Association for Computing Machinery
Welcome to the June 29, 2015 edition of ACM TechNews, providing timely information for IT professionals three times a week.

Updated versions of the ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).


2015 STEM Index Shows Gender, Racial Gaps Widen
U.S. News & World Report (06/29/15) Alan Neuhauser

The results of the 2015 U.S. News/Raytheon STEM Index show that, despite the efforts of corporations and non-profits, the gender and racial gaps in science, technology, engineering, and math (STEM) have failed to improve and in some cases have gotten worse. According to the index, women continue to trail men in the number of STEM degrees granted, exam scores, and general interest in STEM fields. The same holds true for race, with White and Asian students and graduates overwhelmingly outperforming Back, Hispanic, and Native American students on all three metrics; this is especially the case for Black students, who have increased the number of STEM degrees they earn between 2000 and 2014. Despite this, the share of all STEM degrees earned by Black students has declined over this period, while the share earned by White students has increased. "There either hasn't been much progress or it's declined," says David Miller, a psychologist at Northwestern University who studies STEM trends. Several researchers say the continuing trends are due to persistent cultural attitudes that cause many to view STEM fields as the exclusive preserve of a certain kind of White or Asian male, which actively discourages women and minorities from seeking STEM education and employment.

Watson's Next Feat? Taking on Cancer
The Washington Post (06/28/15) Ariana Eunjung Cha

IBM's Watson supercomputer is being trained to find personalized cancer treatments by comparing patients' disease and treatment histories, genetic data, scans, and symptoms against the vast corpus of medical knowledge. Whereas human specialist teams could take weeks accomplishing this task, Watson can make such recommendations in minutes. The project began with Watson being fed names, ages, genders, medications, lab tests, imaging results, and notes from each visit for thousands of leukemia patients treated at a specific facility over the past few years. Researchers also selected key journal articles from the past for Watson to reference and started giving it access to newly published content. Watson produces a list of therapy possibilities and rates them according to low, medium, or high confidence. "I see technology like this as a way to really break free from our current healthcare system, which is very much limited by the community providers," says the University of Texas' Lynda Chin. Watson's evolution is rooted in its creators' priority to make it capable of reading and understanding natural language, generating hypotheses, and locating and parsing evidence to support or refute them. Watson also has learned from its failures, its successes, and user feedback to become smarter over time.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Automatic Bug Repair
MIT News (06/29/15) Larry Hardesty

Researchers at the Massachusetts Institute of Technology's (MIT) Computer Science and Artificial Intelligence Laboratory presented a new system that can automatically repair software bugs at ACM's Programming Language Design and Implementation conference this month. The system, dubbed CodePhage, repairs software flaws by automatically importing functionality from other applications. CodePhage works by first feeding a pair of inputs to the program being repaired, called the recipient, to see which behaviors generate crashes and which do not. It then runs these same inputs past the "donor" program to see how it handles them. The system records the donor's responses using a symbolic expression that enables it to rewrite the donor's responses into the recipient's code once it finds a set of constraints that ameliorate the issue. It then incorporates this new code into the recipient. The MIT researchers say they have tested CodePhage on seven common open source programs and it has been successful in all instances. CodePhage imported repairs from between two and four donors for each of the programs and was able to implement its repairs within only a few minutes. "The longer-term vision is that you never have to write a piece of code that somebody else has written before," says MIT professor Martin Rinard.

Hospital ICUs Mine Big Data in Push for Better Outcomes
The Wall Street Journal (06/25/15) Laura Landro

Some hospitals are using a big-data approach to mine massive volumes of intensive-care unit (ICU) data to help providers and patients realize improved outcomes. Researchers are sifting through years of medical records and information from multiple sources to uncover previously unknown correlations so more trouble spots and potential solutions can be identified. In collaboration with integrated-systems scientists from the Massachusetts Institute of Technology and human-factors experts at Aptima, Beth Israel Deaconess Medical Center has embarked on the Risky States initiative to measure risk levels in a specific ICU at any given time. An analysis of two years' worth of data on all ICU patients enabled the project team to identify situations that raise risk, including a high number of admissions, greater numbers of sicker patients, a higher percentage of nurses with less than a year of ICU experience, and a high patient-to-nurse ratio. The team then identified about 30 harms during and following the risky states. An application automatically extracts such data from electronic medical records and enables clinical staff to enter additional unit concerns or issues about a specific patient. A visual dashboard then estimates a risk score and presents it to the ICU staff in real time on monitors in the unit and on mobile devices.

Should a Driverless Car Decide Who Lives or Dies?
Bloomberg (06/25/15) Keith Naughton

With driverless vehicle technology development underway and being fast-tracked by major automakers, a key question concerns whether such technology should make ethical choices. Many car manufacturers are looking to Stanford University's Center for Automotive Research (CARS), which is focused on programming cars to make ethical decisions. It is hoped that self-driving vehicles will anticipate and avoid collisions, but in cases in which accidents are unavoidable, the car may have to choose between two evils, such as swerving onto a crowded sidewalk to avoid being rear-ended by another vehicle, or staying put and imperiling the driver. Among the questions ethicists are wrestling with are whether rules guiding autonomous vehicles should stress the greater good and save the most lives, while assigning no value to the individuals involved. "[Driverless cars are] going to set the tone for all social robots," predicts California Polytechnic University's Patrick Lin. "These are the first truly social robots to move around in society." CARS director Chris Gerdes this summer will be testing driverless vehicles programmed to observe ethical rules to make split-second decisions. One rule governs when it is appropriate to violate traffic laws and cross a double yellow line to make room for cyclists or double-parked vehicles.

Breakthrough in Graphene Production Could Trigger Revolution in Artificial Skin Development
University of Exeter (06/25/15)

University of Exeter researchers say they have discovered a new way to produce graphene that is significantly less expensive, and easier, than previous methods. The researchers used their new technique to create the first transparent and flexible touch sensor that could be applied to artificial skin or in robot manufacturing. Exeter professor Monica Craciun believes the discovery could lead to a graphene-driven industrial revolution. The technique involves growing graphene in an industrial cold wall CVD system, called nanoCVD, which is based on a concept already used for other manufacturing processes in the semiconductor industry. The researchers say their method grows graphene 100 times faster than conventional methods, reduces costs by 99 percent, and has enhanced electronic quality. "The extremely cost-efficient procedure that we have developed for preparing graphene is of vital importance for the quick industrial exploitation of graphene," says former Exeter professor Thomas Bointon. Exeter professor Saverio Russo notes, "this breakthrough will nurture the birth of new generations of flexible electronics and offers exciting new opportunities for the realization of graphene-based disruptive technologies."

IU Researchers Participate in LinkedIn Project to Tackle Economic, Employment Challenges
IU Bloomington Newsroom (06/24/15) Kevin Fryling

The LinkedIn Economic Graph Challenge courts researchers, academics, and data-driven thinkers to propose strategies for using LinkedIn data to address major economic problems. Indiana University (IU) researchers participating in the project will concentrate on the macro-evolution of industries to forecast economic trends and offer career guidance to professionals. IU Bloomington professor Yong-Yeol Ahn says the initiative aims to plot out the "entire global economy" by the end of the year. The project will tap Ahn's background in developing mathematical models to map complex systems including the brain, social networks, and culture. The study will seek to map the links between industries throughout the world, as well as the flow of people between these industries, while also identifying patterns in people's mobility and anticipating future employment trends. The researchers will identify emerging markets by harnessing earlier work by Ahn and colleagues that determined human capital flows toward areas of the greatest economic opportunity. "Users will not only know where they can find jobs but also predict bright spots on the horizon...and acquire the skills, or tools, they will need to get where they want to go," Ahn predicts.

Widespread Backing for U.K. Robotics Network
The Engineer (United Kingdom) (06/24/15)

The Engineering and Physical Sciences Research Council (EPSRC) has launched the U.K. Robotics and Autonomous Systems Network (UK-RAS Network) in an effort to boost Britain's efforts in robotics research. The goal of the robotics network is to bring its academic capabilities under national coordination, with hopes of driving innovation and growth in the industry, according to EPSRC. The UK-RAS Network will foster academic and industry collaboration. "To get it right we need to draw on the expertise of the U.K.'s research base and the ambition of industry," says Minister of State for Universities and Science Jo Johnson. "By working collaboratively, this network will only help to accelerate growth of a high-tech sector and pave the way for new high-value, skilled jobs--a win-win scenario for the U.K." Several universities are founding members of the network, which will work to build relationships with industry by supporting interdisciplinary mobility and secondments, as well as developing proof-of-concept projects and running design challenges. The network will coordinate activities at EPSRC-funded RAS-dedicated facilities and centers of doctoral training, and organize the U.K. Robotics conference, workshops, and exhibitions.

UW Researchers Demonstrate System to Transmit Power Over Wi-Fi
The Daily of the University of Washington (06/24/15) Arunabh Satpathy

Wi-Fi can power connected devices using a new system developed by a team at the University of Washington (UW). Called Power Over Wi-Fi (PoWiFi), the system can charge low-power devices up to 28 feet away. UW graduate student Vamsi Talla and colleagues say the system does not interfere with the Internet connection. PoWiFi makes use of a Wi-Fi router with special software. The system also has a hardware component that consists of sensors, including camera and temperature sensors and coin-cell batteries, built specifically for powering low-power devices. "On the router side, it's essentially software," Talla says. "We have to harvest the signals, so we have built custom sensors we want to wirelessly power." The team tested the system in six homes, which uncovered issues involving leakage and the range of the camera and temperature sensors. UW professor Joshua Smith says the possibilities of greater range could be investigated with the addition of an amplifier. Still, the researchers believe the system is ready for initial applications, and note the technology has the ability to scale. "If we can get better ranges, the possibilities are limitless, " Talla says.

What 24,000 Facebook Confession Posts Tell Us About College
Motherboard (06/23/15) Clinton Nguyen

A Tufts University study analyzed about 24,000 Facebook posts on the Tufts Confessions page dating back to late 2013 and organized them into 13 topic groups. The site enables students to anonymously post messages about student life. Soubhik Barari, author of the study and a Tufts computer science student, ran Facebook posts through a natural-language processing program he wrote in Python and separated the post contents into topical groups using Latent Dirichlet allocation. Barari also used topic modeling to identify trigrams, or keyword triplets that roughly correspond to a topic label. The study found Tufts students felt lonely most frequently, in 22 percent of the sampled posts. Keywords associated with loneliness include "want," "sometimes," or "talk," Barari found. "It's the sum of the entire neighborhood that constructs an interpretable basis when our model converges and gives meaning to a topic," he says. Barari believes similar studies can be expanded to other campus communities to analyze what is critically different between them.

Distributed Technique for Power 'Scheduling' Advances Smart Grid Concept
NCSU News (06/24/15) Matt Shipman

North Carolina State University (NCSU) researchers have developed a new technique for scheduling energy in electric grids, advancing the smart grid concept by coordinating the energy being produced and stored by both conventional and renewable sources. Existing approaches to scheduling are highly centralized and do not scale up well. In addition, the rise of on-site energy storage technologies makes power scheduling calculations significantly more complex. To address these problems, the NCSU researchers developed technology that utilizes distributed computing power to replace the traditional control center with a decentralized approach. When each device communicates with its immediate neighbors, the technology can calculate and schedule how much energy it will need to store, how much to contribute to the network, and how much to draw from the network. "Collectively, this distributed technique can determine the optimal schedule for the entire grid," says NCSU Ph.D. student Navid Rahbari-Asr. In addition, he says this distributed technique would help protect the privacy of homeowners and other power generators, because they would not be sharing their energy production, storage, and consumption data with a control center.

'Valleytronics' Development Could Lead to New Approaches for Spintronics and Quantum Computing
IEEE Spectrum (06/23/15) Dexter Johnson

Researchers at the University of Bath have made a surprising discovery in the field of valleytronics, which involves using the wave quantum number of an electron in a crystalline material to encode data. The use of valleytronics in silicon has been associated with the loss of electron speed and decoherence, which can ruin the quantum state of quantum computers. The team examined the behavior of electrons in the valleys of silicon-on-insulator quantum wells when exposed to a magnetic field. In polarized valleys it should be twice as difficult to polarize spins, but the researchers found at a low enough electron density it becomes easier when valleys are frozen. "This is a striking demonstration of how interactions between electrons lead to qualitatively new behavior," says Bath researcher Kei Takashina. The finding could impact the development of silicon-based spintronic devices or interface complementary metal-oxide semiconductor technology with silicon-based quantum information processing.

Abstract News © Copyright 2015 INFORMATION, INC.
Powered by Information, Inc.

To submit feedback about ACM TechNews, contact:
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe