Welcome to the March 10, 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).

Representation of a neural network Brainlike Computers Are a Black Box. Scientists Are Finally Peering Inside
Jackie Snow
March 7, 2017

Researchers at the Heinrich Hertz Institute in Germany say they have developed a method for examining neural networks as they operate and visualizing how they reach conclusions. The Institute's Wojciech Samek and colleagues produced software that goes backward through neural nets to see where a certain decision was reached, and to what degree the decision shaped the results. In this way, researchers can quantify how much individual inputs influence the software's overall conclusion, assigning them a score. Using this information, scientists can visualize a superimposed mask highlighting the most influential areas. Samek says this technique could help cut the amount of data required to train neural nets, and probe errors when they occur in results. Harvard University's Sara Watson stresses ongoing research in this area is essential as the number of decisions algorithms make in people's daily lives continues to grow.

Full Article
IBM Fits a Bit on an Atom, Eyeing Ever-Smaller Devices
IDG News Service
Stephen Lawson
March 8, 2017

Researchers at IBM have used a single atom to write and read one bit of data via magnetism, which could lead to storage devices with hundreds of times greater density than currently available solutions. The researchers note much denser storage could significantly miniaturize computing devices and data centers. IBM's Christopher Lutz says solid-state magnetic storage does not require atomic manipulation, and his team found after storing the bit that two atoms could be separated by only one nanometer and still be read independently, which makes storage of about 600 terabits per square inch possible. Researchers could use IBM's breakthrough to develop new high-density storage that functions outside a lab, likely using a small number of atoms that can remain stable at room temperature. Lutz notes the study of data density and miniaturization is important as chip fabrication hits its physical limits.

Full Article

A grouping of microbes Counting Microbes on a Smartphone
A*STAR Research
March 8, 2017

Researchers at the Agency for Science, Technology and Research (A*STAR) Bioinformatics Institute and James Cook University Singapore have developed the APD Colony Counter App suite, an Android-based smartphone app they say provides a cost-effective and accurate alternative to the traditionally labor-intensive task of counting bacterial colonies on a culture. The researchers developed an algorithm that provides the most accurate colony counts by creating better contrasts for colony detection than other image-processing algorithms. They currently are using the app to streamline their microbial analysis service, which they conduct for other scientists. However, the app also is available for use on data outside of the A*STAR lab. The researchers say the program could bring more flexibility to researchers around the world. "By having apps that anyone can access anywhere, I'm hoping that we're going to bring back the spatial freedom for scientists to make discoveries anytime, anywhere," says A*STAR researcher Samuel Gan.

Full Article
Baidu's Artificial Intelligence Lab Unveils Synthetic Speech System
Technology Review
March 8, 2017

Baidu's artificial intelligence research lab has developed Deep Voice, a text-to-speech system that can learn to talk with little to no human interference in a matter of hours. Most text-to-speech tools involve recording a large database of speech from one individual and then rearranging the utterances into new phrases. The Baidu researchers say Deep Voice employs deep-learning methods to transform text into phonemes, and then applies a speech synthesis network to replicate the sounds. Each stage of the process follows deep learning, so once trained, Deep Voice has little need for human modification. Deep Voice has no control over stresses on the phonemes, their duration, or the natural frequency of the sound, enabling Baidu to change the voice of the speaker and the emotion the word evokes. The researchers say real-time speech synthesis is possible with Deep Voice, which can be quickly re-educated on new datasets with no human involvement.

Full Article
New Material Makes It Possible to Record Data With Light
Tamara Besedina
March 10, 2017

Researchers at ITMO University in Russia, Leipzig University in Germany, and the Eindhoven University of Technology in the Netherlands have developed a method, involving changing the light parameters, to generate excitons that are fully controllable and able to record information at room temperature. Excitons act as a transitional form between photons and electrons, and could be used to create compact optoelectronic devices for rapid recording and processing of an optical signal. The new approach to information recording involves changing the distance between the crystal layers of metal-organic frameworks (MOF) to switch the interlayer excitons "on" and "off." The researchers locally heated the crystal with a laser, and in the place of exposure the layers stuck together and the luminescence of excitons disappeared while the rest of the crystal continued to shine. Using this process the researchers were able to record one bit of information and keep it for several days.

Full Article
Textile-Based Wearable Electronics and Fashion Displays
March 3, 2017

Researchers at the Korea Advanced Institute of Science and Technology (KAIST) have developed clothing-based light-emitting devices using organic light-emitting diodes (OLEDs). The KAIST team demonstrated reliable fabric-based OLEDs and high-luminance fiber-based polymer light-emitting diodes (PLEDs). The fabric-based OLEDs are encapsulated by flexible, transparent, multilayer barrier films, and were designed for long-term reliability, with measured operational lifetimes exceeding 1,000 hours under conditions of 30 degrees Celsius and 90-percent relative humidity. In addition, the researchers successfully operated the OLEDs on rough fabrics through the thermal lamination of thin planarization sheets. The fiber-based PLEDs were developed using a dip-coating method to coat polymer layers concentrically onto fibers. "This technology will accelerate the commercialization of fiber-based wearable displays because it offers low-cost mass production using roll-to-roll processing, a type of technology applied to create electronic devices on a roll of flexible plastic or metal foil," says KAIST professor Kyung Cheol Choi.

Full Article

A simulation of an autonomous car preparing to start Sharing the Fares
MIT News
Peter Dizikes
March 6, 2017

The potential for urban ride-sharing is similar for a variety of cities all over the world, according to a study co-authored by researchers at the Massachusetts Institute of Technology (MIT). The study tracked scores of taxis in New York, San Francisco, Singapore, and Vienna and found more than 90 percent of rides exceeding five minutes in duration could be easily shared. "We found this kind of global law of ride-sharing," says Carlo Ratti, director of MIT's Senseable City Lab. The researchers monitored more than 13,000 taxis for 12 months in New York City, using factors such as a city area, the number of taxi trips, and average traffic speed to determine what proportion of rides are shareable, given some modest rules about the flexibility passengers might have in terms of waiting for a ride. The researchers found more than 99 percent of trips in New York are shareable.

Full Article
The AI Debate Must Stay Grounded in Reality
Prospect Magazine
Vincent Conitzer
March 6, 2017

Addressing concerns about the long-term societal impact of rapidly advancing artificial intelligence (AI) technologies should be the focus of a balanced debate that includes the participation of AI researchers who tend to avoid long-term speculation in favor of more immediate issues of concrete technical progress, writes Duke University professor Vincent Conitzer. "Research communities work best when they include people with different views and different sub-interests," he says. Conitzer notes in addition to traditional AI scientists, the debate is attracting economists concerned with AI-fueled unemployment, legal scholars focused on regulation of autonomous vehicles and other technologies, and philosophers studying AI's moral and ethical ramifications. "While I am quite skeptical of the idea that truly human-level AI will be developed anytime soon, overall I think that the people worried about this deserve a place at the table in these discussions," he says.

Full Article
Smartphone Interruptions: Are Yours Relentless and Annoying?
Rutgers Today
Todd B. Bates
March 6, 2017

Researchers at Rutgers University have developed a model that can predict a person's receptiveness to smartphone interruptions by incorporating personality traits that could facilitate better notification management. "Preferably, your smartphone would recognize your patterns of use and behavior and schedule notifications to minimize interruptions," says Rutgers professor Janne Lindqvist. He co-created the two-stage predictive model which first measures user availability, then gauges user interruptibility. The researchers culled more than 5,000 smartphone records from 22 individuals over four weeks, and predicted how busy they were. The team used major personality traits to help anticipate interruptibility, and found users' willingness to be interrupted is influenced by their location. Lindqvist says his team's next focus will be to optimize the predictive model to tailor smartphones to user preferences. The researchers will present their study in May at the ACM CHI Conference on Human Factors in Computing Systems (CHI 2017) in Denver, CO.

Full Article

A woman’s hands on a computer keyboard CODeLLA Aims to Teach Latina Girls Another Vital Language: Coding
NBC News
Carmen Pelaez
March 8, 2017

The Miami, FL-based CODeLLA project was launched in 2013 to encourage Latina girls ages 8-12 to explore careers in science, technology, engineering, and math (STEM). CODeLLA is an eight-week coding and technology entrepreneurship immersion program that aims to teach the girls programming skills, as well as enable them to see a future in STEM fields. "Learning to code ascertains our girls are fluent in one of the most widely used languages of today's world," says CODeLLa co-founder Josie Goytisolo. Last fall, CODeLLa held the first She Innovates Tech conference, which aimed to inspire the next generation of innovators by connecting them with female technology leaders such as Alicia Abella, assistant vice president of Cloud Technologies at AT&T, who was a keynote speaker at the conference. "I try and get the girls to recognize that they can actually change the world by getting a science and engineering degree," Abella says.

Full Article
Artificial Intelligence and Robots to Make Offshore Wind Farms Safer and Cheaper
University of Manchester
March 9, 2017

A consortium led by the University of Manchester in the U.K. is investigating artificial intelligence (AI) and other technologies to enhance offshore wind farms. The Crown Estate estimates up to 90 percent of the cost of offshore operation and maintenance goes to the need for site access in remote and hazardous environments. Reducing the cost and effort of such maintenance is the goal of a project to probe the use of sensors, robots, virtual reality models, and AI. Predictive and diagnostic methods will enable early problem detection, while robots and advanced sensors will minimize human intervention. "The U.K. has world-leading expertise in the technologies and science in this area, but they have often operated separately," says Manchester professor Mike Barnes. "The U.K. Engineering and Physical Sciences Research Council has supported this project to bring them together for the first time to make a real step change in this industry."

Full Article
Robot Uses Social Feedback to Fetch Objects Intelligently
News from Brown
Kevin Stacey
March 6, 2017

Researchers at Brown University have developed an algorithm that helps robots get better at retrieving objects. The new algorithm enables the robot to quantify how certain it is that it knows what a user wants. When the certainty is high, the robot will hand over the object. If the certainty is lower, the robot guesses about what the user wants and then asks for confirmation by hovering its gripper over the object and asking, "this one?" The question enables the robot to make important inferences based on the answer. The researchers tested the system by asking participants to interact with a Baxter robot. The team set Baxter to never ask questions, ask a question every time, or to ask questions only when uncertain. The tests found asking questions with the new algorithm was better in terms of accuracy and speed compared to the other two conditions.

Full Article
Facebook's AI Chief: Machines Could Learn Common Sense From Video
Technology Review
Tom Simonite
March 9, 2017

In an interview, Yann LeCun, director of Facebook's artificial intelligence (AI) research group, says software could learn common sense by studying video. "One of the things we really want to do is get machines to acquire the very large number of facts that represent the constraints of the real world just by observing it through video or other channels," LeCun says. "That's what would allow them to acquire common sense, in the end." Of particular interest to researchers is predictive learning, in which a system attempts to anticipate future actions from a few frames of video. "If we can train a system to do this we think we'll have developed techniques at the root of an unsupervised learning system," LeCun says. "That is where...a lot of interesting things are likely to happen. The applications for this are not necessarily in vision--it's a big part of our effort in making progress in AI."

Full Article
The VR Book
ACM Online Courses

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: [email protected]