Association for Computing Machinery
Welcome to the September 9, 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).


Google's DeepMind Makes Progress in Computer-Generated Speech
Financial Times (09/08/16) Richard Waters

Researchers at Google's U.K.-based DeepMind unit say they have made considerable progress in developing computer-generated speech, claiming tests on their WaveNet system with human listeners demonstrated significant reduction of a gap in quality between modern computers and human speech. A source familiar with the WaveNet research says the system diverges from existing text-to-speech solutions by concentrating on the actual sound waves being produced, instead of using human voice recordings to reassemble the sounds to match the language being spoken. WaveNet employs a neural network to analyze raw waveforms and attempts to model probable patterns. The source says the system is highly complex, and must digest at least 16,000 waveform samples every second, producing vast volumes of data. DeepMind says WaveNet's ability to model sound waves enables it to create speech that imitates any human voice, and that it can produce short piano compositions by sampling classical music. The researchers note computerized speech generation has garnered less interest than natural language recognition in the recent artificial intelligence race. "Allowing people to converse with machines is a longstanding dream of human-computer interaction," the DeepMind team says.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Quantum Computing May Have Scored in Australian Research Funding Round
ZDNet (09/08/16) Asha McLean

The Australian government is allocating more than $200 million in funding to nine new Australian Research Council Centers of Excellence, including centers devoted to developing advanced quantum technologies. The University of Queensland center will build quantum machines for practical applications and create quantum materials, engines, and imaging systems. Possible applications include material simulators, diagnostic technologies, and geographical surveying tools. The University of New South Wales (UNSW) received funding for its ARC Center of Excellence for Quantum Computation and Communication Technology, which was opened in April. UNSW researchers currently are working to build the world's first quantum computer in silicon. The university already has developed a way to write and manipulate a quantum code using two quantum bits in a silicon microchip, and a team of engineers has built a quantum logic gate in silicon. UNSW also will use data modeling to predict Australia's future climate extremes at the new ARC Center of Excellence for Climate Extremes. Several other Australian universities will establish centers devoted to astrophysics, biodiversity, gravitational wave discovery, exciton science, and low-energy electronics.

Montreal Universities Land Historic $213M Investment for Computer and Brain Research
Montreal Gazette (Canada) (09/06/16) Karen Seidman

The Canadian government has announced a cumulative $213-million grant from the Canada First Research Excellence Fund to Montreal universities to invest in quasi-human artificial intelligence (AI) development, medical technologies, and brain disease disablement research. "These major investments will help these institutions emerge as leaders in these fields and attract the most brilliant researchers here," says Marie-Claude Bibeau, Canada's minister of international development and la francophonie. The University of Montreal, Montreal Polytechnical University, and HEC Montreal plan to use the grant money to establish an institute called IVADO for the purpose of advancing AI and big data in the quest for computers with quasi-human-level performance. McGill University professor Yoshua Bengio says IVADO will concentrate on the science of optimization, while the funding also will help entrench "a mini Silicon Valley in Montreal." Meanwhile, Montreal Polytechnical will use its Montreal TransMedTech Institute to transform diagnosis and medical intervention research to expedite the availability of new medical technologies or interventions. McGill will apply its grant to becoming a global brain research hub via the Healthy Brains for Healthy Lives initiative, tapping its neuroscience expertise to make mental-health and neurological disorders less burdensome and possibly even curable.

The Exascale Computing Project Awards $39.8M to 22 Projects
HPC Wire (09/07/16) Tiffany Trader

The U.S. Department of Energy's (DoE) Exascale Computing Project (ECP) on Wednesday announced its first round of funding, fully underwriting 15 application development proposals and providing seed funding for seven more. The $39.8-million award round will affect 22 projects and 45 research and academic organizations, chosen for their significance to society and their potential for advancement via exascale computing. Co-design capabilities also weighed significantly in the selection process because integration and co-design are critical to ensuring the ECP can fulfill its objective of a production exascale system that is 50 to 100 times faster than current high-performance computing (HPC) systems. "A key element of the ECP's mission is to deliver breakthrough HPC modeling and simulation solutions that confidently deliver insight and predict answers to the most critical U.S. problems and challenges in scientific discovery, energy assurance, economic competitiveness, and national security," says ECP director Paul Messina. ECP's primary goals include developing a wide-ranging set of modeling and simulation applications that support the DoE's scientific, engineering, and nuclear security programs. Other key objectives include creating productive exascale computing hardware and software by 2023, readying multiple DoE facilities to house exascale machines by 2023, and maximizing HPC's benefits to advance U.S. science and commerce.

Only 10 Percent of the Global Cybersecurity Workforce Are Women
SC Magazine (09/05/16) Danielle Correa

Women account for an estimated 10 percent of the global information security workforce, according to a new report from U.K.-based nonprofit CREST. Twenty-six percent of information technology (IT) professionals worldwide are women, and only 18 percent of computer science undergraduate degrees in the U.K. are awarded to female students. CREST suggests the lack of female applicants to IT jobs can be attributed to the perception of the cybersecurity industry being inhospitable to women. However, most attendees at the 2016 Diversity Workshop who participated in the study say they have not experienced sexist or inhospitable IT workplaces. Several attendees expressed concern the industry is too quick to emphasize technical skills over all other abilities; workshop attendees agreed that hiring from a diverse pool of applicants would lead to a workforce with a more diverse range of skills and experiences. To drive change in the industry, respondents say education, awareness, and support for women with an interest in the field can increase gender diversity. Outreach programs targeting school-age children and college students could be particularly effective. "Although most of our workshop attendees agreed that cybersecurity is welcoming to women, the perception from outside the industry is much the opposite," says CREST president Ian Glover. "It is clear that this is one of the major challenges we face."

Beam Me Up to the Video Conference
Fraunhofer-Gesellschaft (09/01/16) Kathleen Schroter

Researchers at the Fraunhofer Institute for Telecommunications' Heinrich Hertz Institute (HHI) in Germany have created a technique for virtually transmitting a fully-sized, three-dimensional (3D) image of a person that viewers can walk around and see from different angles. This was achieved with a stereoscopic camera system that produces well-estimated distances and an accurate 3D impression. "We are currently using more than 20 stereo cameras to map a human," says HHI's Oliver Schreer. The cameras are used in conjunction with algorithms that can rapidly derive depth information from the images, a necessity for calculating the 3D form of a captured person. The computer estimates a virtual model of the human, which is transferred into the virtual scene. "In developing these algorithms, special care has been taken to ensure they work efficiently and fast, so the movements of dialogue partners can very quickly be converted into a dynamic model," Schreer says. He notes the 3D data from the camera images are fused in a matter of seconds, and the system transmits the 3D dynamic model quickly in virtual reality. "Our goal is that in the future a realistic image copy of a human is able to directly interact with the virtual world--for example, to let it grab virtual objects," Schreer says.

Judging a Book Through Its Cover
MIT News (09/09/16) Larry Hardesty

An imaging system that can read closed books is being designed by Massachusetts Institute of Technology (MIT) researchers and colleagues at the Georgia Institute of Technology (Georgia Tech). The prototype was tested on a stack of papers, each with a single letter printed on it, and the system correctly identified the letters on the top nine sheets. The system uses MIT-developed algorithms that obtain images from individual sheets, while an algorithm from Georgia Tech interprets the often misshapen or incomplete images as individual letters. The prototype has a standard terahertz camera that emits short bursts of radiation, and a sensor to detect their reflections so the MIT algorithm can measure the distance to the book's individual pages. The algorithm also filters out noise caused by the sensor's electronics and other disruptions, and it can accurately infer the distance from the camera to the top 20 pages in a stack. MIT Media Lab researcher Barmak Heshmat believes the system could be employed to analyze any materials arranged in thin layers. "The Metropolitan Museum in New York showed a lot of interest in this, because they want to, for example, look into some antique books that they don't even want to touch," he says.

The Enigma Machine Takes a Quantum Leap
University of Rochester NewsCenter (09/06/16) Peter Iglinski

A quantum enigma machine developed by researchers at the University of Rochester, the Massachusetts Institute of Technology (MIT), and the U.S. National Institute of Standards and Technology was used to test the theory of quantum data locking in message encryption. A concept advanced by MIT professor Seth Lloyd, quantum data locking proposes using photons to relay messages securely. The quantum enigma machine, developed by Rochester researcher Daniel Lum and professor John Howell, can generate photons that travel through free space and into a spatial light modulator (SLM), which modifies the properties of the individual photons to properly encode the message into flat but tilted wavefronts that can be concentrated to unique points according to the tilt. Lum says the SLM also distorts the photons' shapes into random patterns. The sender and the recipient of the encrypted message both know the keys that identify the implemented scrambling operations, so the recipient can use his own SLM to flatten the wavefront, re-focus the photons, and translate the changed properties into the message. Lum says his team also employed the uncertainty principle to securely lock in six classical data bits using only one bit of an encryption key.

Quantum Computing Has the Cybersecurity World White-Knuckled
IDG News Service (09/07/16) Katherine Noyes

Cybersecurity experts are anxious about modern encryption being cracked by a quantum computer, with Global Risk Institute adviser Michele Mosca warning in a new report of a one-in-seven chance that quantum computing will break basic public-key cryptography tools by 2026. Because quantum computing relies on quantum bits that can exist as both 0 and 1 simultaneously, such systems are far faster and more efficient than conventional computers--with the downside of also being able to rapidly factor large numbers, which underpin the security of many cryptographic solutions. "Although the quantum attacks are not happening yet, critical decisions need to be taken today in order to be able to respond to these threats in the future," Mosca warns. He says existing cybersecurity systems are unprepared for a quantum hack. "There is a pending lethal attack, and the clock is ticking to design and deploy the cure before the threat is realized," Mosca cautions. He says the short-term goal should be designing "cryptographically agile" systems that can swiftly swap one cryptographic tool for another. Longer-term, Mosca says "quantum-safe" cryptography tools, including protocols that operate on conventional technologies and shield against quantum attacks, need to be created.

Team of Robots Learns to Work Together, Without Colliding
Georgia Tech News Center (09/06/16) Jason Maderer

New algorithms created by Georgia Institute of Technology (Georgia Tech) researchers enable robots to move within inches of each other, without colliding, to complete their task. The team, led by roboticist Magnus Egerstedt, director of Georgia Tech's Institute of Robotics and Intelligent Machines, is the first to create such minimally invasive safety algorithms. A set of safe states and barrier certificates ensure each bot stays in its own safe set throughout the entire maneuver, the researchers note. In a demonstration involving four robots, the machines approach from four different areas, meet in the middle, circle counterclockwise within inches of each other, then fan out into opposite directions. In another demonstration, eight robots perform the same task, this time circling clockwise before dispersing. Instead of keeping their distance and taking the long way around their neighbors, the robots move very independently wherever they wish. "Robots are very conservative--they want to make sure they're safe," Egerstedt says. "You couldn't pack the interstate with self-driving cars with today's technology." He says a minimally invasive safety controller similar to the algorithms could be used for the next generation of air traffic control.

Researchers Use Hardware to Accelerate Core-to-Core On-Chip Communication
NCSU News (09/06/16) Matt Shipman

A new chip design developed by researchers at North Carolina State University (NCSU) and Intel significantly accelerates core-to-core communication. Software instructions for multiple processors to work in a coordinated way are replaced with built-in hardware that coordinates communication between cores. NCSU professor Yan Solihin says execution times--from start to finish--are twice as fast or faster with the approach, called the core-to-core communication acceleration framework (CAF). The CAF design makes use of a queue management device (QMD), a small attachment to the processor network on a chip. Capable of simple computational functions, the QMD effectively keeps track of communication requests between cores without having to rely on software routines. Moreover, Solihin says the QMD can be used to aggregate data from multiple cores, expediting some basic computational functions by as much as 15 percent. "We are now looking at developing other on-chip devices that could accelerate more multi-core computations," Solihin notes.

New 'Gel' May Be Step Toward Clothing That Computes
LiveScience (09/06/16) Edd Gent

A team of researchers from the University of Pittsburgh has designed a theoretical model of a material that is capable of computational pattern recognition using an oscillating gel. They say the combination of stimuli-responsive materials and unconventional computing could one day be used in wearable sensors. This computational material is composed of a chemical gel overlaid with a piezoelectric element that continually expands and contracts in response to certain chemical reactions. When the gel oscillates, the mechanical motion is transferred to the piezoelectric beam, which generates a voltage. The researchers say when multiple units of the gel are wired together, the electrical signals enable them to communicate and synchronize. By encoding the material with electrical polarities representing certain patterns of pixels, the system is able to recognize patterns in images. The researchers note the gel's computing mimics processes seen in the human body, such as the retina's ability to compress visual information before sending it to the brain. Research indicates the material can analyze patterns in pressure, chemical stimulation, or light, with possible applications including shoe insoles and skin for robotic arms.

Science Makes First Study of Know-It-All Internet Commenters
The Stack (UK) (09/05/16) Martin Anderson

The first study of intransigent commenters in social networks has been conducted via a method developed by researchers at Stanford University and Microsoft. Their analysis system was trained on 5,000 annotated posts from Reddit communities, but the extracted principles were applied to millions of other comments on the site. Among the knowledge tasks set up by Stanford's Ethan Fast and Microsoft's Eric Horvitz was determining whether dogmatic commenters are uniformly stubborn across a range of subjects, and for this purpose they produced 10 million analyzed posts from 2007-2015 across 1,000 users. The study found some of the most intransigent commenters express their convictions across less likely subjects, and appear to be just bad-tempered. "For example, among the users who are dogmatic on politics, they are also disproportionately dogmatic on unrelated subreddits such as science, technology, IAmA, and AskReddit," the researchers say. The study indicates dogmatic commenters post often and in favorite communities, but they also appear to most value placing a first or at least early comment, as they are "not as inclined to engage with discussion, once it has begun." Fast and Horvitz suggest the results of their computational model could help users engage in more pro-social behavior in online communities.

Abstract News © Copyright 2016 INFORMATION, INC.
Powered by Information, Inc.

To submit feedback about ACM TechNews, contact: