations,' aims to develop an interactive AI system that leverages the collective strengths of human crowd workers and machine learning models. The researchers will use crowdsourcing platforms to recruit experts and non-experts to perform tasks, making the system more robust and ...
Researchers designed custom-fit PAP masks using computational modelling to reduce air leakage and increase comfort. The masks were tested against commercially available ones, showing promising results.