Association for Computing Machinery
Welcome to the September 11, 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).

HEADLINES AT A GLANCE


Vint Cerf Wants Your Help Re-Imagining the Internet
InformationWeek (09/10/15) David Wagner

Google chief Internet evangelist and former ACM president Vint Cerf is starting a new project to solicit ideas from the public about how to improve the Internet, specifically to address issues such as education and what Cerf calls the "Digital Dark Age." Cerf, co-recipient in 2004 of the ACM A.M. Turing Award, kicked off the project by posting an open letter on social media site Lettrs.com. Cerf says he chose the format of a letter because he feels modern communication methods are too short and do not involve a lot of thought, compared to the careful consideration people once put into writing letters. In his letter, Cerf discusses his anxieties about the Digital Dark Age, the looming issue that, as software and technology advance, media recorded using older technology becomes unreadable as that technology falls out of use. Cerf also writes about how he hopes the Internet can change education for the better, in particular his excitement about the advent of Massively Open Online Courses (MOOCs). Cerf then asks others to follow his example and post their own letters with ideas about how to improve the Internet. He plans to take the best responses and share them online after presenting them at the inaugural Silicon Valley Comic Con next March.


Get Ready to Live in a Trillion-Device World
Computerworld (09/11/15) Sharon Gaudin

In an interview, University of California, Berkeley professor Alberto Sangiovanni-Vincentelli predicts within 10 years people will live in a world completely suffused with tiny sensors. "The entire environment is going to be full of sensors of all kinds," Sangiovanni-Vincentelli says. "Chemical sensors, cameras, and microphones of all types and shapes. Sensors will check the quality of the air and temperatures. Microphones around your environment will listen to you giving commands." Sangiovanni-Vincentelli predicts there also will be sensors within people's bodies, monitoring their health and thoughts so they do not even have to speak to interact with the connected environment. He dismisses concerns about the ramifications such a world would have for personal privacy by saying, "We've already lost it...everything is already recorded somewhere. What else is there to lose?" However, he acknowledges numerous challenges remain before such a world can be realized. More robust wireless networks and wireless communication protocols will be needed, business models will need establishing to create the trillions of tiny sensors, better security will be needed, and a larger cloud will be needed to handle all of the data such devices will generate. "It's actually exciting," Sangiovanni-Vincentelli says. "In the next 10 years, it's going to be tremendous."


New Limit to the Church-Turing Thesis Accounts for Noisy Systems
Phys.org (09/10/15) Lisa Zyga

It has previously seemed physical systems could violate the Church-Turing Thesis--a hypothesis developed by computer scientists Alonzo Church and Alan Turing in the 1930s and that in a sense defines what can be a computer--but new research from Princeton University and National University in Santiago, Chile suggests this is not the case. The presence of noise in a physical system can lead to unpredictable behaviors that seem incomputable, and would therefore violate the Church-Turing Thesis. "The physical interpretation of the 1930s Church-Turing thesis asserts that a physical system cannot be harnessed to perform computations that cannot be performed (in principle) by a standard computer," the researchers write in a new study. However, in their study, the researchers found computing the behavior of noisy systems is relatively straightforward. They did so by presenting a new limitation on the ability of physical systems to perform computation in the form of their memory. Using their new definition of memory to formulate what they call the "space-bounded Church-Turing thesis," they found the limit to a computer simulating a physical system is that they must share about the same amount of working memory. The finding could have implications for the development of so-called "hypercomputers," devices with computational capabilities well beyond those of existing computers.


Making IoT Configuration More Secure and Easy-to-Use
University of Southampton (United Kingdom) (09/10/15)

Researchers at the University of Southampton say they have identified easy-to-use techniques that will make the process of configuring Internet of Things (IoT) devices more secure. The team compared four approaches that use smartphone touchscreens to let users enter secure passwords. Two of the techniques were more traditional in that they connect a smartphone and an IoT device through a USB or audio cable, via the smartphone's headphone socket. The third approach, a Wi-Fi-only technique, has a smartphone create a special temporary Wi-Fi network, an ad-hoc network to which an IoT device automatically connects before being redirected to the correct permanent network. In the final option, a smartphone and an IoT device exchange information through light: the IoT device reads the flashing black and white binary pattern on the smartphone's screen to learn the password. In a presentation at the ACM UbiComp 2015 conference this week in Japan, the team said the audio cable and Wi-Fi-only interaction techniques were more usable than the other approaches.


AI Software Goes Up Against Fourth Graders on Science Tests
Technology Review (09/09/15) Tom Simonite

Allen Institute for Artificial Intelligence researchers are developing Aristo, a program designed to give machines a measure of common sense about the world. The software uses reasoning algorithms to answer questions using knowledge harvested from study guides and the Internet. The researchers are measuring their system by giving it standardized tests written for school children, and they want to convince other artificial intelligence (AI) researchers to adopt the method. "We can put our understanding of progress in AI and in natural language on an objective footing," says Allen Institute CEO Oren Etzioni. In early October, the Allen Institute will launch a contest challenging researchers to make software to address eighth-grade science questions. Aristo currently cannot pass a fourth-grade science test, which requires a score of 65 percent. The program can only solve multiple-choice questions, which make up two-thirds of the test, with about 75-percent accuracy on those questions that do not involve diagrams, and about 45-percent accuracy on those that do, according to Etzioni. New York University professor Ernest Davis thinks a better benchmark would be to create exam-style questions specifically for machines. "Standardized tests for humans don't get very good coverage of the kinds of problems that are hard for computers," Davis says.


Android Widgets May Boost Effectiveness of Sleep-Monitoring Apps
Penn State News (09/08/15) Matt Swayne

Pennsylvania State University (PSU) researchers have developed SleepTight, an Android sleep-monitoring app widget that serves as a data-capturing tool and provides visual reminders of the user's activities and sleep patterns. The researchers note traditional automated sensing systems can lower the capture burden to collect a lot of data, but they can lead to less engagement with the data and less self-awareness. "We thought that maybe the widget could ease the capture burden, as well as ease the access burden," says PSU professor Eun Kyoung Choe. The researchers studied 22 participants, and found those who used the widget version of the app were more likely to enter daily sleep diary information into the app than those who did not use widgets. Sleep diary adherence was 92 percent for the participants who had the widgets installed on their app compared to 73 percent for those who used the app without the widget. In addition, participants using the widget version viewed the sleep summary page more than participants who used the full app version. "This result indicates that the lock screen and home screen widgets reminded participants to view the sleep summary page and offered a shortcut to the sleep summary page," the researchers say. They presented their findings this week at the ACM UbiComp 2015 conference in Japan.


New Wearable Technology Can Sense Appliance Use, Help Track Carbon Footprint
University of Washington News and Information (09/08/15) Jennifer Langston

University of Washington (UW) researchers have developed MagnifiSense, wearable technology that can sense what devices and vehicles the user interacts with throughout the day. The researchers say the technology could help track the user's carbon footprint, enable smart home applications, or assist with elder care. In a study, MagnifiSense correctly classified 94 percent of users' interactions with 12 common devices after a quick one-time calibration. The system includes a sensor worn on the wrist that uses unique electromagnetic radiation signatures generated by electrical components or motors in those devices to identify when the user completes certain actions. In one 24-hour test, MagnifiSense correctly identified 25 out of 29 interactions with various devices and vehicles. The researchers combined three off-the-shelf sensors that use inductors, which proved to be the most accurate without being so power-hungry that wearing them would be impractical. In addition, the sensors capture a broad frequency range that enables the system to differentiate between electromagnetic radiation emanating from the unique combinations of different electronic components. The researchers also developed signal-processing and machine-learning algorithms to help the system correctly match those patterns with a certain kind of device. "The next steps are really to look at what other devices we can detect and work on a prototype that's wearable," says UW professor Shwetak Patel.


How Hashtags and @ Symbols Affect Language on Twitter
Georgia Tech News Center (09/10/15) Jason Maderer

New research on language in social media found people tend to follow many of the same communication etiquette rules as they do in speech. Researchers from the Georgia Institute of Technology (Georgia Tech) report tweeters are often more formal when they use hashtags, which enables messages to reach more people. Meanwhile, tweeters tend to use non-standard words such as "smh" or "nah" with the @symbol to address smaller audiences. People also use Twitter-specific terms to show their regional identity or tech savviness. In general, the researchers found heavy social-media users reserve standard English for the right social situations. The review of 114 million geotagged messages from 2.77 million users over three years helps explain a puzzle about language on social media, says Jacob Eisenstein, a professor in Georgia Tech's School of Interactive Computing. "Since social media facilitates conversations between people all over the world, we were curious why we still see such a remarkable degree of geographical differentiation in online language," says Eisenstein, who led the research.


A Target-Centric Approach to Cloud Security
Government Computer News (09/08/15) Patrick Marshall

Images can be protected in cloud storage by controlling access with signatures, such as file hashes and JPEG header metadata. However, those measures are not able to verify signatures within the network environment. Manchester Metropolitan University's Rob Hegarty and Nottingham Trent University's John Haggerty have developed a system called XDet, which associates a single block of data with a file and creates reference points within that block as a signature. Once a block signature is created, XDet monitors the HTTP/TCP data streams to look for the signed blocks. Hegarty and Haggerty acknowledge the reassembling of files does impose some overhead, which they are working to mitigate. When a mismatch of blocks either entering or leaving the network is detected, the connection to the server will either be terminated immediately or the mismatch will be logged for later action, depending on how XDet is configured. The researchers plan to enhance XDet in several ways, such as by extending it to sign other types of files in addition to images. Hegarty says the system also can be configured for file auditing if linked to a logging program.


Linguists Use the Bible to Develop Language Technology for Small Languages
University of Copenhagen (09/07/15) Carsten Munk Hansen

Professor Anders Sogaard from the University of Copenhagen and colleagues are using the Bible and Wikipedia to develop language technology for people who do not speak English or the world's other major languages. As part of project LOWLANDS: Parsing Low-Resource Languages and Domains, the team is using texts that have been translated into many languages. The Bible has been translated into more than 1,500 languages, while as many as 129 languages are represented by more than 10,000 Wikipedia articles. "We teach the machines to register what is translated with what is in the different translations of biblical texts, which makes it possible to find so many similarities between the annotated and unannotated texts that we can produce exact computer models of 100 different languages," Sogaard says. He notes the team has made the models available for other researchers and developers, enabling them to develop language technology for smaller languages. Sogaard says the language technology will enable people who speak languages such as Faroese, Welsh, Galician, or Yoruba to talk to their mobile phone, and use search engines or machine-translation systems.


Elephant or Teapot? Why Occlusion Is the Next Great Frontier for Image Recognition
The Stack (UK) (09/07/15) Martin Anderson

The intensity of research in the field of image recognition has delivered significant gains in recent years, driven in large part by neural networks and advances in database classification. Scientists at Facebook AI Research (FAIR) say classic challenges such as image classification, edge detection, object detection, and semantic segmentation are so near to being solved that it is time for the field to turn to its next great challenge: occlusion. Occlusion occurs when it is difficult to identify a given object because it is somehow obscured or obstructed in the frame, either by other objects or by tricks of perspective. In a recent paper, FAIR researchers asked human subjects to complete vector outlines for subjects in images in which they were occluded. The task involved, for example, distinguishing between a musician and an instrument they were holding in front of them or distinguishing the head of a stag from a thicket of branches. The researchers note humans were far better at identifying the occluded objects than artificial-intelligence systems. Occlusion is expected to be a tricky issue to tackle because its most extreme manifestations yield optical illusions that frequently confound even human subjects, with one example being the Kaniza Triangle.


Fortifying Computer Chips for Space Travel
Berkeley Lab News Center (09/04/15) Kate Greene

Computer chips used in spaceships need to be robust to withstand excessive radiation from high-energy sources such as the sun or cosmic rays. To ensure a spaceship's microprocessors are resilient, they undergo testing at places such as the Lawrence Berkeley National Laboratory. The lab's 88-Inch Cyclotron is used to accelerate ions to high energies along a circular path. Berkeley research coordinator Mike Johnson says some of the electronics now being tested by the lab are destined for the U.S. National Aeronautics and Space Administration's new Mars-bound spacecraft called Orion. The processors are mounted in a vacuum chamber facing a "cocktail beam" whose intensity of ions can be adjusted depending on the planned application. In the case of Orion, several radiation safeguards have already been put in place, such as using older and larger microprocessors with larger transistors that are less sensitive to interaction with an ion. The chips are encased in significant radiation shielding. Johnson says a 10-micron by 10-micron microbeam is now available at the lab, but within a year his team expects to reduce its size to the sub-micron level to better pinpoint radiation problems in chips.


Online Security Braces for Quantum Revolution
Nature (09/08/15) Chris Cesare

Powerful quantum computers able to break the security of the Internet are predicted to become a reality in a decade or so, but researchers say preparations need to be taken now. The Dutch General Intelligence and Security Service presented a scenario it calls "intercept now, decrypt later," in which a malicious attacker could start intercepting and storing financial transactions, personal emails, and other sensitive encrypted traffic and then unscramble it all once a quantum computer becomes available. Researchers believe new public-key cryptosystems are secure enough to protect secrets from quantum computers for all practical purposes. One such system is lattice-based cryptography, in which the public key is a grid-like collection of points in a high-dimensional mathematical space. A second option, known as McEliece encryption, hides a message by first representing it as the solution to a simple linear algebra problem. However, these systems require up to 1,000 times more memory to store public keys than existing methods, although some lattice-based systems have keys not much bigger than those used by existing RSA encryption. PQCRYPTO, a European consortium of quantum-cryptography researchers in academia and industry, released a preliminary report this week recommending the McEliece system, which has resisted attacks since 1978, for public-key cryptography.


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


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