Welcome to the June 6, 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).
HEADLINES AT A GLANCE
Smaller Chips May Depend on Vacuum Tube Technology
The New York Times (06/05/16) John Markoff
Vacuum tubes may enjoy a resurgence as computer chips continue to shrink, and researchers at the California Institute of Technology (Caltech) are laying the technological groundwork for the future of scaling using this concept. Caltech Nanofabrication Group director Axel Scherer and students have built circuits that operate like vacuum tubes but are one-millionth their normal size. They have invented a minuscule tube composed from metal, which can turn on and off the flow of electrons between four smaller probes. The switches use quantum tunneling to switch the flow on and off without any electron leakage, and Scherer thinks this phenomenon could enable modern vacuum tube circuits to consume less power and function faster than chips with transistors. "Effects that are currently problems in scaling are precisely those that we would like to use for switching in these next-generation devices," Scherer says. He notes unlike silicon, which is a semiconductor, his tubes can be made from a wide spectrum of conducting materials such as gold, platinum, tungsten, and molybdenum, which will serve to streamline the switches at the atomic level. Scherer doubts the tubes will immediately dethrone transistors, but Boeing is funding his research because of their potential use in space and aviation.
Google Developing Panic Button to Kill Rogue AI
InformationWeek (06/04/16) Dawn Kawamoto
Researchers at Google's DeepMind subsidiary and Oxford University are collaborating to create a panic button to interrupt a potentially rogue artificial intelligence (AI) agent. The researchers propose a framework to enable humans to repeatedly and safely interrupt an AI agent's reinforcement learning. In addition, this interruption could be done while simultaneously blocking an AI agent's ability to learn how to prevent a human operator from turning off its machine-learning capabilities or reinforcement learning. The researchers studied AI agents working in real time with human operators, considering scenarios in which the operators would need to press a button to prevent the agent continuing with actions that either harmed it, its operator, or the environment around it, and teach or lead the AI agent to a safer situation. "However, if the learning agent expects to receive rewards from this sequence, it may learn in the long run to avoid such interruptions; for example, by disabling the red button--which is an undesirable outcome," according to the researchers. When Google acquired DeepMind in 2014, the DeepMind founders imposed a buyout condition that Google would create an AI ethics board to follow advances that Google would make to the AI industry.
A Method to Image Black Holes
MIT News (06/06/16) Larry Hardesty
Researchers at the Massachusetts Institute of Technology (MIT) and Harvard University have developed Continuous High-resolution Image Reconstruction using Patch priors (CHIRP), an algorithm that could help astronomers produce the first image of a black hole. The algorithm would stitch together data collected from radio telescopes as part of the Event Horizon Telescope project, which aims to turn the entire planet into a large radio telescope dish. The telescope uses a technique called interferometry, which combines the signals detected by pairs of telescopes, so the signals interfere with each other. The MIT and Harvard researchers found that if the measurements from three telescopes are multiplied, the extra delays caused by atmospheric noise cancel each other out. Although this means each new measurement requires data from three telescopes instead of two, the increase in precision makes up for the loss of information. CHIRP is able to produce a more reliable image by using a model that is slightly more complex than individual points but is still mathematically tractable. CHIRP also uses a machine-learning algorithm to identify visual patterns that tend to recur in 64-pixel patches of real-world images. The researchers note those features were used to further refine the algorithm's image reconstructions.
Aussies on the Verge of Bionics 'Holy Grail' Ahead of Human Trials of Brain-Machine Interface Technology
News.com.au (06/02/16) Nick Whigham
Melbourne University researchers are developing a brain-machine interface designed to enable patients with paralysis and epilepsy to regain bodily control. The team inserts a stentrode, a small biocompatible implant, into a blood vessel next to the brain, which records the brain's electric activity while the information is interpreted by a decoding algorithm. To assuage fears of mind-control scenarios, lead researcher Tom Oxley says the device can only record and transmit activity from the brain and cannot implant information. The research was funded in part by the U.S. Defense Advanced Research Projects Agency (DARPA) after Oxley reached out to DARPA director Colonel Geoffrey Ling. According to Oxley and his team, the next step in the research will be conducting human trials in the next year. Oxley says the challenge will be in demonstrating a human's ability to actively use and control the transmitted signal. "The beginning is probably going to be slow," he notes. "We are aiming for basic control of a couple different directions on a computer screen with a cursor and then with that we hope to use that to manipulate mobility assist devices such as exoskeletons."
Researchers Uncover a Flaw in Europe's Tough Privacy Rules
The New York Times (06/03/16) Mark Scott
Europeans are legally entitled to ask search engines to delete links about themselves from online search results under certain circumstances, but that right's effectiveness may be in doubt, according to a new multi-university study. The researchers estimated in about 33 percent of examined cases, they were able to determine the names of people who had asked for links to be removed via basic coding. They created an algorithm that analyzed 283 links to U.K. news articles Google had removed from search results, cross-referencing the names in each report with each article's headline on Google.co.uk. The researchers say if a link to the online article did not appear in the results, they could infer the individual they had searched for had asked for the link to be deleted. New York University Shanghai's Keith Ross says his team identified 80 people's names within 103 articles. "This poses a threat to whether the 'right to be forgotten' can be maintained in the long term," Ross says. "If a hacker can easily find 30 or 40 percent of people's names from delisted articles, what is the point?" The researchers say the findings will likely put more pressure on European authorities to push online search engines to broaden the right to be forgotten across all of their global domains.
New Devices, Wearable System Aim to Predict, Prevent Asthma Attacks
NCSU News (06/01/16) Matt Shipman
A wearable system of sensors developed by researchers from North Carolina State University's (NCSU) ASSIST Center could help people with asthma predict and prevent asthma attacks by tracking environmental and physiological data. The Health and Environmental Tracker (HET) uses a variety of sensors connected to a wristband and a patch. "Our goal was to design a wearable system that could track the wellness of the subjects and in particular provide the infrastructure to predict asthma attacks, so that the users could take steps to prevent them by changing their activities or environment," says NCSU professor Alper Bozkurt. He says the patch binds to the user's chest and tracks movement, heart and respiratory rates, blood oxygen levels, skin impedance, and wheezing. Environmental conditions such as ozone levels, humidity, and temperature are monitored by the wristband, and a spirometer measures lung function similar to a peak flow meter, but is self-powered and collects more accurate measurements. Bozkurt says all collected data is then transferred to a computer to be recorded and analyzed. HET incorporates nano-enabled sensors with extremely low power consumption levels, which gives the devices long battery lives. The team plans to begin testing HET on a larger subject population this summer, and will develop software to alert users to potential asthma attacks.
Software Turns Webcams Into Eye Trackers
News from Brown (06/01/16) Kevin Stacey
Brown University researchers have created WebGazer.js, software designed to help website owners and developers determine which parts of a Web page are attracting users' eyes. The software turns integrated computer webcams into eye trackers that can infer where on a Web page a user is looking. "We're using the webcams that are already integrated in users' computers, which eliminates the cost factor, and it's more naturalistic in the sense that we observe people in the real environment instead of in a lab setting," says Brown University Ph.D. candidate Alexandra Papoutsaki. The code, which is embedded on a website, prompts users to give permission to access their webcams. After permission is given, the software uses a face-detection library to locate the user's face and eyes, and then converts the image to black and white, enabling it to distinguish the whites of the eyes from the iris. The system then uses a statistical model calibrated by the user's clicks and cursor movements, assuming a user looks at the spot where they click, so each click tells the model what the eyes look like when they are viewing a particular spot. It requires three clicks to get a reasonable calibration, after which WebGazer.js can accurately infer the location of the user's gaze in real time.
In a New Method for Searching Image Databases, a Hand-Drawn Sketch Is All It Takes
University of Basel (05/31/16)
Researchers at Switzerland's University of Basel have developed vitrivr, a tool for searching for digital images and videos by using hand-drawn sketches. Vitrivr enables users to create a sketch on a tablet or interactive paper, which will be used to find similar image and video results. The vitrivr system matches results to the hand-drawn queries by looking for similarities in color, shapes, and directions of movement in a video. Searches can be further specified by using search terms and examples of images and videos. The system is open source and free to access by international researchers. The scalability of vitrivr means it can be applied toward researchers' large multimedia collections, and it is already in use by Switzerland's Federal Office of Sport to analyze movement patterns in sports clips and to search collections of digital watermarks with the Basel Paper Mill. The European Union and the Swiss National Science Foundation's iMotion research project also employs vitrivr as the search engine for its large-scale video collections.
PC Hardware Is Physically Leaking Your Encryption Keys
Motherboard (06/01/16) Michael Byrne
The ethereal layer of electromagnetic radiation, heat, and noise surrounding computer hardware could potentially reveal information about the operations within the machine and the security measures employed to protect user data, according to researchers from the Israel Institute of Technology. "We have identified multiple side channels for mounting physical key-extraction attacks on PCs, applicable in various scenarios and offering various trade-offs among attack range, speed, and equipment cost," the researchers say. They employed targeted acoustic data, electric data, and electromagnetic data, and the hacks were effective in every case. Extracting an RSA key by analyzing coil whine required an hour of up-close recording, monitoring electric potential took a second and could be accomplished with a simple touch, and only an antenna was needed to monitor radiation. After capturing the data, the researchers note the rest of the hacking via hardware leakage requires just signal processing. Still, they consider some countermeasures, including hardware muffling and shielding schemes and introducing ciphertext randomization into the software itself.
A Smartphone to Help Keep the Elderly Safe
Rutgers Today (05/31/16) Jeff Tolvin
A team of researchers at Rutgers University is developing a smartphone app that could help keep senior citizens from falling. The tool would assess balance during daily activities, such as standing up, sitting down, and walking. The app uses sensors in a strapped-on smartphone to measure body movements such as sway, jerkiness, or acceleration, which can trigger loss of balance, and detects the rate of change in movements using the device's gyroscope. The researchers tested the app on six individuals ranging in age from 76 to 94 and found isolating movement patterns immediately before a change in balance could help predict when a person is about to fall. The team says they will use the recorded sway movements and acceleration of motion, along with blood pressure readings, to build a predictive model of falling. "What I love about this type of application is that it is objective--and the data never lie," says Rutgers professor William Craelius. "It should allow us to establish a baseline for patterns of movement and a repository of readings that we can easily manage in a database."
Algorithm Could Help Detect and Reduce Power Grid Faults
Binghamton University (05/31/16)
Binghamton University researchers say they have proved the Single Spectrum Analysis (SSA) algorithm is the best tool to help authorities remotely detect and locate power grid faults. Although the SSA algorithm, theoretically, is an optimal approach for accurate and quick detection, it has not been adopted in real-world engineering applications. The Binghamton researchers adapted and improved the algorithm for the new application in power grid areas. The time and location of anomalies within the grid currently are determined by well-known formulas such as the Event Start Time algorithm, which calculates differing arrival times of power changes in different geographic locations. The Binghamton researchers used simulation data generated by the Power System Tool box to prove the SSA algorithm is faster and more robust at finding changes in the power grid from generator or transmission line problems. SSA also could be used to predict problem areas in the future. "Being able to detect subtle changes in the power grid promptly, our approach has the potential to predict future problems by including a power system model," says Binghamton professor Yu Chen.
David Dill: Why Online Voting Is a Danger to Democracy
Stanford Report (06/03/16) Ian Chipman
In an interview, Stanford University professor and Verified Voting Foundation founder David Dill says computer scientists and security experts almost universally agree online voting is a threat to democracy. "Computers are very complicated things and there's no way with any reasonable amount of resources that you can guarantee that the software and hardware are bug-free and that they haven't been maliciously attacked," Dill warns. "The problems are growing in complexity faster than the methods to keep up with them." Dill contends the opportunity for hacking online voting systems is much broader compared to touchscreen voting, in part because neither voters nor election officials would see signs of tampering due to the secret ballot. Dill also notes the ability to rig an election is much harder with current paper-based voting systems, in comparison to touchscreen or online voting. Just as worrying to him are the many people and organizations that would benefit from election fraud, which range from political zealots to organized crime to adversarial nation-states. "If you have an election system where fraud can be committed and...that fraud is undetectable, then you don't really have a reason to trust the outcome of the election," Dill notes. He says Verified Voting is working to raise both the public and lawmakers' awareness of the dangers of online voting so they can make educated decisions.
Abstract News © Copyright 2016 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: [email protected]
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.