Articles tagged with Computational Science
The Argonne team developed a high-performance, iterative reconstruction system for noninvasive imaging at synchrotron facilities. Their novel optimizations enabled reconstruction of large 3D volumes in under 3 minutes using 24,576 GPUs.
Stony Brook University has installed a new supercomputer, Ookami, powered by the HPE Apollo 80 system and Fujitsu A64FX processor, offering a balance of high performance and power efficiency. The system is supported by Bright Cluster Manager software and will be available for researchers nationwide to test new computing technologies.
Researchers at Skoltech predicted high-entropy alloys with improved mechanical properties, such as hardness and fracture resistance, in the W-Mo-B system. The study aims to develop new hard materials that can withstand higher temperatures or pressures.
Eduardo Lopez Atencio and his collaborators are analyzing agent search methods in a constrained network environment to help the US Army workforce stay focused on strategic goals. This analysis will be applied to forecast workforce behaviors across an extended planning horizon.
Scientists from Nanyang Technological University (NTU Singapore) have developed an AI system that recognizes human hand gestures with high precision. The system combines skin-like electronics with computer vision and achieves accuracy even in poor environmental conditions.
Researchers found a Hopfion structure, a repeating pattern present throughout nature, in ferroelectric nanoparticles. This discovery highlights the interconnectedness of scientific fields and could lead to breakthroughs in energy storage devices and information systems.
Researchers developed Q-scores to assess true resolution at every point in cryo-EM maps, enabling accurate interpretation of atomic models. The approach validated on large molecules, achieving high-resolution maps close to 1.75 angstroms, and demonstrates improved confidence in molecular interpretations.
Gordon examines secure computation to leverage user data without compromising privacy. He investigates three key tasks: handling large volumes of data with few parties, interacting with a large number of parties, and storing dynamic data in a distributed plane.
The Tapis Project aims to simplify access to powerful supercomputers and manage data from diverse sources. The platform offers event-driven computing, streamlining workflows and enabling hands-free analysis.
Scientists developed computational models of word meanings and tested them against real-time brain activity in volunteers, revealing how the brain combines words to make sense in context. The study shows how word constraints directly affect meaning interpretation, highlighting the neural mechanisms behind spoken language understanding.
Researchers from Skoltech found that professional players move around more often and intensely than amateurs, while sitting still during game events. Machine learning methods correctly predicted a player's skill level in 77% of cases.
Computing scientists at the University of Alberta developed a methodology combining machine-learning algorithms to recognize depression using acoustic cues. The technology could lead to apps tracking mood indicators and providing support for individuals.
Tanu Malik's project aims to establish reproducibility of real-world computational and data science. Her system, Sciunit container, makes it easier to reproduce and compare scientific experiments by aligning data, programs, and environments.
Researchers at the University of Maryland introduced a new way of combining perception and motor commands using hyperdimensional computing theory, which can fundamentally alter AI tasks. This integration enables robots to fuse their perceptions with motor capabilities, creating a more efficient and faster way to complete tasks.
A new study finds that half of scientists will drop out of academia after five years, contrasting with the 1960s when a higher fraction spent their careers in academia. The study reveals a steadily growing temporary workforce and a 35-percentage-point rise in researchers never credited as primary authors.
The U.S. Department of Energy's Office of Science has awarded 62 projects under the Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program for 2019. These projects aim to tackle some of the world's most challenging science problems using cutting-edge computational methods and resources.
Kenneth Miller received the 2018 Swartz Prize in Theoretical and Computational Neuroscience for his contributions to modeling cerebral cortex function. His research has led to a reliable model of synaptic plasticity, confirmed by experimental data.
A $3 million grant will support a two-year effort to develop a strategic plan for a Cyberinfrastructure Center of Excellence at USC ISI. The project aims to enhance existing cyberinfrastructure and provide a forum for knowledge sharing among large facilities and users.
The Biophysical Society has announced the winners of its Outstanding Poster Award, recognizing outstanding scientific achievements in biophysics. The student winners were selected for their research on HERG kinetics, while the postdoctoral winner was recognized for her work on cardiac resynchronization therapy.
Two new $1 million NSF grants will support RENCI's efforts to develop a high-performance platform for scientific data processing and detect unintentional data errors. The projects aim to improve scientific productivity, increase confidence in outcomes, and facilitate seamless integration of available resources and networks.
Researchers at Tel Aviv University and Tsinghua University have discovered microscale superlubricity, achieving ultra-low friction and wear between dissimilar materials under high loads. This breakthrough has enormous implications for computer hard discs and ball bearings, leading to potential energy savings and wear prevention.
The Society for Industrial and Applied Mathematics (SIAM) has awarded 16 distinguished mathematicians for their outstanding research and contributions to the field of applied mathematics and computational science. The recipients were honored at the SIAM Annual Meeting Prizes and Awards Luncheon in Portland, Oregon.
A computational model suggests that older memories are affected in depression, with longer episodes causing more extensive memory deficits. The study's findings imply that major depressive disorder could have far-reaching consequences, including the loss of remote memories that do not recover even after the depression subsides.
The Texas Advanced Computing Center (TACC) is developing innovative software solutions to enhance scientific productivity. Their interactive parallelization tool, IPT, enables researchers to convert serial code into parallel code using tens of thousands of processors, achieving significant speed-ups and user productivity enhancements. ...
A new study on turtle brain evolution reveals that their brains have changed slowly and constantly over 210 million years, with modern turtles showing a wide variety of brain shapes and sizes. The first turtles with fully formed shells were likely living on land, contradicting previous hypotheses about their origins.
The U.S. Department of Energy's INCITE program awards large-scale research campaigns to researchers, allocating 5.95 billion core-hours on three powerful supercomputers. Domain scientists and computational experts provide support to aid in code development, optimization, and data analysis.
Scientists at Yale University and international collaborators have developed new memsistor devices that can store information and regulate electrical current. The discovery offers great potential for applications in computing, particularly in neuromorphic and logic circuits.
Lim's work impacts public through real-world applications, including neural fiber mapping and Human Connectome Project. He received multiple awards for his research on tensors, hypermatrices, and computational geometry/topology.
Chris Wilmer, assistant professor at the University of Pittsburgh, has received the AIChE Young Investigator Award for his outstanding research in computational molecular science and engineering. His work focuses on large-scale molecular simulations to find promising materials for energy and environmental applications.
GFlow is a computational software that makes wildlife habitat connectivity modeling vastly faster, more efficient and superior in quality and scope. It can solve ecological problems spanning large landscapes and predict the spread of diseases like Zika.
Researchers have developed an ultrafast technique to film the propagation of guided light and read its spatial profile across a multilayered structure. This breakthrough enables the design and control of confined plasmonic fields, crucial for future optoelectronic devices.
A new video game, Colony B, helps scientists analyze how microbes relate to lifestyle habits and health. The game correlates player data with survey information from the American Gut project, revealing potential links between microbial composition and human health.
The National Science Foundation has awarded $110 million to XSEDE to expand its services and support for advanced computational and data-enabled research. XSEDE will continue to provide services to over 6,000 scientists and engineers while adding innovative elements in response to evolving user demands.
Researchers aim to develop a recipe for solving runaway electron problems using simulations and data from worldwide experiments. The Simulation Center for Runaway Electron Avoidance and Mitigation will explore causes and solutions for relativistic runaway electrons traveling at nearly the speed of light.
Scientists with Berkeley Lab developed a way to chemically assemble transistors and circuits that are only a few atoms thick, yielding functional structures large enough for real-world applications. This breakthrough helps pave the way for scalable and repeatable atomic electronics or more computing power in smaller areas.
A computer program can now accurately retrieve product images based on hand-drawn sketches, revolutionizing online shopping. This technology enables users to sketch their desired products and instantly find them online.
Researchers at University of Oklahoma are developing a higher-resolution model that uses Stampede supercomputer to predict severe hail storms with greater accuracy. The goal is to provide 2-hour warnings for storms like the May 1995 Fort Worth Mayfest storm, which caused over $2 billion in damage.
Researchers from MIPT and MSU created a computer model predicting agrochemical activity using machine learning and Kohonen self-organizing maps. The model showed high predictive power, accurately classifying molecules into pesticides or plant growth regulators with 87% accuracy.
Terrence Sejnowski receives the Swartz Prize for his significant cumulative contribution to theoretical models and computational methods in neuroscience. He has made major research discoveries, including learning models for birdsong and neuroeconomics.
Scientists at the University of Missouri have developed a free online genetic research tool called RNAMiner. The platform allows users to upload genomic data and analyze it through multiple steps, providing results within hours rather than days or months.
A partnership between a cognitive scientist and an environmental scientist has enhanced high school students' learning of basic spatial reasoning, as well as solving complex problems involving climate change and other STEM subjects. The program uses computer-based layered maps to facilitate visualization and reasoning.
Scientists developed a smart keyboard that senses typing patterns to distinguish individual users, providing an additional layer of security for computer systems. The device harnesses energy from typing to power itself or other devices, while its surface coating repels dirt and grime.
A new tool called KnIT has shown promise in mining all public medical literature and generating hypotheses that can lead to breakthroughs in cancer research. The tool was tested on the protein p53, which is a critical tumor suppressor protein, and accurately predicted the existence of proteins that modify it.
Computer scientists at MIT and Israel's Technion have discovered an unexpected source of information about the world's languages: native speakers' writing habits in English. The system analyzed over 1,000 essays to predict typological features of 14 languages, achieving 72% accuracy on average.
A team of scientists has developed a model that uses open data and network science to evaluate the logistics of disaster relief. The proposed tool can quantify the reachability of critical loci within a geographic area and assess the effects of infrastructure destruction scenarios.
NERSC recognized four award recipients for their exceptional impact on scientific understanding and society. Victor Ovchinnikov was awarded an Early Career Award for his work on computational modeling of conformational transitions in large biological molecules.
A new wireless surveillance system, TernCam, reduces battery life without compromising image quality. The system enables real-time visual data collection and has successfully transmitted signals in severe weather conditions.
The Penn State-led team aims to improve existing hydrological predictions and study the coupling of water with its environment. The goal is to create a strategic tool to assist in decision-making for extreme weather events.
A study by Georgia Tech researchers found that people in the US and South Korea prefer wearing e-textiles on their wrists and forearms, with minimal acceptance for other locations. Cultural differences play a significant role in perceptions of wearable technology, with Americans focusing on ease of use and South Koreans prioritizing mo...
A new computational tool called ExpressionBlast enables searches based on experimental values, revealing links between diseases and treatments. Researchers used the tool to uncover clues about SIRT6, a potentially important drug target involved in immune response, metabolism, and gender-specific gene regulation.
Researchers at CU-Boulder demonstrated that brains use a pointer system similar to computers to understand novel situations. The brain's connections between the prefrontal cortex and basal ganglia play a key role in processing sentences with unfamiliar words.
Researchers at Disney Research built computer models by monitoring artists as they sketch human faces, capturing each artist's drawing style, strokes, and feature selection. The system successfully generated multiple distinct styles similar to hand-drawn sketches.
Scientists are facing a digital bottleneck in assembling genetic data, which is being produced at an unprecedented rate. Computing solutions are needed to address this issue, including clever algorithms and harnessing more computing power. The article highlights the need for innovative solutions to overcome this challenge.
Researchers used a computer model of integrin to explore its molecular machinery and interactions with the cell environment. The model suggests that integrins cluster together in response to external forces and recruit more integrins when subjected to higher forces.
The Planck observatory's first 15 months of data reveal that the universe is 100 million years older than previously thought, with more matter and less dark energy. Scientists used supercomputing at NERSC to create detailed maps of the relic radiation from the big bang.
Georgia Tech has received a $561,130 contract from DARPA to create an algorithmic framework for supercomputing systems that require less energy than traditional high-speed machines. The goal is to enable devices in the field to perform calculations previously requiring room-sized supercomputers.
Researchers connect 250 years of organic chemical knowledge into a giant computer network, optimizing syntheses of drug molecules and identifying suspicious chemical recipes. The Chematica system learns from experience and can test every possible synthesis, finding truly optimal ways to make desired chemicals.
Researchers developed an expert system to diagnose thyroid disease using computerized screening of medical data, achieving an accuracy rate of 93.5%, significantly improving detection rates.
Scientists at University of Warwick solve 40-year-old Fourier Transform Mass Spectrometry phasing problem, improving data quality for pharmaceuticals, healthcare, and environmental management. The breakthrough method doubles resolution and sensitivity at no extra cost.
The Society for Industrial and Applied Mathematics named 35 individuals as part of its 2012 Class of Fellows, recognizing their outstanding research in applied mathematics and computational science. The selected fellows are academics and professionals who have made significant contributions to the field.