Articles tagged with Computer Modeling
Researchers at Duke University have developed a machine learning algorithm that incorporates known physics into neural networks, allowing for new insights into material properties and more efficient predictions. The approach helps the algorithm attain transparency and accuracy, even with limited training data.
The European drought event from 2018 to 2020 was the most intense in over 250 years, affecting approximately one third of the land area. The drought's total duration was unusually long, starting in April 2018 and ending in December 2020.
A team of University of Illinois researchers, led by Charles Gammie, has captured the first direct visual evidence of a supermassive black hole at the center of the Milky Way. The image reveals a dark central region surrounded by a bright ringlike structure, providing valuable clues about the workings of such giants.
Researchers developed an algorithm to identify users' basic needs from their Instagram posts, images, and captions. The study analyzed 86 profiles in Spanish and Persian, achieving promising accuracy and complementary information between visual and textual cues.
Xie aims to develop algorithms that reflect human domain knowledge and reasoning patterns, increasing trust in machine learning models. He also explores algorithmic fairness, considering multiple perspectives on what's considered fair, to address concerns about biased decision-making.
A study by Cornell University researchers found that online retail images of models have statistically lighter skin tones than videos of the same product and model. The discrepancy was also linked to 'tokenism,' where one model with darker skin is used to represent a range of diversity.
The study found that even if Earth's climate stopped warming, it would be difficult to rebuild the ice shelf once it has fallen apart. The researchers suggest that the ice shelf may not recover unless the future climate cools considerably. This has significant implications for sea-level rise and the stability of polar ice sheets.
Researchers aim to overcome physical limitations of telecommunications by creating immediate response regardless of distance. The eTouch project leverages Model-Mediated Teleoperation and Edge Computing to enable haptic feedback perception without noticeable delay.
A team of international researchers created digital models of disease interactions to identify key proteins and signaling cascades in seasonal allergies. They found that inhibiting PDGF-BB protein was more effective than existing treatments, suggesting a potential breakthrough in personalized medicine.
A new method reduces computational complexity of traffic models, making them operate more efficiently. The modified algorithm breaks down complex forecasting questions into smaller problems that can be solved in parallel, significantly reducing run time. This approach also allows for a good enough solution within an error bar, rather t...
MIT researchers develop ExSum, a framework to formalize explanations of machine-learning models into quantifiable rules. This allows for testing assumptions about model behavior and reveals unexpected insights, such as negative words having sharper contributions to model decisions.
A team of physicists and chemists at the University of Surrey used computer modeling to show that quantum mechanics can cause errors in DNA replication, leading to mutations. The researchers found that protons can tunnel through energy barriers, causing mistakes in the pairing of DNA bases.
A new model, GCM, uses graph neural networks to capture user behavior interactions and factorization machine to model feature interactions, demonstrating effectiveness on public datasets. The research advances context-aware recommender systems by leveraging graph learning for strong user and item representations.
A severe rainstorm over Kauai in April 2018 resulted in $180 million in damage and 532 homes destroyed. The study found that supercell thunderstorms triggered the deluge, which set a new US 24-hour rainfall record of nearly 50 inches.
The study suggests that Venus' atmosphere plays a crucial role in determining its rotation speed, with fast winds dragging along the surface and slowing it down. This has significant consequences for the sweltering Venusian climate, with average temperatures of up to 900 degrees Fahrenheit.
A team of NYU Tandon researchers investigated the variability of mortality prediction models when applied to different hospitals and geographic regions. They found that these models exhibit a lack of generalizability due to dataset shifts in race and clinical variables, leading to disparities in performance across racial groups.
Scientists develop models that complement simulations using reinforcement learning and numerical methods to predict climate change, turbulent flows, and morphogenesis. This approach enables faster and more energy-efficient predictions, solving complex problems in engineering and climate applications.
Researchers developed a novel framework to characterize weakly chaotic dynamics in complex systems with many constituent parts. By investigating Lyapunov spectrum scaling, they identified emerging quasi-conserved quantities that shed light on quantum computation and physical models.
A new study provides evidence that long-term warming of the Amundsen Sea, a key contributor to global sea level rise, is linked to rising greenhouse gases. The research suggests that ocean temperatures in the region have been rising since before records began and are expected to continue if greenhouse gas emissions increase.
Researchers found that modeling studies have been assuming colonoscopies are more accurate than they really are, detecting fewer small adenomas. However, colonoscopy still appears effective in finding precursors of cancer.
Dystonia is characterized by involuntary movements and postures, limiting daily activities. A new study maps specific brain networks for treatment success in patients with cervical and generalized dystonia. The findings reveal distinct stimulation sites depending on the type of dystonia, offering a more targeted approach to improving t...
A $150,000 grant from the NV Energy Foundation will support DRI's development of a Weather and Research Forecast model to simulate weather, fire, and smoke for firefighting operations. The tool will provide critical air quality forecasts and risk assessment for specific locations.
Researchers have developed a model to predict Vibrio vulnificus abundance in the canal by analyzing rainfall, water temperature, dissolved nutrients and organic matter. The study found that warmer waters due to climate change may lead to an increase of twice or three times current levels of bacteria by the end of the century.
Impact craters reveal insights into planetary bodies' evolution, structure, and composition. By studying impact craters on various planets and asteroids, researchers can determine the formation timeline of these celestial bodies and gain knowledge about their interiors.
Researchers have created the Thesan simulation, a cubic volume spanning 300 million light years across, to study cosmic reionization and galaxy formation. The simulation aligns with observations and sheds light on key processes, such as how far light can travel in the early universe.
Researchers propose that chaos terrains on Europa's surface could shuttle oxygen to the moon's subsurface ocean, where it could sustain life. The computer model showed that 86% of oxygen is transported through the ice via a 'porosity wave', potentially raising hopes for finding life in Europa's ocean.
Researchers at Ural Federal University have developed a method to significantly accelerate the synthesis of aluminum-based alloys using computer modeling. The new approach allows for control over the internal structure and physical properties of the alloy, enabling the creation of materials with desired characteristics.
Researchers developed a new reading method for SOT-RAMs that can nullify the readout disturbance, reducing it by at least 10 times. The method involves creating a bi-directional read path, cancelling out the disturbances produced by spin currents.
A new predictive model, LAMAP, assesses landscape variables to identify potential archaeological sites. The model was tested in Alaska's Tanana Valley and successfully predicted high-potential areas for hunter-gatherer campsites.
A recent study found that large US firms have lower product diversity, with companies manufacturing similar products becoming more common. This decrease in product diversity could be due to various factors such as narrowing consumer demand and data biases in the analysis.
Scientists discovered that earthquakes influence tectonic plate movement, altering frequency and patterns of quakes. This finding suggests improved earthquake risk models can be developed by incorporating feedback mechanisms after an earthquake.
A new virtual ligand-assisted screening method enables efficient searching of ligands for optimal catalysis, accelerating reaction design. By simplifying ligand description with two metrics, researchers identify promising ligands and focus on real testing.
MIT engineers mapped airplane contrails over the US in 2020 and found a 20% drop in coverage compared to prepandemic years. The team's computer-vision technique can help predict where contrails form, allowing airlines to reroute planes and reduce aviation's climate impact.
A recent study found that sleep bruxism can lead to temporomandibular joint disorders due to increased mechanical loading. The research discovered that specific combinations of tooth shape and location during grinding significantly impact the risk of TMJ problems.
Researchers at MIT and Harvard University applied cognitive science theories to human-robot interaction, finding that humans need to see variation in robot behavior to build accurate mental models. Theories suggest that strategic variation can reveal concepts that might be difficult for a person to discern otherwise.
Adversarially robust models capture aspects of human peripheral processing, with results showing similarity in image transformations and perception alignment. The study's findings shed light on the goals of peripheral processing in humans and could help improve machine learning models.
Computer simulations reveal metformin's effectiveness in targeting diabetes and some cancers, but also its negative consequences for others. Researchers emphasize the need for precision medicine to individualize treatment based on patient profiles.
A study using machine learning identifies climatic thresholds driving vegetation distribution, highlighting the importance of extreme climate conditions for savannas and deciduous forests. The findings provide valuable insights for improving process-based vegetation models and their coupling with Earth System Models.
Researchers at University of Texas at Austin create first-ever biologically authentic computer model of HIV-1 virus liposome, shedding light on replication and infectivity. The study reveals key characteristics of the liposome's asymmetry and its role in shaping macroscopic properties.
Researchers have replicated and expanded on previous work to show that tics associated with Tourette syndrome have a fractal pattern, which can predict disease severity. This discovery could lead to a diagnostic tool for doctors to analyze tic patterns and diagnose patients with Tourette's
Researchers used AI to predict flood damage in the US, finding a high probability of flood damage for more than 1.01 million square miles across the country. The study suggests that recent FEMA maps do not capture the full extent of flood risk, with 84.5% of reported damage not within high-risk flood areas.
The study of MUNC long non-coding RNA reveals the importance of experimentally determining its structure to identify functional domains. The researchers found that two structural domains, including six common 'hairpins,' were crucial for regulating gene expression and muscle cell differentiation.
A new study by the University of Edinburgh suggests that social media bots pose less of a threat to spreading harmful messages and misinformation. Bots were found to have very limited relation to users' stance on different topics, making up less than 10% of accounts affecting users' views.
A research team led by Professor Àlex Arenas from Universitat Rovira i Virgili has discovered a theoretical explanation for the degradation of neural networks, allowing for predictions on when they will stop functioning. This breakthrough can help in understanding neurological diseases and optimizing network systems.
A new study examines the toxic impact of bio-based substances and innovative technologies on the environment, finding both biosurfactants and microgels as highly promising candidates for use in sustainable products. The results highlight the importance of integrating green toxicology into bioeconomy strategies.
Researchers adapted cancer models to predict effective drug combinations for treating Covid-19 at different stages of the disease. The study identified existing therapeutics that might be suitable for treating severe cases, and could help lower Covid-19 related deaths.
KAUST researchers simulate microgrid cyberattacks to assess impact and develop detection methods. Effective methods identify anomalous conditions associated with attacks, enabling swift isolation of affected subsystems.
A new study published in The Cryosphere finds that warm seawater intrusion under glaciers may cause much higher rates of melting at the glacier bottom. This could lead to projected ice sheet volume loss being 10-50% higher, or more than doubling over the next century.
Researchers developed a computational model that simulates COVID-19 transmission and forecasts case numbers & hospitalizations. The model incorporates various factors such as vaccination rates, mask use, lockdowns, and breakthrough infections to predict the spread of variants.
A new $1 million project at UCF aims to understand how raindrops interact with hypersonic shock waves. Researchers will use computer simulations and experiments to predict conditions for safe hypersonic travel. The knowledge gained could prevent damage and improve rocket launch accuracy.
Researchers developed an algorithm that incorporates customer behavior into recommendation systems, making more accurate and personalized suggestions. The technique, known as tensor decomposition, analyzes data in multiple dimensions to capture complex patterns and relationships.
Researchers created three computer models to optimize housing and service provider locations, ensuring former offenders can access necessary services efficiently. The tool addresses limited public transportation options, a major barrier to successful reentry.
Research at Rensselaer Polytechnic Institute finds that over 90% of people who use online delivery services would likely revert back to their original way of shopping. Delivery service users access products in four categories: groceries, food, home goods, and other items, with grocery deliveries having the highest proportion of new ado...
Researchers at Harvard SEAS developed a new way to simulate tens of thousands of bubbles in foamy flows. This allows for predictive simulations in scales ranging from microfluidics to crashing waves, opening up possibilities for industrial applications such as food production and drug development.
Researchers found that only certain types of planets can form large moons in respect to their host planets. They propose that smaller planets are better candidates to host fractionally large moons. This study provides constraints for future observations and sheds light on the formation of Earth's unique moon.
Researchers from UOC-led OptimalSharing@SmartCities project will analyze inhabitants' mobility patterns and demands to design more efficient shared transport practices. The project aims to develop agile optimization algorithms capable of processing large volumes of data in real-time for dynamic system coordination.
USC researchers develop highly accurate model to predict Covid-19 risk by combining anonymized cellphone location data with mobility patterns. The system shows a 50% improvement in accuracy compared to current systems, enabling targeted policy decisions at neighborhood or zip code levels.
The MIT team developed a computer model that can perform sound localization tasks as well as humans, and adapts to real-world environments. The model uses convolutional neural networks and was trained on over 400 sounds, including human voices and animal sounds.
A UOC study found that non-code tasks are highly significant in open-source systems, complementing programming work. Researchers analyzed 100 npm projects to gain a deeper understanding of collaboration dynamics.
A team of NTU Singapore scientists has developed a predictive computer programme using wearable technology data to detect depression risk. The programme, Ycogni model, achieves an accuracy of 80% in detecting individuals at high or low risk of depression.