ACM TechNews
Association for Computing Machinery
Welcome to the May 4, 2016 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).


Moore's Law Running Out of Room, Tech Looks for a Successor
The New York Times (05/04/16) John Markoff

A global coalition of computer chip manufacturers' decision to stop relying on Moore's Law to produce smaller, faster, and less expensive processors may force an industry reconsideration of non-silicon materials or new design concepts. "The end of Moore's Law is what led to this," says Georgia Institute of Technology researcher Thomas M. Conte. "Just relying on the semiconductor industry is no longer enough." The upcoming International Technology Roadmap for Semiconductors from the Semiconductor Industry Associations of the U.S., Europe, Japan, South Korea, and Taiwan is expected to signify widespread conviction that a rethink is in order. Meanwhile, IEEE on Wednesday announced a plan to create a new forecasting system, the International Roadmap for Devices and Systems, to track a broader range of computer technologies. Quantum computing could be one possible alternative, while another is graphene, which could enable smaller and faster transistors that also consume less power than silicon. The eventual end of Moore's Law was foreseen as early as 2005, when researchers began to fear processors' heat output was becoming excessive. The cessation of the decline of per-transistor costs of chips--a factor in the swift development of new computer technologies--is another sign the industry thinks signals Moore's Law is about to end.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

IBM Makes Quantum Computing Available in the Cloud
Computerworld (05/04/16) Sharon Gaudin

IBM Research is making its quantum processor freely available to the public via IBM's cloud to any desktop or mobile device. "This moment represents the birth of quantum cloud computing," says IBM Research's Arvind Krishna. The cloud-enabled quantum computing platform, called the IBM Quantum Experience, is designed to let people use individual quantum bits (qubits) to run algorithms and experiments on IBM's quantum processor. "There are several opportunities for material and drug design, optimization, and other commercially important applications where quantum computing promises to offer significant value beyond what classical computers can offer," says Jay Gambetta, manager of the Theory of Quantum Computing and Information Group at IBM's T.J. Watson Research Center. Analyst Charles King says IBM's five-qubit process should be powerful enough to handle a variety of research and other computations. Providing public access also should help validate work being done on quantum computing algorithms and applications, which previously could only be run in simulations. "If the project succeeds and leads to a clearer understanding of quantum computing, as well as workable larger systems, it will definitely be remembered as a game changer," King says. Moreover, analyst Earl Joseph notes the experiment gives the public an opportunity to help figure out how to program quantum computers.

China Sets Ambitious Goal to Reach Exascale by 2020
HPC Wire (05/02/16) Tiffany Trader

Speaking at the 12th High-Performance Computing (HPC) Connections Workshop, Beihang University professor Depei Qian revealed new information regarding HPC development in China and exascale plans in the works under China's 13th five-year plan (2016-2020). Qian described the progression of two 100-petaflop systems: the next iteration of Tianhe-2 and the upcoming Sunway system, with indications one or both will be introduced in June. Both systems were developed under China's 12th five-year plan, which also included new operation models and mechanisms for CNGrid, China's national HPC environment, and developing cloud-like application villages over CNGrid to promote applications. The 13th five-year plan puts into motion one of the most ambitious exascale programs in the world, which, if successful, would produce an exaflops supercomputer by the end of 2020 within a 35-MW power limit. To realize this, China is funding research into novel high-performance interconnects with three-dimensional chip packaging, silicon photonics, and on-chip networks. The exascale prototype will be about 512 nodes, offering 5 to 10 teraflops/node, 10 to 20 Gflops/watt, and point-to-point bandwidth greater than 200 Gbps. The goal of the exascale program is technology transfer, and China will work to field high-end domain-oriented servers based on exascale system technologies that utilize advances at the node, the interconnect, scalable I/O, storage, energy savings, reliability, and application software.

Women in Tech Band Together to Track Diversity, After Hours
The New York Times (05/03/16) Mike Isaac

Eight prominent women activists from Silicon Valley on Tuesday unveiled Project Include, an employee diversity campaign that seeks to draw attention to diversity issues in technology as the industry struggles with criticism about the composition of its workforce. The project's participants intend to obtain pledges from tech firms to track the diversity of their workforces over time and share that information with other startups. The initiative will concentrate on startups that employ 25 to 1,000 workers, in the hope of encouraging them to consider equality sooner. "If companies start early with diversity and inclusion, they don't have to bolt it on later, which is much harder," says project participant and former Google engineer Erica Baker. The project also will court the participation of venture capital firms to advise and mentor the startups. After a seven-month interval to define and track metrics, Project Include will publish results detailing the progress--or lack thereof--of the startups. Fellow Project Include collaborator Freada Kapor Klein says the prevailing industry culture promoting meritocracy above all is doing a disservice to women and minorities. The group's members hope the campaign will lead to more open discussions about inclusion. "A lot of these conversations are very uncomfortable for people," says former venture capitalist and project participant Ellen Pao. "These are exactly the types of uncomfortable conversations that we need to have."
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Can an Online Game Help Create a Better Test for TB?
NPR Online (05/02/16) Esther Landhuis

An online puzzle released on Monday marked the launch of "EteRNA Medicine," a challenge in which players of the Web-based game EteRNA attempt to design a sensor molecule that could ease tuberculosis (TB) diagnosis. The game was developed by Stanford University professors Rhiju Das and Purvesh Khatri, with the goal of getting nonexperts to design RNA tools that can disrupt the RNA mechanisms in a disease or virus. Das and Khatri say with every puzzle they solve, EteRNA players learn new principles of RNA design. Khatri notes the jumping-off point for the creation of EteRNA Medicine was their realization of "how awesome it would be to sequentially merge our two approaches--to use public data to find a diagnostic marker for a disease, and then use the public's help to develop the test." Das believes game players could help generate a simpler TB diagnostic test than is currently available. The latest EteRNA puzzle teaches players how to create RNA molecules that switch their folding patterns when other chemicals or molecules are present. Puzzle solutions are scored by actual molecules, which is a shift from other projects that used computer models to compare players' predictions. The winning RNA designs are selected by player vote, and then synthesized in a Stanford biochemistry lab. The EteRNA game has so far drawn more than 100,000 participants.

New Health Sensing Tool Measures Lung Function Over a Phone Call, From Anywhere in the World
UW Today (05/02/16) Jennifer Langston

University of Washington (UW) researchers have developed SpiroCall, a health-sensing tool that can accurately measure lung function over a phone call. The researchers note SpiroCall meets the medical community's standards for accuracy, having produced results that come within 6.2 percent of results from clinical spirometers used in hospitals and doctor's offices. "We wanted to be able to measure lung function on any type of phone you might encounter around the world--smartphones, dumb phones, landlines, pay phones," says UW professor Shwetak Patel. "With SpiroCall, you can call a 1-800 number, blow into the phone, and use the telephone network to test your lung function." In 2012, researchers from UbiComp Lab introduced a smartphone version called SpiroSmart. Users take a deep breath and exhale, and a phone's microphone senses sound and pressure from that exhalation and sends the data to a central server, which uses machine-learning algorithms to convert the data into standard measurements of lung function. Over the past four years, the UW team has collected data from more than 4,000 patients who have had their lung function measured using both SpiroSmart and a commercial spirometer. They report comparative data has improved the performance of the algorithms. The researchers will present a paper on their work this month at the ACM CHI 2016 conference in San Jose, CA.

Hacking Into Homes: 'Smart Home' Security Flaws Found in Popular System
University of Michigan News Service (05/02/16) Nicole Casal Moore

University of Michigan researchers have developed a way to hack into the leading "smart home" automation system and get the PIN code to a home's front door. The method, a "lock-pick malware app," was one of four attacks the researchers used on an experimental set-up of Samsung's SmartThings. The researchers performed a security analysis of the SmartThings' programming framework and conducted successful proof-of-concept attacks to show the impact of the flaws they found. For example, they demonstrated a SmartApp that eavesdropped on someone setting a new PIN code for the door lock, and then sent the PIN in a text message to a potential hacker. The app was disguised as a battery-level monitor and only expressed the need for that capability in its code. The researchers also showed an existing, highly rated SmartApp could be remotely exploited to virtually make a spare door key by programming an additional PIN into the electronic lock. A different SmartApp was shown to be able to turn off "vacation mode," which enables the user to program the timing of lights, blinds, and other household features to help secure the home while the owner is away. The researchers note one common security flaw is the platform grants its SmartApps too much access to devices and to the messages those devices generate.

Headaches Likely to Grow Over Auto Cybersecurity Concerns
Network World (04/27/16) Michael Cooney

Issues related to automobile cybersecurity are likely to become more pronounced in the near future as experts attempt to eliminate risks, but federal auto cybersecurity standards are not expected until at least 2018. A new report from the U.S. Government Accountability Office (GAO) also cites expert opinion that it will be about five years before technologies for securing and authenticating the legitimacy of communications via in-vehicle networks are incorporated during vehicle design and production. According to the report, today's vehicles often have multiple interfaces that make vehicle systems susceptible to cyberattacks. The onboard diagnostics port is one point of exploitation stakeholders are concerned with, while GAO also noted wireless attacks, such as those targeting weaknesses in autos' built-in cellular-calling capabilities, present the biggest risk to passengers. Remote cyberattacks on auto systems are a source of concern, but some experts say attacks comparable to current hacking demonstrations would be difficult to accomplish. Most industry experts polled by GAO agreed automakers should place safety-critical systems and non-safety-critical systems on separate in-vehicle networks and restrict communication between them. Two U.S. industry groups are spearheading a push to set up an Automotive Information Sharing and Analysis Center to collect and analyze intelligence information and support a forum for members to anonymously share threat and vulnerability information with one another.

RUB Researchers Develop Secure Audio Captchas
Ruhr-University Bochum (Germany) (04/29/16) Raffalea Romer

Ruhr University Bochum (RUB) researchers have developed a Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA) that makes use of human speech comprehension. Traditional CAPTCHAs contain small boxes with barely legible letters and symbols designed to distinguish human Internet users from machines. However, partially sighted people may be unable to read these CAPTCHAs, and must instead rely on audio CAPTCHAs, which consist of a synthetically generated and distorted sequence of words, letters, or digits the user must transcribe. However, audio CAPTCHAs often do not work very well, as human users have difficulties in solving them, whereas computers often show superior performance on the task. The RUB researchers' audio CAPTCHA system presents a sequence of words to the listener, some of which make sense, while the rest is gibberish. "A human will be able to spot and recognize the meaningful words," says RUB researcher Hendrik Meutzner. "The machine will find the task difficult, because the meaningful words and the gibberish sounds exhibit very similar characteristics in the time-frequency domain." For this type of audio CAPTCHA, the human success rate was 60 percent, compared to 14 percent for machines.

Creating a Prosthetic Hand That Can Feel
IEEE Spectrum (04/28/16) Dustin J. Tyler

Researchers at the Louis Stokes Cleveland Veteran Affairs Medical Center and partners have developed a "myoelectric" haptic system that is designed to work with a prosthetic hand. The team has tested the system on amputee Igor Spetic, who previously had electrodes implanted in his right arm that can stimulate different nerve fibers and produce realistic sensations. Spetic was able to reach into a bowl in front of him and pick up a cherry by its stem 93 percent of the time with the haptic system turned on, compared to 43 percent with it turned off. Spetic reported feeling as though he was grabbing the cherry, not just using a tool to grab it. As soon as the researchers turned on the stimulation, Spetic said, "it is my hand." The researchers installed thin-film force sensors in the prosthesis' index and middle fingers and thumb, and used the signals from those sensors to trigger the corresponding nerve stimulation. Work on the haptic interface partly falls under the U.S. Defense Advanced Research Projects Agency's new Hand Proprioception and Touch Interfaces program, and the team is now working toward having a fully implantable system ready for clinical trials within five years.

Computers Might Just 'See' Like Humans After All
Motherboard (04/29/16) Jordan Pearson

An international team of researchers from France, Iran, and Australia recently compared one of the most advanced deep-learning systems on the market to human vision. They concluded machines and humans experience greater difficulty in recognizing the same kinds of images, suggesting deep-learning networks actually follow similar "internal mechanisms" to human vision. The researchers hope other scientists will use their study to take lessons from the brain's visual processes and use them to make computer vision better. In addition, the researchers note if deep-learning computer models do work like the human brain, they also could be used to test hypotheses about how the brain works. "One of the main goals of artificial intelligence is to replicate different aspects of human behavior including vision," say researchers Timothee Masquelier and Saeed Kheradpisheh. Still, researchers already know a lot about how the brain processes information, and that knowledge can be used to improve computers. For example, modeling the data "spikes" caused by individual static inputs so motion over time can be accounted for would improve a computer's ability to process video. This type of research also could be used to implement feedback loops and enable computers to pay closer attention to certain parts of an image.

Computer Memory Finds Ally in Biology
Boise State University Update (04/25/16) Kathleen Tuck

Nucleic acid memory (NAM) could offer a better way to store digital information, according to researchers at Boise State University. A team led by professor Will Hughes is examining DNA's stability under extreme conditions. Hughes calls DNA the storage material of life in general. "Because of its physical and chemical properties, it also may become the data storage material of our lives," he says. The group is subjecting DNA strands to temperatures varying from -20 degrees Celsius to 100 degrees Celsius, and to a variety of ultraviolet exposures to see if they can still retain their information. They have found much less data is lost with NAM than with the current state of the industry. The researchers say nucleic acid far surpasses electronic memory in retention time, while also providing greater information density and energy of operation. Hughes will participate in a workshop this month at the U.S. Intelligence Advanced Research Projects Activity to discuss a portable DNA hard drive that would have 500 terabytes of searchable data. "When information bits are encoded into polymer strings, researchers and manufacturers can manage and manipulate physical, chemical, and biological information with standard molecular biology techniques," the researchers note in a paper published in April.

Promoting Women in Science
Amsterdam Science (04/19/16) Hamideh Afsarmanesh; Noushine Shahidzadeh

In an interview, Utrecht University professor Lynda Hardman discusses the pressing need to promote participation of women in science in the Netherlands. Hardman, who also is a member of the management team of Centrum Wiskunde & Informatica, the national research institute for mathematics and computer science in the Netherlands, cites cultural and psychological factors underlying the underrepresentation of women, including stereotypes that women are less science-oriented than men and a lack of female role models comfortable with science. "If the teachers are not able to supply this role model, then parents can get involved in sharing their knowledge and interest in science with the girls," Hardman says. She also says female candidates for high-ranking scientific positions should be rated based on quality instead of quantity, with women receiving more coaching and mentoring. Hardman envisions prestigious international associations contributing to the promotion of women in science, and notes Informatics Europe (an organization for which she serves as president) "gives guidance for heads of departments, whereas the Association for Computing Machinery [ACM] is more oriented to individual members." In particular, ACM Women Europe hosts events for younger female professionals at various educational levels so they can network and discuss issues together. Hardman says the institution of a code of best practices for women and information and communications technology by the European Commission does not ensure positive outcomes. She stresses local management should strive to apply the existing code of best practices to grow women's share of science roles.

Abstract News © Copyright 2016 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