Association for Computing Machinery
Welcome to the June 10, 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


It's a Fantastic Time to Graduate in the U.S. as an Engineer or Computer Scientist
Quartz (06/08/15) Max Nisen

Recent college graduates are doing well, according to a survey from the National Association of Colleges and Employers. The report covers 65 percent of the 266,119 people who graduated with a bachelor's degree in the country last year, and offers an early look at the job market. Graduates with degrees in engineering technology top the employment category with a full-time employment rate of 82.9 percent, followed by computer science at 73.2 percent, and communications technology at 71.8 percent. The employment rate is 62.2 percent for engineering, and 61.2 percent for those who majored in security. Engineering graduates also led the salary category with a mean starting salary of $64,891, followed by computer science at $62,194, and engineering technology at $57,090. The salary for mathematics majors is $52,821. Overall, the employment rate for the class of 2014 is 55 percent, and the average starting salary is $48,190.


Injectable Electronics Could Form Basis for Brain Implants
IDG News Service (06/09/15) Tim Hornyak

Flexible electronics injected directly into brain tissue could lead to new applications in medicine and brain-machine interfaces, according to research from Harvard University. The technology could be used to directly monitor and stimulate brain cells. Researchers from Harvard collaborated with experts from China's National Center for Nanoscience and Technology to test the technology. The team used tiny, flexible mesh structures made of conductive polymer strands and embedded with transistors and electrodes. The researchers rolled up the mesh, inserted it into a syringe, and injected it into the brain tissue of mice, where it unfurled. Mostly porous, the mesh can expand to fill biological cavities. The mesh integrated with the brain tissue and the mice did not show significant signs of immune reaction to the material after five weeks. External computers were connected to the mesh via extremely thin electrical wires, but the research team wants to make the technology wireless in the future. The technology could potentially be developed to treat brain damage from stroke and Parkinson's disease, according to Harvard's Charles Lieber.


When Google Self-Driving Cars Are in Accidents, Humans Are to Blame
The Atlantic (06/08/15) Adrienne LaFrance

In its latest report on the safety record of its fleet of self-driving cars, Google says in six years the vehicles have been involved in only 12 minor accidents, all of which were the fault of humans. The accidents include eight instances of the self-driving cars being rear-ended, one of a vehicle being merged into, and one incident in which a vehicle was struck from the side. A final incident, which did not result in any damage, involved the mirror of a non-Google car grazing a sensor on an automated Lexus. One of the incidents detailed in the report was the fault of a Google employee who was driving one of the self-driving vehicles manually when he rear-ended another vehicle. In the report, Google also discusses the improvements it has made to its cars, including the ability to distinguish ambulances from regular traffic and react accordingly. Chris Urmson, director of the self-driving car program, says the extensive testing Google has done will give its latest prototypes, "the equivalent of about 75 years of typical American driving experience." Google's new prototypes will be tested on the streets of Mountain View, CA, this summer.


Just Add Water: Stanford Engineers Develop a Computer That Operates on Water Droplets
Stanford Report (06/08/15) Bjorn Carey

Stanford University researchers have developed a synchronous computer that operates using the physics of moving water droplets. The researchers say the droplet computer theoretically can perform any operation that a conventional electronic computer can analyze, although at a significantly slower rate. "Our goal is to build a completely new class of computers that can precisely control and manipulate physical matter," says Stanford professor Manu Prakash. Almost every computer program requires several simultaneous operations, each conducted in a step-by-step process. A clock ensures these operations start and stop at the same times, which guarantees the information synchronizes. However, creating a clock for a fluid-based computer required a rotating magnetic field. The researchers built arrays of tiny iron bars on glass slides, and then laid blank glass slides on top with a layer of oil in between. The researchers then injected magnetic nanoparticle-infused water droplets into the system and turned on the magnetic field. Every time the field flips, the polarity of the bars reverses, signaling one clock cycle. Every water droplet moves exactly one step forward with each cycle. The presence or absence of a droplet represents the 1s and 0s of binary code, and the clock makes sure all of the droplets move in synchrony, allowing the system to run virtually forever without any errors.


Closing the Computer Science Gender Gap: How One Woman Is Making a Difference in Many Lives
The Conversation (06/08/15) Maria Klawe

Maria Klawe, president of Harvey Mudd College and former president of ACM, says the decline of women's participation in computer science (CS) can be reversed. Klawe cites statistics showing that although nearly a third of CS degrees were awarded to women as recently as the mid-1990s, today that number is only 14 to 18 percent. Female participation in CS education was low when Klawe joined Harvey Mudd in 2006, when only 10 percent of CS majors were women. Since then, the school has radically altered its CS curriculum to make it more appealing to women and those with little previous CS experience and now women make up 40 percent of CS majors. The changes the school made include a redesign of its introductory CS class, which is now the most popular class the university offers, and splitting the course into two sections: one for students with experience and one for those without experience. The faculty now assigns CS students to team-based projects, which cuts down on isolation, and offers students access to summer research opportunities, which has proven to increase female students' confidence and interest in the field. Klawe says more needs to be done to attract not just women, but minorities, to CS, and this should include teaching CS at much earlier ages.


Using Minecraft to Unboggle the Robot Mind
News from Brown (06/08/15) Kevin Stacey

Brown University researchers are developing an algorithm designed to help robots better plan their actions in complex environments with the aid of the virtual worlds based in the Minecraft video game. The algorithm changes standard robot-planning algorithms using goal-based action priors, sets of objects and actions in a given space that are most likely to help an agent achieve a goal. The researchers say Minecraft provides an ideal environment to test how well the algorithm learns action priors and uses them in the planning process. "There's a huge space of possible actions somebody playing this game can do, and it's really cheap and easy to collect a ton of training data," says Brown professor Stefanie Tellex. The researchers constructed small domains in a model of Minecraft, and then placed a character into the domain and gave it a task to solve. The algorithm-powered character had to try different strategies in order to learn the task's goal-based priors. After the algorithm completed several trials of a given task to learn the correct priors, the researchers moved to a new domain that it had not yet experienced. They showed that, equipped with priors, the Minecraft agents could solve problems in unfamiliar domains much faster than agents powered by traditional planning algorithms.


Tim Cook Says Lack of Diversity in Tech Is 'Our Fault'
Mashable (06/08/15) Christina Warren

In an interview during an orientation session for recipients of Apple's Worldwide Developer Conference Scholarship Program, Apple CEO Tim Cook discussed the importance of diversity in the technology industry and what Apple is doing to become a more diverse company. Cook called the idea that women do not want to be involved in tech a "cop-out" and said the current lack of diversity in the tech sector is "our fault," explaining the tech community itself has not done enough to reach out to women and demonstrate technology's appeal. He said one way to correct this is with more female role models and noted Apple is pursuing several strategies to reach out to girls and women at every level from junior high to college. Cook said Apple also is reaching out to African Americans by strengthening its ties to historically black colleges. He said what he sees as one of the biggest obstacles to making tech more diverse is "the appalling silence of the good people." Cook said he considers increasing diversity to be vital to Apple's future because doing so will enable it to make better products and be a better company.


As Open Source, Apple's Swift Language Could Soar
IDG News Service (06/08/15) Joab Jackson

Later this year, Apple will release the source code underlying its Swift programming language, a move that could bring the new language to more users. Once Swift becomes open source, programmers will be able to compile Swift programs to run on Linux as well as OS X and iOS. The source code will also include the Swift compiler and standard library, and community contributions will be "accepted--and encouraged," according to Apple. Analysts say Apple's decision to release Swift's source code is based on the realization that open source is a strategy that will send Swift to other platforms, and could be a path to wider adoption. "Open source has a lot of benefits for vendors," says IDC analyst Al Hilwa. He notes open source can provide greater visibility for a product and generate positive feedback in the development community. For example, Microsoft has benefited by open sourcing several pieces of infrastructure, including key parts of its .Net framework. Google's Android OS also could benefit from Apple moving Swift to open source because it is based on Linux. Apple also announced it is getting ready to release a second version of Swift, which will feature improved error handling, additional protocol extensions, and modular optimization.


Counting People With Wi-Fi
The UC Santa Barbara Current (06/08/15) Sonia Fernandez

University of California, Santa Barbara (UCSB) researchers have shown that a Wi-Fi signal can be used to count the number of people in a given space. The researchers say the technology could be used in a wide range of applications, including energy efficiency and search-and-rescue missions. They emphasized the technology does not require people to carry Wi-Fi-enabled devices for them to be counted. "Our approach can estimate the number of people walking in an area, based on only the received power measurements of a Wi-Fi link," says UCSB professor Yasamin Mostofi. The technique involves putting two Wi-Fi cards at opposite ends of a target area. Then, using only the received power measurements of the link between the two cards, the researchers can estimate the number of people walking in that area. The researchers say they have successfully tested the technique with up to nine people in both indoor and outdoor settings. The people-counting method relies in large part on the changes of the received wireless signal. The presence of people weakens the signal in the direct line of sight between the Wi-Fi cards. In addition, human bodies scatter the signal, resulting in a phenomenon called multi-path fading, when they are not in the direct line-of-sight path.


A Handy Collaborator
The Economist (06/06/15)

Although robots are good at precise, complex activities, they struggle at tasks humans find trivial, such as navigation and planning. University of Bristol researcher Walterio Mayol-Cuevas wants to use people for navigation and planning, leaving robots with the freedom to perform the more precise tasks they are better at performing. Bristol researchers have developed an intelligent hand-held robot with a grass trimmer base and a tentacle-like robotic arm that can move freely in any direction. The arm is equipped with motors that move a carbon-fiber rod, at the end of which a variety of tools can be attached. A second prototype of the robotic system has a spinning brush at its tip, which could serve as a cleaning tool. The robots employ basic gestures to communicate with their users, such as pointing at places where they would like to work. Tests indicate the devices reduce both the time required to complete a task and the difficult of doing so. Mayol-Cuevas says hand-held robots eventually will help experienced operators do a better job, and let unskilled users address difficult tasks. The two prototypes currently rely on external computers and motion-capture sensors to provide the information needed to guide their human operators. In the future, the researchers plan to change that by incorporating a Kinect motion-tracking system.


A New Grasp on Robotic Glove
Harvard Gazette (06/05/15) Kat J. McAlpine

Harvard University researchers have demonstrated a proof-of-concept prototype and completed experimental testing on a soft robotic glove that could help people suffering from loss of hand motor control to regain some of their daily independence. The technology incorporates social and psychological elements of design that promote translation and seamless adoption by its intended users. The researchers incorporated the patients' feedback at every stage of development of the glove to maximize its potential for translation. "In addition to glove function, we found that people cared about its appearance, which could have a big impact on whether or not the glove would be a welcome part of their daily routine," says Wyss Institute for Biologically Inspired Engineering researcher Kevin Galloway. The researchers adapted the mechanics of the glove to make it more comfortable and natural feeling to wearers. The glove is equipped with soft, multi-segment actuators that support the range of motions performed by biological fingers. The researchers now are working to improve the glove control system to enable it to detect the intent of the wearer. "The current goal is to refine the overall system sufficiently so we can begin a feasibility trial with multiple patients later this year," says Harvard Biodesign Lab founder Conor Walsh.


Team Develops Camera That Uses Sensors With Just 1,000 Pixels
Carnegie Mellon University (06/02/15)

Researchers at Carnegie Mellon University (CMU) and Columbia University have developed LiSens, a new camera that uses a sensor with just 1,000 pixels to produce images and videos at nearly a mega-pixel resolution. The researchers say the image is created by focusing the scene onto a digital micro-mirror array (DMD), an array of tiny mirrors that can direct light toward or away from the sensor, and then focusing the DMD onto the low-resolution sensor. LiSens uses a linear array of sensing elements such that each pixel on the sensor can add together light on an entire column of the DMD. Each measurement taken by the line-sensor is a coded sum along a line in the scene, and given multiple such measurements with different codes, it is possible to recover the image focused on the DMD at its full resolution, according to the researchers. LiSens, which is based on a single-pixel camera (SPC) that uses a single photonic detector to sense the scene, can be interpreted as a multi-pixel extension of the SPC. LiSens delivers measurement rates that are up to 1,000 times that of the SPC, which enables the new technology to sense scenes at much higher spatial and temporal resolutions.


The Good, the Bad, and the Robot: Experts Are Trying to Make Machines Be 'Moral'
California Magazine (06/04/15) Coby McDonald

As robots and artificially intelligent software become more capable and autonomous, many experts say at some point robots will need to have the ability to make moral decisions. There currently are two broad ideas for how to instill a robot with morality. The first strategy is called top down, and would involve explicitly programming the robot with moral guidelines to follow. However, experts say this approach has potential downsides. For example, trying to compute the moral consequences of its actions could prove too much for a robot's processor, and even the most rigorous guidelines have flaws and loopholes that could lead to undesirable consequences. The other approach, bottom up, would involve having the robot use a form of machine learning to "learn" moral behavior, possibly by observing human media. Moral issues will also start to crop up around how people perceive machines as they become more and more human-like. The law will have to determine the legal status of such robots and how to handle the potential legal repercussions of their autonomous actions. This will especially be true as robots and artificial intelligence become more complex and begin to exhibit emergent behavior, actions they were not specifically programmed to carry out.


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


To submit feedback about ACM TechNews, contact: technews@hq.acm.org
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe