HiPACC Data Science Press Room. From: UCI

The Data Science Press Room highlights computational and data science news in all fields *outside of astronomy* in the UC campuses and DOE laboratories comprising the UC-HiPACC consortium. The wording of the short summaries on this page is based on wording in the individual releases or on the summaries on the press release page of the original source. Images are also from the original sources except as stated. Press releases below appear in reverse chronological order (most recent first).

September 2, 2014 — Sierra Nevada freshwater runoff could drop 26 percent by 2100, UC study finds

Warm climate = thirsty plants = less water runoff
The Sierra Nevada snowpack runoff will diminish as a warmer climate encourages more plant growth at higher temperatures, a UC Irvine and UC Merced study has determined. Credit: Matt Meadows /UC Merced
UCI/UCM 9/2/2014—By 2100, communities dependent on freshwater from mountain-fed rivers could see significantly less water, according to a new climate model recently released by researchers at UC Irvine and UC Merced. As the climate warms, higher elevations that are usually snow-dominated see milder temperatures; plants that normally go dormant during the winter snows remaining active longer, absorbing and evaporating more water, reducing projected runoff. Using water-vapor emission rates and remote-sensing data, the authors determined relationships between elevation, climate and envirotranspiration. Greater vegetation density at higher elevations in the Kings basin with the 4.1 degrees Celsius warming projected by climate models for 2100 could boost basin evapotranspiration by as much as 28 percent, with a corresponding 26 percent decrease in river flow. The study findings appear in Proceedings of the National Academy of Sciences. Scientists have recognized for a while that something like this was possible, but no one had been able to quantify whether it could be an effect big enough to concern California water managers.

UCM release
UCI Release

August 26, 2014 — Existing power plants will spew 300 billion more tons of carbon dioxide during use

Committed to global warming: 300 gigatons of CO2
A coal-burning power plant at the Turceni Power Station in Romania. Credit: Robert and Mihaela Vicol
UCI 8/26/2014—Existing power plants around the world will pump out more than 300 billion tons of carbon dioxide over their expected lifetimes, significantly adding to atmospheric levels of the climate-warming gas, according to UC Irvine and Princeton University scientists. Using a new mathematical technique called commitment accounting, their study is the first to quantify how quickly these “committed” emissions are growing—by about 4 percent per year—as more fossil fuel-burning power plants are built. “These facts are not well known in the energy policy community, where annual emissions receive far more attention than future emissions related to new capital investments,” the paper states. The study was published in the August 26 issue of the journal Environmental Research Letters.

View UCI Data Science Press Release

July 18, 2014 — Pair awarded NSF grant to study ‘crowdprogramming’

UCI 7/18/2014—Informatics professor André van der Hoek and postdoctoral scholar Thomas LaToza have received a four-year, $1.4 million grant from the National Science Foundation for their research into what they call “crowdprogramming.” Crowdsourcing leverages the power of mass input by individuals to complete tasks that were previously too labor-intensive to be feasible; van der Hoek and LaToza propose applying those same principles to software development. The pair will explore whether crowdprogramming can be achieved and, if so, in what form, under what conditions, and with what benefits and drawbacks. They will also create a publicly available platform, CrowdCode, that will offer a tool set specifically designed to address the intricacies of crowdprogramming.

View UCI Data Science Press Release

July 2, 2014 — ‘Deep learning’ makes search for exotic particles easier

UCI, 7/2/2014— Fully automated “deep learning” by computers greatly improves the odds of discovering particles such as the Higgs boson, beating even veteran physicists’ abilities, according to findings by UC Irvine researchers published today in the journal Nature Communications. Machine learning is a branch of computer science where, rather than computers being programmed to do a difficult task, computers learn automatically from examples. Currently, physicists devise by hand mathematical formulas that they apply to the data to derive the features they’re looking for, which are then fed to machine learning programs. However, by employing recent advances in deep learning, in which computers learn automatically at multiple processing levels, the UCI team eliminated the need for the time-consuming manual creation of those formulas in the search for these fleeting particles. In computer experiments using carefully structured simulated data, the UCI researchers’ methods resulted in a statistically significant 8 percent increase in the detection of these particles. Fully automated deep learning techniques could be employed in experiments scheduled for 2015 at the Large Hadron Collider.

View UCI Data Science Press Release

May 19, 2014 — Greenland will be far greater contributor to sea rise than expected

Greenland ice far deeper threat to sea level rise
A glacier in the Sukkertoppen ice cap in southwest Greenland flows down a rocky canyon like those mapped in a new UCI-NASA study. Hundreds of previously unknown coastal canyons buried under the ice could contribute to far higher sea level rise than previously predicted. Credit: Michael Studinger/NASA
UCI 5/19/14—Greenland’s icy reaches are far more vulnerable to warm ocean waters from climate change than had been thought, according to new research by UC Irvine and NASA glaciologists. The work, published today in Nature Geoscience, shows previously uncharted deep valleys stretching for dozens of miles under the Greenland Ice Sheet. To obtain the results, UC Irvine associate project scientist Mathieu Morlighem developed a breakthrough method that for the first time offers a comprehensive view of Greenland’s entire periphery. To reveal the full subterranean landscape, he designed a novel “mass conservation algorithm” that combined the previous ice thickness measurements with information on the velocity and direction of its movement and estimates of snowfall and surface melt. The difference was dramatic. What appeared to be shallow glaciers at the very edges of Greenland are actually long, deep fingers stretching more than 100 kilometers (almost 65 miles) inland. “We anticipate that these results will have a profound and transforming impact on computer models of ice sheet evolution in Greenland in a warming climate,” the researchers conclude.

View UCI Data Science Press Release

May 12, 2014 — Fighting off virtual attacks

Fighting off virtual attacks
UC Irvine computer science professor Michael Franz has devised a way to individualize software programs to help keep hackers from inflicting widespread damage. Credit: Steve Zylius/UC Irvine
UCI 5/12/14—Imagine a cyber world in which hackers, identity thieves, spammers, phishers, foreign spies and other miscreants have a much tougher time plying their trade. Thanks to UC Irvine computer science professor Michael Franz and his research group, such a world is closer to a reality. Franz, director of UC Irvine’s Secure Systems & Software Laboratory, is borrowing the idea of “biodiversity” from nature and applying it to the software that runs on digital devices from smartphones to supercomputers. His promising ideas have already won a U.S. patent and he has been awarded more than $11 million as a principal investigator for UC Irvine—including more than $7 million as sole principal investigator—from the Defense Advanced Research Projects Agency, the U.S. intelligence community, the Department of Homeland Security, and other funding entities.

View UCI Data Science Press Release

May 9, 2014 — Encouraging cross-campus collaboration

New Initiative for Data Science at UC Irvine
The Water UCI initiative will harness the brainpower of university experts, government entities, public agencies, the private sector, and user groups to identify effective solutions to global water issues. Credit: Hoang Xuan Pham / UC Irvine
UCI 5/9/14—Multidisciplinary projects focused on data science, medical humanities and water will be funded under new Interschool Academic Initiative program, introduced in fall 2013 to identify, develop and support areas of multidisciplinary excellence. Recently, three new efforts were announced as part of that program: the Initiative for Data Science at UC Irvine, the Medical Humanities Initiative, and Water UCI. They join previously launched interschool initiatives in sustainability and exercise science. In an increasingly data-centric world, electronic information has become a critical element in modern research, education, medicine and business. The data science initiative will bring together faculty from almost every school to explore overlapping interests in the area of data science/big data, including methodology, infrastructure, theory, policy and education.

View UCI Data Science Press Release