Association for Computing Machinery
Welcome to the June 5, 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


The Emerging Science of Human Computation
Technology Review (06/04/15)

The wisdom of the crowd has become a valuable resource with a number of successful projects and services tapping human cognition, and scientists are looking to get even more out of the emerging field of human computation. A group of computer scientists, crowdsourcing pioneers, and visionaries led by Pietro Michelucci at the Human Computation Institute, has created a roadmap for research into human computation. The team, which met last year to discuss the increasing importance of crowdsourced cognition, believes human computation has considerable promise, as long as human cognition can be efficiently harnessed on a global scale. The scientists note human computation systems have been used to address some complex problems, and describe the kinds of projects they would like to see, such as an initiative to help people considering suicide or suffering from depression. Still, the researchers believe some significant issues must be addressed, including the ethical, legal, and social implications of human computation, and the optimal division of labor between humans and machines. They also call for a national center devoted to human computation.


New Website Can Identify Birds Using Photos
Cornell University (06/04/15)

Cornell University researchers have developed the Merlin Bird Photo ID, a bird photo identifier that can recognize 400 of the most commonly seen birds in the United States and Canada. The program is able to correctly identify the bird in the picture among the top three results about 90 percent of the time, according to Cornell researcher Jessie Barry. In addition, the website is designed to keep improving as more people use it. Users upload an image of a bird and tell Merlin where and when the picture was taken. Users also must draw a box around the bird and click on its bill, eye, and tail, which helps orient the computer vision system. Merlin uses artificial intelligence techniques combined with millions of data points submitted by users to determine the most likely species, including photos and sounds. "Computers can process images much more efficiently than humans--they can organize, index, and match vast constellations of visual information such as the colors of the feathers and shapes of the bill," says Cornell professor Serge Belongie. The system also relies on more than 70 million sightings recorded by bird watchers in the eBird.org database.


Code.org Inks 11 New Partnerships to Help Expand Computer Science Education
GeekWire (06/03/15) Taylor Soper

Code.org this week announced 11 new partnerships with organizations dedicated to helping the computer science education advocate reach even more schools than it could on its own. The new partners include Amplify Education, Beauty and Joy of Computing, Bootstrap, Codecademy, CodeHS, Globaloria, the National Math and Science Initiative, Project Lead the Way, Technology Education And Literacy in Schools, ScratchEd, and Tynker. Even before these new partnerships, Code.org worked with more than 70 of the largest U.S. school districts to help train teachers and develop computer science curricula. "As we expand in high school, we work region by region, and we can't do it all,” says Code.org co-founder Hadi Partovi. "We're leading a movement and we need partners to help." The new partners offer alternative ways to teach computer science, instead of relying solely on Code.org's programs. Although the new partners offer a range of new strategies for teaching computer science, the underlying theme is a commitment to adding computer science classes to schools. The new partnerships follow Code.org's agreement with the College Board and the U.S. National Science Foundation, both of which were announced earlier this year.


Rat Brain Cells Power a Computer
The Hindu (India) (06/04/15) Mohit M. Rao

Indian Institute of Science researchers have demonstrated that cultured rat brain cells on a glass plate can read signals from an infrared-enabled robot, process the problem of obstacles, and give an appropriate, accurate solution. The researchers took the cells and cultured them on a specialized glass plate covered with multiple electrodes that can detect small spikes in voltages generated by the cells. The cultured cells start to grow dendrites, and then the connected cells form a network that shows spontaneous electrical activity through tiny voltage spikes. The spikes are interpreted through a novel electronics platform that can detect and send electrical impulses to the cultured brain tissue through the embedded electrodes. Similar to the brain, the cultured tissue develops a coding system to decode the electrical spikes, and training is imparted through instructions coded as electrical spikes. The researchers conducted an experiment in which an infrared robot senses obstacles. The impulses are fed through the computer to the cells, which process the information, and the resulting voltage spikes, which represent commands for front, back, left, and right, are translated into codes for the robot. The experiment ran for 10 minutes, with obstacles moved at random, and the researchers discovered the robot was able to navigate successfully about 98 percent of the time.


Brain's Reaction to Certain Words Could Replace Passwords
Binghamton University (06/02/15)

Binghamton University researchers are studying how the way the brain responds to certain words could be used to replace conventional passwords. As part of the study, the researchers observed the brain signals, called brainprints, of 45 volunteers as they read a list of 75 acronyms. The researchers recorded the brain's reaction to each group of letters, focusing on the part of the brain associated with reading and recognizing words, and found that participants' brains reacted differently to each acronym. In fact, the reactions were sufficiently different that a computer system was able to identify each volunteer with 94-percent accuracy. The results suggest brainwaves could be used to verify a person's identity. Brain biometrics are an attractive security possibility because they are cancellable and cannot be stolen by malicious means, according to Binghamton University professor Sarah Laszlo. "In the unlikely event that attackers were actually able to steal a brainprint from an authorized user, the authorized user could then 'reset' their brainprint," she says. Although brainprint systems might not be mass-produced for low security applications, they could have other important security applications, notes Binghamton University professor Zhanpeng Jin.


MIT Cheetah Robot Lands the Running Jump
MIT News (05/29/15) Jennifer Chu

A robotic cheetah built by Massachusetts Institute of Technology (MIT) researchers is now the first four-legged robot to run and jump over obstacles autonomously. The team reports the robot has successfully cleared obstacles up to 18 inches tall, which are more than half of its own height, at an average running speed of five miles per hour. The robot uses a three-part algorithm to plan its path based on data from a visual system. The first component of the algorithm enables the robot to detect an obstacle and estimate its size and distance. The second component allows the cheetah to adjust its approach while nearing the obstacle, and the third determines its jumping trajectory. MIT professor Sangbae Kim calls the running jump a truly dynamic behavior. "You have to manage balance and energy, and be able to handle impact after landing," he says. Kim's team will demonstrate the cheetah's running jump at the U.S. Defense Advanced Research Projects Agency's Robotics Challenge in June.


App Provides Insight Into the Quantum World of Coupled Nuclear Spins
Technische Universität München (06/03/15)

Munich Technical University (TUM) researchers have developed an intuitive graphical representation of the information contained in matrices describing the properties of coupled nuclear spins in arbitrary quantum states. The new mathematical procedures control the behavior of nuclear spins in a targeted manner with maximum efficiency. The quantum properties of coupled nuclear spins are described mathematically using density matrices. "These are abstract columns of numbers that require very much experience to recognize the information contained within them," says TUM professor Steffan Glaser. The researchers have created the discrete representation of operators for spin systems process, which maps the density matrix onto three-dimensional drop-like objects reflecting all quantum mechanical interactions and entanglements between the spins at a given point in time. The researchers also created SpinDrops, an app that illustrates the creation, deformation, and rotation of spin-spin correlations under the influence of controllable magnetic fields in real time. "This program provides intuitive and comprehensible access to the fascinating world of quantum control theory for anyone dealing with the optimal control and utilization of quantum phenomena," Glaser says.


Stanford Scientists Show fMRI Memory Detectors Can Be Easily Fooled
Stanford News (06/03/15) Bjorn Carey

Researchers at Stanford University have developed software capable of reading a functional magnetic resonance imaging brain scan and determining with a high degree of accuracy whether the person is recalling a memory. This invention has sparked a great deal of interest in various sectors, from marketing to criminal justice. However, in a recent study, the researchers demonstrated that with only a few minutes of training, individuals can learn to trick the software into thinking they are recalling a memory when they are not, or that they are not recalling a memory when they actually are. The algorithm developed by researchers in the lab of professor Anthony Wagner looks for activity in parts of the brain associated with the formation and recall of memory and in previous studies has been able to detect a memory in the brains of cooperative subjects with 75- to 95-percent accuracy. However, the researchers wanted to see how it would perform against those actively trying to conceal their memory state. They found when subjects were given brief training and told to try to deceive the algorithm, its accuracy dropped to 50 percent, or that of random chance. Wagner says the result should cause people to reconsider the use of the technology in serious situations, such as legal cases.


Self-Programming Machines Next Phase of Computer Science: Wozniak
CIO (05/29/15) Rebecca Merrett

Speaking at the World Business Forum in Sydney, Australia this week, Apple co-founder Steve Wozniak said computers that learn in unsupervised ways and do not need to be told what to do in advance is the next phase of computer science. This kind of artificial intelligence (AI) is important to the future of computing because until now, humans have simply pre-programmed computers to carry out certain functions and tasks, according to Wozniak. Although computer power and speed has advanced dramatically in recent years, simply having computers complete computations faster is not the same as developing a method for solving a problem on its own. However, Wozniak does have concerns about AI's potential implications on humanity, fears that are echoed by other technology leaders such as Stephen Hawking, Elon Musk, and Bill Gates. "Once machines have independent thought and can program themselves, that's the turning point," Wozniak says. He also notes hopefully computers will only reach human level, so they will never be smarter or better than mankind. Wozniak observes various AI advances are bringing the industry closer to where computers seem to be doing the things that we call thinking.


UT Arlington Computer Scientist's Research to Better Control Cyber-Physical Systems
University of Texas at Arlington (06/03/15) Herb Booth

University of Texas at Arlington professor Taylor Johnson recently received a two-year grant from the U.S. Air Force Research Laboratory to model, predict, monitor, and control potential emergent behavior in cyber-physical systems. Emergent properties are those that spontaneously emerge in the execution of engineered systems, especially when those systems have software and physical components. "We need methods to determine for a given system, if emergent behaviors are possible, could an attacker exploit these emergent behaviors at the software level or insert physical signals that would lead to global specifications being violated, such as drones colliding?" Johnson says. The work aims to make cyber-physical systems more stable and predictable. "Future work in artificial intelligence, satellite navigation, and unmanned vehicles will benefit greatly from Dr. Johnson's findings," says University of Texas at Arlington College of Engineering dean Khosrow Behbehani. Specifically, the research seeks to define emergent behavior rigorously using formal specifications, verify the existence or absence of emergent behavior with formal verification, control to achieve desired emergent behavior, and evaluate the findings in a case study.


NYU Researchers: One Big-Data Picture Is Worth a Thousand Words on Human Rights
New York University (06/01/15)

New York University (NYU) researchers are developing state-of-the-art visualization tools that they say will enable human rights advocates to convey complex information and vividly display their view of a better world. The researchers recently received a $250,000 MacArthur Foundation grant, which will allow them to expand upon their pilot study and provide several human rights organizations with hands-on technical assistance and training. "Our pilot research revealed a low level of statistical literacy within the human rights advocacy world, and there remains a lack of understanding of how both data and data visualization could be deployed most effectively to support human rights messages," says NYU professor Oded Nov. The researchers feel the human rights community is eager to embrace data-driven analysis while making sure the focus remains on the human story. "There has been a recent explosion of data relevant to important social and economic issues," says NYU professor Enrico Bertini. "If all that data can be visualized, simply and efficiently, organizations can identify patterns and trends and harness them to tell powerful stories that the public needs to hear. We can help get the word out about areas without access to clean water, which is a basic human right, or map civilian casualties in war-torn areas."


Developer Tracks Real-Time Locations of Facebook Messenger Users
IT Pro (05/29/15) Joe Curtis

Harvard University researcher Aran Khanna has developed Marauders Map, a browser extension that allows users to discover the identity of Facebook Messenger users, their location, and previous movements. The Marauders Map exploits Facebook's default location settings on iOS and Android, which users must manually disable, and uses a global positioning system to place users on a map. "By simply looking at the cluster of messages sent late at night you can tell exactly where his [another user's] dorm is, and in fact approximately where his room is located in that dorm," says Khanna. By examining messages from the past few days, the Marauders Map can develop a profile of users' weekly schedules and predict where they might be at any given time. In addition, users can track the location of others even if they are not friends if the user had sent messages to a group chat. "With this code you can see for yourself the potentially invasive usage of the information you share, and decide for yourself if this is something you should worry about," Khanna notes.


Software Designed to Plan NASA Space Missions Applied to Online Advertising
PhysOrg.com (05/29/15) Rob Matheson

A Massachusetts Institute of Technology startup is turning marketing into an exact science, claiming it can accurately measure the return on investment in online marketing for clients. Boston-based DataXu has developed software that can quickly analyze user data--such as previously viewed advertisements and video--and data from third-party providers on a user's interests, demographics, and purchase history. The software was initially designed to help the U.S. National Aeronautics and Space Administration plan missions to Mars, and the analysis it generates can help map out the tastes and buying intentions of customers. In less than one millisecond, the software can determine which advertisement from a pool of clients will most likely ensure that a user will click through or eventually make a purchase, while calculating the cost of the advertisement. Moreover, the software can learn patterns for ads that work best for certain users or groups of users. "We're helping companies compete in a world that's full of data," says DataXu co-founder Bill Simmons. The software makes about 1.5 million decisions per second to decide if an ad should be placed before a specific user via the process of real-time bidding. DataXu chooses from a pool of about 100,000 ads provided by some 1,000 clients, to select the ad for each opportunity that will best fit the user.


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