Welcome to the April 3, 2017 edition of ACM TechNews, providing timely information for IT professionals three times a week.

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

Facebook headquarters Tech Jobs Are Thriving Nationwide--Up to 7.3M
USA Today
Jon Swartz
April 3, 2017

The U.S. technology job market has expanded 2 percent to about 7.3 million employees since 2016, with 6.9 million employed by technology companies, according to CompTIA's latest annual Cyberstates report. Tech employees earned an average of $108,900 last year, more than double the national average. CompTIA's Tim Herbert says the growth in tech jobs amounts to about 4 percent of the U.S. workforce, and $1.3 trillion in revenue. The report also found the increasing migration of both tech and non-tech companies to cloud computing and other software services is accompanied by concurrent job growth. The increase and allure of tech-related jobs has spurred groups such as Black Girls Code to lobby for more training and access to historically underserved potential employees. The report noted surges in tech employment in California, New York, North Carolina, Texas, and Michigan.

Full Article
Why Asimov's Laws of Robotics Should Be Updated for the 21st Century
March 29, 2017

Tom Sorell, a professor at the University of Warwick in the U.K., suggests Isaac Asimov's famed "Three Laws of Robotics"--envisioned as moral safeguards against robots harming humans--need updating for the modern era, in which robots of increasing sophistication are being employed to conduct ever-more-complicated tasks designed to protect and care for people. For example, Sorell says the laws are problematic when applied to human-guided military drones designed for remote killing, which can be argued to be both adhering to and violating the First Law. The laws are appropriate when the robot's purpose is to keep seniors safe, but robotics often falls into a spectrum of assistive tools designed to promote seniors' independence. The First Law would be violated, for example, if a robot allowed its human to make independent decisions leading to injury. However, Sorell contends human autonomy must be maintained by both robots and other people.

Full Article
Brain-Controlled Muscle Stimulation Makes Paralyzed Man Move Limbs Again
International Business Times
Judy Cordova
March 31, 2017

Researchers at Case Western Reserve University have used neuroprosthesis to help a 56-year-old paralyzed man regain movement of his limbs. Scientists have long sought a method for using cortically-controlled neuroprosthesis for intracortical brain-computer interfaces. The Case Western researchers implanted two small recording chips in the patient's motor cortex, as well as 36 electrodes in his right arm. The patient was able to coordinate reaching and grasping movements using his own arm and hand. The brain-implanted chips measure the neurons firing when the patient thinks about moving his limb. The signals are processed by an algorithm and then transmitted to the electrodes in the patient's upper and lower arm. "What we are doing is circumventing the spinal cord injury," says Case Western's Bolu Ajiboye. The team notes this study is an example of how neuroprosthetics can restore function and sensation for people with high-cervical spinal cord injuries.

Full Article

co-authors Daphne Yao, Fang Liu, and Gang Wang Virginia Tech Researchers: Android Apps Can Conspire to Mine Information From Your Smartphone
Virginia Tech News
Amy Loeffler
April 3, 2017

Researchers at Virginia Polytechnic Institute and State University (Virginia Tech) say they have conducted the first-ever large-scale systematic study of how the trusty applications on Android phones are able to talk to one another and trade information. The Virginia Tech researchers found these threats fall into two major categories--a malware app that launches a cyberattack, or apps that enable collusion and privilege escalation. As part of the study, the researchers developed DIALDroid, a tool that performs inter-app security analysis. The researchers studied 110,150 apps over three years, including 100,206 of Google Play's most popular apps and 9,994 malware apps from Virus Share. The team found the biggest security risks were some of the least utilitarian, such as those related to personalization of ringtones, widgets, and emojis. The researchers were to present their findings today at the ACM Asia Computer and Communications Security Conference (ASIACCS 2017) in Dubai, United Arab Emirates.

Full Article
IBM Technology Creates Smart Wingman for Self-Driving Cars
Network World
Michael Cooney
March 30, 2017

IBM announced it has patented a machine-learning technology that defines how to shift control of a self-driving car between a human driver and vehicle control in case of a potential emergency. IBM's system uses onboard sensors and artificial intelligence to ascertain potential safety issues and determine whether the autonomous vehicle should hand over control to a driver. "The technology would be a smart wingman for both the human and the self-driving vehicle," says IBM Research's James Kozloski. Another patented technology is designed to enable safer interaction between human-driven and self-driving cars by applying sensor-facilitated automatic driver modeling to monitor driver behavior. The system makes assumptions about a car and forwards those observations to other autonomous vehicles in the vicinity. "The whole idea with both patents is to enable a safer environment for humans," Kozloski says.

Full Article

automated face detection, illustrative photo Finding Faces in a Crowd
Carnegie Mellon News (PA)
Byron Spice
March 30, 2017

Researchers at Carnegie Mellon University (CMU) have developed an improved method for coding important context from an image that enables a significant advance in detecting tiny faces. When applied to benchmarked datasets of faces, the new method reduced errors by a factor of two, and 81 percent of the faces found using the CMU methods proved to be actual faces, compared with up to 64 percent for prior methods. The researchers say the technique could have multiple applications, such as performing headcounts to calculate the size of a crowd. In addition, detecting small items in general will become increasingly important as self-driving cars must monitor and evaluate traffic conditions in the distance. The new method uses "foveal descriptors" to encode context in a way similar to how human vision is structured. The researchers also improved the ability to detect tiny objects by training separate detectors for different scales of objects.

Full Article

Rows of compressed images A Faster Single-Pixel Camera
MIT News
Larry Hardesty
March 29, 2017

Researchers at the Massachusetts Institute of Technology (MIT) say they have developed a new computational technique that makes image acquisition using compressed sensing 50 times as efficient. The researchers say the technology, when applied to a single-pixel camera, could reduce the number of exposures needed from thousands to only dozens. Compressed-sensing imaging systems do not require lenses, which could make them useful in harsh environments or in applications that use wavelengths of light outside the visible spectrum. The new compressed-sensing technique depends on time-of-flight imaging, in which a short burst of light is projected into a scene, and then ultrafast sensors measure how long the light takes to reflect back. The MIT researchers also describe a procedure for computing light patterns that minimize the number of exposures. Using synthetic data, the researchers compared the performance of their reconstruction algorithm to that of existing compressed-sensing algorithms.

Full Article

Brain scans MRI Brain Scans Train Machines to See the World More Like Us
New Scientist
Matt Reynolds
March 29, 2017

Researchers at Harvard University have trained algorithms to process images more like humans do. The researchers analyzed how regions of the brain's visual cortex responded to images containing four different types of objects, including humans, animals, buildings, and food. The data was taken from a volunteer who viewed more than 1,200 images while a functional magnetic resonance imaging machine measured their brain response. The researchers found the different objects had their own corresponding pattern of brain activity, and the strength of the signals indicated how difficult each image was to classify. The team used this information to train its machine-learning algorithms. The researchers note this method could enable relatively basic machine-learning models to approach the accuracy of state-of-the-art neural networks. The researchers next plan to examine how cells in rat brains react to different images, with the goal of developing systems that more closely mimic human decision-making.

Full Article

A GPS station near Mount St. Helens Wall Street Algorithm Helps Scientists Track 'Slow Slip' Earthquakes
Brooks Hays
March 28, 2017

Researchers at the University of Washington (UW) have adapted Wall Street's relative strength index, a measure of a stock price's rate of change, to identify the signatures of slow slips among seismic noise in the gradual movement of Earth's tectonic plates. The researchers used the algorithm to analyze data collected from 2005 to 2016 by 200 global-positioning system (GPS) stations along the West Coast's Cascadia subduction zone. The researchers found the algorithm successfully identified the same slow slips previously found during more comprehensive surveys. The new algorithm will help geologists study a variety of geophysical phenomena, especially in places with few GPS stations. "We might use the method to look at the seismic effects of groundwater extraction, volcanic inflation, and all kinds of other things that we may not be detecting in the GPS data," says UW researcher Brendan Crowell.

Full Article
Making America's Power Grid Much, Much Smarter
Vanderbilt University
Heidi Hall
March 29, 2017

Researchers at Vanderbilt University, Washington State University, and North Carolina State University are working to reinvent and protect the U.S.'s power grid. They will present their first solutions, including the Resilient Information Architecture Platform for Decentralized Smart Systems platform, this month at the ACM/IEEE International Conference on Cyber-Physical Systems (ICCPS 2017) in Pittsburgh. The goal is to build an underlying, open source software platform to support decentralized applications that boost the power grid's resilience and protect it from dangers. In addition, the platform is designed to quarantine problems while the rest of the system runs normally and to permit seamless transitions among power sources as needed. The new microgrid controller uses fog computing to calculate and monitor voltages and phase angles and employs inverters to change those when necessary.

Full Article

Actors with holodeck in background Star Trek's Holodeck: From Science Fiction to a New Reality
The Conversation
Fabio Zambetta
March 28, 2017

The Holodeck featured on "Star Trek: The Next Generation" has inspired many researchers in artificial intelligence (AI), human-computer interaction, and computer graphics, and those fields are converging to make the Holodeck a practical reality, writes Fabio Zambetta, a senior lecturer at RMIT University in Australia. Strides have been made with virtual reality (VR) and augmented reality (AR), but the technology is limited by inherent restrictions in bulky tools, real-world engagement, and less-than-natural interfaces. Meanwhile, progress in speech recognition, language translation, sensors, and information searching is being incorporated into hands-free or wearable user interfaces. The primary tool for real-world Holodeck programs is AI embedded within virtual characters, although recent AI and machine-learning innovations have yet to become as sophisticated as those needed to enable Holodeck-quality AIs. One forecast envisions current VR and AR advances facilitating the refined headset necessary for a Holodeck experience within the next five to 10 years.

Full Article
It's Alarmingly Easy for Machines to Control Us
Katharine Schwab
March 22, 2017

A new artwork exhibit at the Science Gallery Dublin in Ireland illustrates how easy it is for machines to use humans by persuading people to internalize external interfaces. The installation employs electric muscle stimulation to encourage visitors to crank a lever and feed the machine more electricity. "At first the machine tells you when to do it, but at some point you and the machine are synchronized, and it becomes a kind of physical meditation exercise," says Pedro Lopes, a human-computer interaction researcher at the Hasso Plattner Institute in Germany. "That's the moment you internalize." Lopes says the experiment highlights a logical gap in people's technology-related anxieties. He notes there is clear concern of losing control to tangible machine interfaces, while more insidious and user-friendly technologies, such as social networks, assert more control over humans than they may realize because they give people the illusion of control.

Full Article
ACM Books

Association for Computing Machinery

2 Penn Plaza, Suite 701
New York, NY 10121-0701

ACM Media Sales

If you are interested in advertising in ACM TechNews or other ACM publications, please contact ACM Media Sales or (212) 626-0686, or visit ACM Media for more information.

To submit feedback about ACM TechNews, contact: technews@hq.acm.org