Computers Math

If a virtual world has ever left you feeling nauseous or disorientated, you’re familiar with cybersickness, and you’re hardly alone. The intensity of virtual reality (VR) — whether that’s standing on the edge of a waterfall in Yosemite or engaging in tank combat with your friends — creates a stomach-churning challenge for 30-80% of users.
In a first-of-its kind study, researchers at the University of Maryland recorded VR users’ brain activity using electroencephalography (EEG) to better understand and work toward solutions to prevent cybersickness. The research was conducted by Eric Krokos, who received his Ph.D. in computer science in 2018, and Amitabh Varshney, a professor of computer science and dean of UMD’s College of Computer, Mathematical, and Natural Sciences.
Their study, “Quantifying VR cybersickness using EEG,” was recently published in the journal Virtual Reality.
The term cybersickness derives from motion sickness, but instead of physical movement, it’s the perception of movement in a virtual environment that triggers physical symptoms such as nausea and disorientation. While there are several theories about why it occurs, the lack of a systematic, quantified way of studying cybersickness has hampered progress that could help make VR accessible to a broader population.
Krokos and Varshney are among the first to use EEG — which records brain activity through sensors on the scalp — to measure and quantify cybersickness for VR users. They were able to establish a correlation between the recorded brain activity and self-reported symptoms of their participants. The work provides a new benchmark — helping cognitive psychologists, game developers and physicians as they seek to learn more about cybersickness and how to alleviate it.
“Establishing a strong correlation between cybersickness and EEG-measured brain activity is the first step toward interactively characterizing and mitigating cybersickness, and improving the VR experience for all,” Varshney said. More

An interdisciplinary team of Cornell and Harvard University researchers developed a machine learning tool to parse quantum matter and make crucial distinctions in the data, an approach that will help scientists unravel the most confounding phenomena in the subatomic realm.
The Cornell-led project’s paper, “Correlator Convolutional Neural Networks as an Interpretable Architecture for Image-like Quantum Matter Data,” published June 23 in Nature Communications. The lead author is doctoral student Cole Miles.
The Cornell team was led by Eun-Ah Kim, professor of physics in the College of Arts and Sciences, who partnered with Kilian Weinberger, associate professor of computing and information science in the Cornell Ann S. Bowers College of Computing and Information Science and director of the TRIPODS Center for Data Science for Improved Decision Making.
The collaboration with the Harvard team, led by physics professor Markus Greiner, is part of the National Science Foundation’s 10 Big Ideas initiative, “Harnessing the Data Revolution.” Their project, “Collaborative Research: Understanding Subatomic-Scale Quantum Matter Data Using Machine Learning Tools,” seeks to address fundamental questions at the frontiers of science and engineering by pairing data scientists with researchers who specialize in traditional areas of physics, chemistry and engineering.
The project’s central aim is to find ways to extract new information about quantum systems from snapshots of image-like data. To that end, they are developing machine learning tools that can identify relationships among microscopic properties in the data that otherwise would be impossible to determine at that scale.
Convolutional neural networks, a kind of machine learning often used to analyze visual imagery, scan an image with a filter to find characteristic features in the data irrespective of where they occur — a step called “convolution.” The convolution is then sent through nonlinear functions that make the convolutional neural networks learn all sorts of correlations among the features. More

When gravitational waves were first detected in 2015 by the advanced Laser Interferometer Gravitational-Wave Observatory (LIGO), they sent a ripple through the scientific community, as they confirmed another of Einstein’s theories and marked the birth of gravitational wave astronomy. Five years later, numerous gravitational wave sources have been detected, including the first observation of two colliding neutron stars in gravitational and electromagnetic waves.
As LIGO and its international partners continue to upgrade their detectors’ sensitivity to gravitational waves, they will be able to probe a larger volume of the universe, thereby making the detection of gravitational wave sources a daily occurrence. This discovery deluge will launch the era of precision astronomy that takes into consideration extrasolar messenger phenomena, including electromagnetic radiation, gravitational waves, neutrinos and cosmic rays. Realizing this goal, however, will require a radical re-thinking of existing methods used to search for and find gravitational waves.
Recently, computational scientist and lead for translational artificial intelligence (AI), Eliu Huerta of the U.S. Department of Energy’s (DOE) Argonne National Laboratory, in conjunction with collaborators from Argonne, the University of Chicago, the University of Illinois at Urbana-Champaign, NVIDIA and IBM, has developed a new production-scale AI framework that allows for accelerated, scalable and reproducible detection of gravitational waves.
This new framework indicates that AI models could be as sensitive as traditional template matching algorithms, but orders of magnitude faster. Furthermore, these AI algorithms would only require an inexpensive graphics processing unit (GPU), like those found in video gaming systems, to process advanced LIGO data faster than real time.
The AI ensemble used for this study processed an entire month — August 2017 — of advanced LIGO data in less than seven minutes, distributing the dataset over 64 NVIDIA V100 GPUs. The AI ensemble used by the team for this analysis identified all four binary black hole mergers previously identified in that dataset, and reported no misclassifications.
“As a computer scientist, what’s exciting to me about this project,” said Ian Foster, director of Argonne’s Data Science and Learning (DSL) division, “is that it shows how, with the right tools, AI methods can be integrated naturally into the workflows of scientists — allowing them to do their work faster and better — augmenting, not replacing, human intelligence.”
Bringing disparate resources to bear, this interdisciplinary and multi-institutional team of collaborators has published a paper in Nature Astronomy showcasing a data-driven approach that combines the team’s collective supercomputing resources to enable reproducible, accelerated, AI-driven gravitational wave detection.
“In this study, we’ve used the combined power of AI and supercomputing to help solve timely and relevant big-data experiments. We are now making AI studies fully reproducible, not merely ascertaining whether AI may provide a novel solution to grand challenges,” Huerta said.
Building upon the interdisciplinary nature of this project, the team looks forward to new applications of this data-driven framework beyond big-data challenges in physics.
“This work highlights the significant value of data infrastructure to the scientific community,” said Ben Blaiszik, a research scientist at Argonne and the University of Chicago. “The long-term investments that have been made by DOE, the National Science Foundation (NSF), the National Institutes of Standards and Technology and others have created a set of building blocks. It is possible for us to bring these building blocks together in new and exciting ways to scale this analysis and to help deliver these capabilities to others in the future.”
Huerta and his research team developed their new framework through the support of the NSF, Argonne’s Laboratory Directed Research and Development (LDRD) program and DOE’s Innovative and Novel Computational Impact on Theory and Experiment (INCITE) program.
“These NSF investments contain original, innovative ideas that hold significant promise of transforming the way scientific data arriving in fast streams are processed. The planned activities are bringing accelerated and heterogeneous computing technology to many scientific communities of practice,” said Manish Parashar, director of the Office of Advanced Cyberinfrastructure at NSF. More

As the COVID-19 pandemic wanes in the U.S., a new study from the University of Maryland School of Medicine (UMSOM) and University of Maryland Medical Center (UMMC) finds that hospitals nationwide may not be adequately prepared for the next pandemic. A 10-year analysis of hospitals’ preparedness for pandemics and other mass casualty events found only marginal improvements in a measurement to assess preparedness during the years leading up to the COVID-19 pandemic. The study was published last month in the Journal of Healthcare Management.
“Our work links objective healthcare data to a hospital score that assesses the ability to save lives in a disaster,” said study lead author David Marcozzi, MD, Professor of Emergency Medicine at UMSOM and Chief Clinical Officer/Senior Vice President at UMMC. “It attempts to fill a glaring gap in the national conversation on the need for improved assessments of and the opportunity for better hospital planning to assure readiness.”
To conduct the research, Dr. Marcozzi, who is also the COVID-19 Incident Commander for the University of Maryland Medical System, and his colleagues first developed and published a surge index tool that linked standard reported hospital information to healthcare preparedness elements. The tool, called the Hospital Medical Surge Preparedness Index (HMSPI), used data from 2005 to 2014 to produce a score designed to predict how well a hospital can handle a sudden influx in patients due to a mass shooting or infectious disease outbreak. Such data included the size of the medical staff, the number of hospital beds, and the amount of equipment and supplies.
Medical surge capacity is an important measure to assess a hospital’s ability to expand quickly beyond normal services to meet an increased demand for healthcare. The Las Vegas mass shooting in 2017, for example, sent more than 500 concertgoers to local hospitals. During the early weeks of the COVID-19 pandemic, New York City hospitals were under siege with 4,000 patients hospitalized. To calculate the HMSPI, researchers input data from four important metrics: Staff: Doctors, nurses, pharmacists, respiratory technicians and others Supplies: Personal protective equipment, cardiac monitors, sterile bandages, and ventilators Space: Total beds and number of beds that current staff can handle Systems: Framework for enabling electronic sharing of files and information between departments and multiple hospitalsIn the new study, Dr. Marcozzi and his colleagues used data from the American Hospital Association’s annual surveys of more than 6,200 hospitals nationwide that were collected from 2005 to 2014. They also employed data from the U.S. Census Bureau to determine population estimates in cities and the Dartmouth Atlas Project to establish the geographic service area of each hospital. They combined the hospital metrics gleaned from the AHA’s annual surveys with the geographic data to calculate HMSPI composite scores for hospitals in each state.
Their evaluation found varying levels of increases in HMSPI scores from 2005 to 2014 in every state, which could indicate that states are becoming better prepared to handle a medical surge. The scores also indicated that ideal readiness had not yet been achieved in any state before the COVID-19 pandemic.
“This is just the starting point. We need to better understand the ability of our nation’s hospitals to save lives in times of crisis,” said Dr. Marcozzi. This information, and follow-up studies building from this work, will be key to better matching states’ healthcare resources to their population to assure optimal care is delivered. Dr. Marcozzi described one follow-up study that would be impactful would be to use data from the COVID-19 pandemic to see whether the index was predictive to indicate which hospitals were most prepared for the pandemic surge based on their patient outcomes.
“This pioneering work is a needed advancement that could allow for a transparent assessment of a hospital’s ability to save lives in a large-scale emergency,” Dr. Marcozzi said. “The COVID-19 pandemic demonstrated that there is still plenty of room for improvement in the ability of our nation’s healthcare system to triage and manage multiple patients in a crisis and that translates into lives lost, unnecessarily. Our research is dedicated to those who lost their lives in this tragedy and other mass casualty events. We can do better.”
National health leadership organizations, such as the U.S. Centers for Medicare and Medicaid Services, the Assistant Secretary for Preparedness and Response, the Joint Commission, and the American Medical Association, as well as state and local emergency planners, could all potentially benefit from the use of HMSPI scores, according to Dr. Marcozzi. The tool could be used to support data-driven policy development and resource allocation to close gaps and assure that individuals get the care they need, when then need it, during a crisis.
Ricardo Pietrobon, MD, PhD, MBA, Adjunct Associate Professor of Emergency Medicine at UMSOM, Nicole Baehr, Manager of Operations at UMMC, and Brian J. Browne, MD, Professor and Chair of the Department of Emergency Medicine, were co-authors on this study. Researchers from the University of Nebraska Medical Center, University of Miami, and the U.S. Department of Veterans Affairs also participated in this research. The study was funded by the Bipartisan Commission on Biodefense.
“The COVID-19 pandemic taught us that we need to be better prepared for the unexpected crisis,” said E. Albert Reece, MD, PhD, MBA, Executive Vice President for Medical Affairs, UM Baltimore, and the John Z. and Akiko K. Bowers Distinguished Professor and Dean, University of Maryland School of Medicine. “Having an important metric like the HMSPI could be a game changer that ultimately saves lives during a surge by helping hospitals identify and fix their vulnerabilities. More

In the United States, low-income and minority students are completing college at low rates compared to higher-income and majority peers — a detriment to reducing economic inequality. Double-dose algebra could be a solution, according to a new study published in roceedings of the National Academy of Sciences of the United States of America (PNAS).
The paper, “Effects of Double-Dose Algebra on College Persistence and Degree Attainment,” is the culmination of a series of studies that followed two cohorts of ninth-grade students over a period of 12 years in the Chicago Public Schools (CPS) where double-dose algebra was introduced in 2003.
The new policy required incoming ninth graders with eighth-grade math scores below the national median to complete two periods of math — one period of algebra, plus an additional period of instruction designed to build foundational prealgebra skills. Research findings showed that, for median-skill students scoring at or above the 50th percentile in the 2003 cohort, double-dose algebra significantly increased semesters of college attended and college degree attainment.
“This provides unique insight for districts that provide extra instruction but are unable to rigorously study the impact of those programs,” said Takako Nomi, Ph.D., associate professor of educational studies at Saint Louis University. Her work focuses on educational policy and equity.
Nomi, who also serves as research affiliate at the University of Chicago’s Consortium on Chicago School Research, led the study. Other authors include Stephen W. Raudenbush, Ed.D., of the department of sociology at the University of Chicago; and Jake J. Smith, of Harris School of Public Policy at the University of Chicago.
A key takeaway from the study is how schools chose to implement the policy matters, Nomi said. Fewer schools adopted the cut-score-based double-dose algebra program in 2004 than in 2003. Most schools that did strongly comply in 2004, did so by placing their median-skill double-dose students in low-skill algebra classrooms, according to the study.
In terms of classroom peer composition, “the impact was largest when schools didn’t group double-dose students with low-skilled students,” Nomi said. Research findings demonstrate that when students were placed in double-dose classes with much lower-skilled peers, the program had no effect. Subsequent research should address the design of optimal policies for lower-skill students, Nomi said. A math intervention far more intensive than double-dose algebra is essential to improve their high school and postsecondary outcomes. The study also notes that ninth-grade students who fail math also tend to fail other core classes.
“It’s not just a math issue,” Nomi said. “The policy of giving extra math is not enough to change the trajectory for the students who struggle the most. It’s important to support struggling students in general.”
This study was supported by grant R305A170602 from the Institute of Education Sciences entitled, “Doubling Up? Understanding the Long-Term Effects of Ninth-Grade Algebra Reform on College Persistence and Graduation.”
Nomi’s research interests include urban education, education policy, inequality in education, school reforms, and college readiness. Nomi is associate director of the Sinquefield Center for Applied Economic Research where she collaborates with top SLU researchers. In a separate study, she’s exploring why low-income and minority students — particularly Black males — are less likely to complete college. She is also a part of a faculty advisory board at SLU’s Geospatial Institute.
Story Source:
Materials provided by Saint Louis University. Original written by Bridjes O’Neil. Note: Content may be edited for style and length. More

Colorado researchers have published new findings in Emerging Infectious Diseases that take a first look at the use of SARS-CoV-2 mathematical modeling to inform early statewide policies enacted to reduce the spread of the Coronavirus pandemic in Colorado. Among other findings, the authors estimate that 97 percent of potential hospitalizations across the state in the early months of the pandemic were avoided as a result of social distancing and other transmission-reducing activities such as mask wearing and social isolation of symptomatic individuals.
The modeling team was led by faculty and researchers in the Colorado School of Public Health and involved experts from the University of Colorado Anschutz Medical Campus, University of Colorado Denver, University of Colorado Boulder, and Colorado State University.
“One of the defining characteristics of the COVID-19 pandemic was the need for rapid response in the face of imperfect and incomplete information,” said the authors. “Mathematical models of infectious disease transmission can be used in real-time to estimate parameters, such as the effective reproductive number (Re) and the efficacy of current and future intervention measures, and to provide time-sensitive data to policymakers.”
The new paper describes the development of such a model, in close collaboration with the Colorado Department of Health and Environment and the Colorado Governor’s office to gage the impact of early policies to decrease social contacts and, later, the impact of gradual relaxation of Stay-at-Home orders. The authors note that preparing for hospital intensive care unit (ICU) loads or capacity limits was a critical decision-making issue.
The Colorado COVID-19 Modeling team developed a susceptible-exposed-infected-recovered (SEIR) model calibrated to Colorado COVID-19 case and hospitalization data to estimate changes in the contact rate and the Re after emergence of SARS-CoV-2 and the implementation of statewide COVID-19 control policies in Colorado. The modeling team supplemented model estimates with an analysis of mobility by using mobile device location data. Estimates were generated in near real time, at multiple time-points, with a rapidly evolving understanding of SARS-CoV-2. At each time point, the authors generated projections of the possible course of the outbreak under an array of intervention scenarios. Findings were regularly provided to key Colorado decision-makers.
“Real-time estimation of contact reduction enabled us to respond to urgent requests to actively inform rapidly changing public health policy amidst a pandemic. In early stages, the urgent need was to flatten the curve,” note the authors. “Once infections began to decrease, there was interest in the degree of increased social contact that could be tolerated as the economy reopened without leading to overwhelmed hospitals.”
“Although our analysis is specific to Colorado, our experience highlights the need for locally calibrated transmission models to inform public health preparedness and policymaking, along with ongoing analyses of the impact of policies to slow the spread of SARS-CoV-2,” said Andrea Buchwald, PhD, lead author from the Colorado School of Public Health at CU Anschutz. “We present this material not as a final estimate of the impact of social distancing policies, but to illustrate how models can be constructed and adapted in real-time to inform critical policy questions.”
Story Source:
Materials provided by University of Colorado Anschutz Medical Campus. Original written by Tonya Ewers. Note: Content may be edited for style and length. More

A team led by researchers from Queen Mary University of London has developed a new artificial intelligence (AI) tool that is able to automatically measure the amount of fat around the heart from MRI scan images.
Using the new tool, the team was able to show that a larger amount of fat around the heart is associated with significantly greater odds of diabetes, independent of a person’s age, sex, and body mass index.
The research is published in the journal Frontiers in Cardiovascular Medicine and is the result of funding from the CAP-AI programme, which is led by Barts Life Sciences, a research and innovation partnership between Queen Mary University of London and Barts Health NHS Trust.
The distribution of fat in the body can influence a person’s risk of developing various diseases. The commonly used measure of body mass index (BMI) mostly reflects fat accumulation under the skin, rather than around the internal organs. In particular, there are suggestions that fat accumulation around the heart may be a predictor of heart disease, and has been linked to a range of conditions, including atrial fibrillation, diabetes, and coronary artery disease.
Lead researcher Dr Zahra Raisi-Estabragh from Queen Mary University of London said: “Unfortunately, manual measurement of the amount of fat around the heart is challenging and time-consuming. For this reason, to date, no-one has been able to investigate this thoroughly in studies of large groups of people.
“To address this problem, we’ve invented an AI tool that can be applied to standard heart MRI scans to obtain a measure of the fat around the heart automatically and quickly, in under three seconds. This tool can be used by future researchers to discover more about the links between the fat around the heart and disease risk, but also potentially in the future, as part of a patient’s standard care in hospital.”
The research team tested the AI algorithm’s ability to interpret images from heart MRI scans of more than 45,000 people, including participants in the UK Biobank, a database of health information from over half a million participants from across the UK. The team found that the AI tool was accurately able to determine the amount of fat around the heart in those images, and it was also able to calculate a patient’s risk of diabetes.
Dr Andrew Bard from Queen Mary University of London, who led the technical development, added: “The AI tool also includes an in-built method for calculating uncertainty of its own results, so you could say it has an impressive ability to mark its own homework.”
Professor Steffen Petersen from Queen Mary University of London, who supervised the project, said: “This novel tool has high utility for future research and, if clinical utility is demonstrated, may be applied in clinical practice to improve patient care. This work highlights the value of cross-disciplinary collaborations in medical research, particularly within cardiovascular imaging.”
Story Source:
Materials provided by Queen Mary University of London. Note: Content may be edited for style and length. More

Using artificial intelligence, UT Southwestern scientists have identified thousands of genetic mutations likely to affect the immune system in mice. The work is part of one Nobel laureate’s quest to find virtually all such variations in mammals.
“This study identifies 101 novel gene candidates with greater than 95% chance of being required for immunity,” says Bruce Beutler, M.D., director of the Center for the Genetics of Host Defense (CGHD) and corresponding author of the study published this week in the Proceedings of the National Academy of Sciences. “Many of these candidates we have already verified by re-creating the mutations or knocking out the genes.” Lead author Darui Xu, a computational biologist at CGHD, wrote the software.
“We’ve developed software called Candidate Explorer (CE) that uses a machine-learning algorithm to identify chemically induced mutations likely to cause traits related to immunity. The software determines the probability that any mutation we’ve induced will be verified as causative after further testing,” Beutler says. His discovery of an important family of receptors that allow mammals to quickly sense infection and trigger an inflammatory response led to the 2011 Nobel Prize in Physiology or Medicine.
“The purpose of CE is to help researchers predict whether a mutation associated with a phenotype (trait or function) is a truly causative mutation. CE has already helped us to identify hundreds of genes with novel functions in immunity. This will improve our understanding of the immune system so that we can find new ways to keep it robust, and also know the reason it sometimes falters,” says Beutler, Regental Professor, and professor of immunology and internal medicine at UT Southwestern.
“CE provides a score that tells us the likelihood that a particular mutation-phenotype association will be verified for cause and effect if we re-create the mutation or knock out the gene,” he says.
CE examines 67 features of the primary genetic mapping data to arrive at an estimate of the likelihood of causation. For some mutations, causation is very clear; for others, less so. Over time, the program “learns” from experiments in which researchers re-create the mutation in a fresh pedigree and verify or exclude the hypothesis of causation. All mutations are made available to the scientific community through a public repository, and the data supporting causation are viewable within the Candidate Explorer program on the CGHD website, Mutagenetix (https://mutagenetix.utsouthwestern.edu/). More