Welcome to the January 22, 2018 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).

To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version. The online version now has a button at the top labeled "Show Headlines."

Bitcoin wrapped in electrical wires There Is Nothing Virtual About Bitcoin's Energy Appetite
The New York Times
Nathaniel Popper
January 21, 2018

The computer power required to create one bitcoin is equivalent to at least as much electricity as the average U.S. household expends in two years, according to financial services firm Morgan Stanley and economist Alex de Vries. The power consumption of computer systems plugged into cryptocurrency networks has climbed with the currencies' appreciation, and advocates argue the energy cost is warranted to secure the networks and support a financial infrastructure free from bank or government interference. The 12.5 bitcoins currently handed out every 10 minutes are valued at about $145,000, but de Vries estimates each bitcoin transaction needs 80,000 times more electricity to process than a Visa credit card transaction. Supporters say the complexity of the mining competition is intentional so no one can dominate the accounting and alter the transaction records, making hacking less likely. However, critics who view cryptocurrency as a speculative bubble warn it contributes to global warming and waste without any real benefits.

Full Article
*May Require Free Registration
Fewer International Students Coming to U.S. for Grad School in Science and Engineering
PRI's The World (MA)
Carolyn Beeler
January 18, 2018

Foreign-born student enrollment in graduate science and engineering programs in the U.S. slipped 6 percent in 2017 after years of gains, and almost all of that decrease was concentrated in computer science and engineering programs, according to a U.S. National Science Board report. The steepest decline was among Indian students, whose enrollment numbers dived 19 percent last year. The National Science Board's Geraldine Richmond notes more than half of technical U.S. graduate students are international, and their presence is important for countering a multi-year erosion of U.S.-born students in science, technology, engineering, and math graduate programs. "We have a research engine that needs to be fueled, and that fuel is really our graduate students," Richmond says. Possible factors behind the enrollment decline include international students' fading hopes of post-graduation employment due to less-friendly U.S. immigration policies. Meanwhile, international student enrollment in Canadian and Australian colleges and universities has risen in the last year.

Full Article

Close up of piano keys Computer Codes Make Sweet Music for Self-Playing Piano
University of Nottingham (United Kingdom)
Jane Icke
January 17, 2018

Researchers at the University of Nottingham, U.K., Mixed Reality Laboratory say they have used unique software to create an interactive self-playing piano performance called "Climb." They note this unique musical composition combines contemporary piano with computer game elements to create a non-linear piece of music accompanied by graphics in which the user ascends a mountain, choosing their own path while dealing with weather, animals, and other obstacles. The system is comprised of 23 fragments or musical events the software plays in different orders. "These fragments are like pieces of a puzzle that can be put together in different ways to create different pieces of music for the pianist to play and the audience to experience," says Nottingham's Adrian Hazzard. In addition, he notes the software can trigger duets and challenges, with the interactive system mimicking a virtual partner who may jump to new points in the score.

Full Article

blank canvas on an easel  and cups of paint brushes in background Microsoft's New Drawing Bot Is an AI Artist
Sarah Perez
January 18, 2018

Microsoft recently debuted a "drawing bot" that can generate images from text descriptions of an object, while adding details not included in the text. The algorithm applies two machine-learning models, including one to create images from text descriptions, and another that employs those descriptions to evaluate the authenticity of the produced images. The models essentially collaborate to create higher-quality images, says Xiaodong He at Microsoft's Deep Learning Technology Center. He notes the bot was fed datasets of paired images and captions so it could understand how to match up words to images; for example, he says it learned to draw a bird when the text says "bird," as well as what an image of a bird should look like. "That is a fundamental reason why we believe a machine can learn," He notes. The bot also can compose images from more complex sentences, as well as filling in gaps when specific details are not provided.

Full Article
New Blueprint for Converging HPC, Big Data
John Russell
January 18, 2018

A group of high-performance computing (HPC) experts has released the Big Data and Extreme-Scale Computing Pathways to Convergence Report as a blueprint for aligning computational infrastructures so HPC and big data are supported. The experts cite big data's role in worsening paradigm splits between traditional "HPC and high-end data analysis" and "stateless networks and stateful services" delivered by end systems. The report says addressing them requires new standards governing the interoperability between data and compute, based on a new, common, and open Distributed Services Platform (DSP). The report's recommendations for decentralized edge and peripheral ecosystems include creating a new hourglass DSP framework, targeting workflow patterns for better data logistics, designing cloud stream processing capabilities for HPC, advocating a scalable approach to Content Delivery/Distribution Networks, and developing software libraries for common intermediate processing tasks. Among the report's actionable conclusions for centralized facilities are recognizing the need to adjust HPC architectures to accommodate machine-learning-driven scientific workloads.

Full Article
New C-BRIC Center Will Tackle Brain-Inspired Computing
Purdue University News
Emil Venere
January 16, 2018

Purdue University will lead the Center for Brian-inspired Computing Enabling Autonomous Intelligence (C-BRIC), a new initiative for developing brain-inspired computing for intelligent autonomous systems such as drones and personal robots. "The center's goal is to develop neuro-inspired algorithms, architectures, and circuits for perception, reasoning, and decision-making, which today's standard computing is unable to do efficiently," says Purdue professor Kaushik Roy, who will lead C-BRIC. He notes C-BRIC also wants to narrow the efficiency gap between high-performance computers, which consume hundreds of thousands of watts of power, and the human brain, which requires only about 20 watts. "C-BRIC will develop technologies to perform brain-like functions with brain-like efficiency," says Purdue professor Anand Raghunathan. The center also will enable next-generation autonomous intelligent systems that can both accomplish end-to-end functions and complete mission-critical tasks without human assistance. The researchers will attempt to reach their goals by improving the theoretical and mathematical frameworks of neuro-inspired algorithms.

Full Article
Professor Designs Algorithms to Make Robots Better Collaborators
Northeastern University News
Jason Kornwitz
January 17, 2018

Researchers at Northeastern University are developing algorithms that enable groups of robots to make good decisions despite inherent communication delays in information sharing. The delays can be detrimental to robotic group behavior, especially if the robots have access to and can act only on "outdated" information. However, the researchers are working to overcome that hurdle via a protocol that can tell robots in which direction to move to perform a collaborative task using only outdated information. The team has developed an algorithm that can improve and accelerate cooperation among three robots with information-sharing delays, and they also are using control theory and human factors engineering to design algorithms that would enable self-driving cars to collaborate with human drivers and each other. In addition, Northeastern professor Rifat Sipahi is working with colleagues at the University of South Florida and Bilkent University in Turkey on an algorithm that would make it easier for a human in one location to control a robot in another location.

Full Article
Trash Talk From an AI Could Help Humans Cooperate Better
Chelsea Gohd
January 18, 2018

Researchers at Brigham Young University (BYU) have developed an algorithm to help artificial intelligence (AI) and humans learn to cooperate. They used 472 games requiring two-player interactions to promote communication between humans and machines. The researchers found light banter, or "cheap talk," doubled the rate of cooperation between AI and human players. Similar to how humans might bluff their way through a game, the AI also was capable of being "all talk." For example, it might say that it was going to make a certain move, only to decide not to follow through with that choice, notes BYU's Jacob Crandall. However, in order to truly work well together, AI must be able to understand and respond to human emotions via verbal and facial cues, cultural practices, and social norms. The researchers say achieving this level of cooperation will likely become a larger focus for AI research in the coming years.

Full Article

Robot head with exposed brain attached to wires New Research Adds Physiology to Computer Models
Penn State News
Erin Cassidy Hendrick
January 15, 2018

Researchers at Pennsylvania State University (PSU) are studying how to make computers work more like human brains, and contend without a physiological complement, a computer program modeling the mind would just be a brain in a vat. The researchers' reasoning for exploring this new theory is that a body's natural needs, state, and inclinations inherently influence the brain. To study this idea, the researchers have programmed a sense of bodily processes within a computer simulation, and are focusing on how the mind responds to physical fatigue. The team is studying this topic through the framework of ACT-R/Phi, which combines a theory of how the human brain works (ACT-R) with a theory of how the body works. The combined ACT-R/Phi program enables the researchers to determine how a human would react in certain situations. Bucknell University professor Christopher Dancy says this strategy offers "a fuller picture of how our normal physiological functions modulate our actions and decisions."

Full Article
Designing Customizable Self-Folding Swarm Robots
IEEE Spectrum
Evan Ackerman
January 18, 2018

Researchers at the University of California, San Diego say they are developing robotic swarms called bristlebots that can be rapidly customized, self-assembled, and self-deployed without human intervention. They note the robots are laser-cut from flat sheets, can fold themselves up, and then move on their own with only the slightest amount of help from humans. The team says the robots are more reliable than previous iterations of similar technology and have the potential to be easily customized for speed, maneuverability, and payload. The bristlebots use vibration motors to shimmy themselves along flat surfaces at up to 23 centimeters a second in a straight line, and the motion characteristics of these bristlebots can be adjusted by changing the length of their legs, making them a good platform to experiment with design changes. "Future work will investigate how the robot structure can be tailored to a given payload or environment," according to the researchers.

Full Article
New Center Headquartered at Carnegie Mellon Will Build Smarter Networks to Connect Edge Devices to the Cloud
Carnegie Mellon News
Sherry Stokes
January 15, 2018

The Computing on Network Infrastructure for Pervasive Perception, Cognition, and Action (CONIX) Research Center headquartered at Carnegie Mellon University (CMU) will spearhead a Semiconductor Research Corporation five-year effort to embed more intelligence within computer networks by developing a framework for networked computing residing between edge devices and the cloud. "CONIX will develop novel architectures for large-scale, distributed computing systems that have immense implications for social interaction, smart buildings, and infrastructure, and highly connected communities, commerce, and defense," says CMU's James H. Garrett Jr. CONIX's research will focus on smart and connected communities, interactive mixed reality, and augmented situational awareness at the edge. The first focus area will investigate instruments for managing and processing millions of sensor feeds in urban environments, the second will combine physical and virtual reality solutions into a collaborative digital teleportation system, and the third will develop on-demand information feeds for decision-makers via aerial drone swarms.

Full Article

D. Fox Harrell, the director of the Imagination, Computation, and Expression Laboratory at MIT’s Computer Science and Artificial Intelligence Laboratory. D. Fox Harrell on His Video Game for the #MeToo Era
MIT News
Larry Hardesty
January 19, 2018

In an interview, Massachusetts Institute of Technology professor D. Fox Harrell discusses the development of Grayscale, a video game designed to make players sensitive to workplace sexual misconduct. Harrell says gamers play employees who receive emails from co-workers with embedded evidence of various kinds of sexism from the Fiske and Glick social-science model. He notes how players respond to these emails is evaluated to assess their performance in navigating "tensions between what is seen as the corporate culture, what would get you ahead, and your own personal thoughts about the sexism that's displayed." Harrell says Grayscale is based on the ICE Lab's Chimeria computational platform, which enables "people to be members of multiple categories or to have gradient degrees of categories and have those categories change over time." Harrell notes the developers want Grayscale to offer "a compelling narrative" that entails understanding co-workers' back stories and personalities.

Full Article
Here's to the Women of Eniac for Giving Us Modern Programming Tools
ACM Publications

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]