Welcome to the December 14, 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."
|
|
|
The AI Boom is Happening All Over the World, and It's Accelerating Quickly
The Verge
Nick Statt
December 12, 2018
A group of experts that includes members of Harvard University, Stanford University, Open AI, and the Partnership on AI industry consortium has released the second annual AI Index. The goal of the Index is to measure the field's progress using hard data and to try to make sense of that progress as it relates to a range of subjects. The latest report finds that commercial and research work in artificial intelligence (AI), as well as funding, is exploding all over the world, with an especially high concentration in Europe, China, Japan, and South Korea. Europe is the largest publisher of AI papers, with 28% of all AI-related publications last year, followed by China at 25%; North America is responsible for 17% of AI-related publications. In terms of the type of AI activity, the report finds that machine learning and probabilistic reasoning are the leading research categories by the number of published papers.
|
Foldable Drone Can Navigate Narrow Holes
UPI
Brooks Hays
December 12, 2018
A drone that can compress itself to pass through narrow spaces and then revert to its previous shape, all while continuing to fly, has been developed by researchers at the University of Zurich in Switzerland and the Swiss Federal Institute of Technology, Lausanne (EPFL). The quadrotor drone has four propellers that rotate independently, mounted on mobile arms that can fold around the drone’s main frame. The drone also is equipped with a control system that adapts in real time to any new position of the arms, adjusting the thrust of the propellers as the center of gravity shifts. Said the University of Zurich’s Davide Falanga, "The final goal is to give the drone a high-level instruction such as 'enter that building, inspect every room and come back' and let it figure out by itself how to do it.”
|
How Do You Feel What You Can't Touch? Scientists Crack the Nerve Code
Sydney Morning Herald (Australia)
Liam Mannix
December 12, 2018
Researchers at Neuroscience Research Australia are exploring the writing of computer code that can be fed into the human nervous system to mimic any desired tactile sensation. Nerves send data to the brain in encrypted code, and Neuroscience Research Australia's Ingvars Birznieks thinks cracking this code could bring advanced prosthetics, brain-computer connections, and other sensory advances closer to reality. Birznieks experimented with such possibilities by inserting a tiny electrode into the nerves in volunteers' wrists, to monitor the data transmitted to the brain; he then used a robot to stimulate volunteers' fingertips and evoke various sensations. The electrode captured the binary "spike" neural signals, with the data encoded in the frequency of the spikes. Birznieks' team fed the code into an artificial intelligence trained to break encryption, and found the timing between the spikes also contained information. Said Birznieks, "We can now recreate these spike codes—or design our own—as we wish."
|
Mapping the Brain's Genetic Landscape
The New York Times
Benedict Carey
December 14, 2018
More than a dozen research centers and many specialists in cell biology, genetics, and bioinformatics have applied advanced computer techniques to huge datasets to plot out the most detailed computer model of the human brain genome to date. The model was partly based on analyses of almost 2,000 brains, from people with and without diagnoses. In one stage, Yale University researchers combined the analyses and calculated the combined added risk they impart into a single coefficient. The model can be searched in a manner similar to a Google Earth-like database, charting specific genetic sites in the brain and identifying active, localized genetic and biological processes, and at what point they occur during maturation. Experts think the model could be used to help clarify the genetic causes of schizophrenia, autism, and other disorders.
*May Require Paid Registration
|
Computer Chip Vulnerabilities Discovered by WSU Researchers
Washington State University
Tina Hilding
December 13, 2018
Washington State University (WSU) researchers found significant vulnerabilities in high-performance computer chips that could lead to failures in modern electronics. The researchers discovered they could damage the on-chip communications system and reduce the chip's lifetime by deliberately adding malicious workloads. The team developed three "craftily constructed deleterious" hacks to test the communications system; the extra workload augmented electromigration-induced stress and crosstalk noise. The team found a limited number of crucial vertical links, which connect the processors in a stack and let them to communicate with each other, were prone to failure. WSU's Partha Pande said, "The role of the communications and the threat had not been clear to the research community before." The researchers aim to ameliorate the vulnerability, with possibilities including automated techniques and algorithms to detect and foil attacks.
|
How Computer Models Can Help Stand-Up Comics
The Boston Globe
Kelly Kasulis
December 13, 2018
Researchers at the University of Alberta in Canada tapped a previous study on human-supplied humor ratings for about 5,000 English words, to run statistical computer models to predict the humor levels of another 45,000. The researchers learned that a word's funniness potentially can be predicted by factoring in what it means, and how it looks and sounds. The team found many of the funniest words relate to things like animals, sex, bodily functions, insults, and curses, while words with certain structures, like a "k" or double "o" sound, were more likely to be funny. Said Alberta's Chris Westbury, "Our models are complicated. They have a lot of predictors in them. It's a first step toward quantifying humor, and that's important because it shows that none of the theories we have [for humor] are good enough."
*May Require Paid Registration
|
Amazon Expands Future Engineer Program to Younger Grades
THE Journal
Dian Schaffhauser
December 12, 2018
Amazon has expanded its "Future Engineer" initiative into kindergarten through eighth grades, offering free online classes and funding summer camps to help students discover the fun of computer science (CS). Previously, the program was open to 100,000 high school students taking Advanced Placement CS courses, and awarding four-year scholarships and internships to learners from underrepresented populations. Amazon aims to fund the latest CS camp scholarships via partnerships with Code.org and Coding with Kids, giving underprivileged schoolchildren the opportunity to learn programming in a hands-on, interactive manner. Non-camp courses are being offered through edhesive, a company committed to providing science, technology, engineering, and math education online.
|
The Making of a Computer-Generated Influencer
The Wall Street Journal
Yoree Koh; Georgia Wells
December 13, 2018
Miquela Sousa is a computer-generated (CG) character created by Brud, a Los Angeles-based startup, which aims to turn her and an expanding group of CG-based social media personalities into a cast of characters that can evolve with the times without aging. Over the past decade, social media "influencers" have become online celebrities through daily postings about a range of topics. These influencers also have become highly sought-after brand ambassadors. Miquela Sousa is a test of the power of social media celebrity; launched on Instagram in 2016, the character posts carefully composed images in a lifelike way, and did not identify herself as artificial until April, as part of a staged drama that played out over a series of posts. Said Cyan Banister of Founders Fund, which has invested in Brud, “You can create the Kardashians without any of the inherent issues that come with being human.”
*May Require Paid Registration
|
Radical New Neural Network Design Could Overcome Big Challenges in AI
Technology Review
Karen Hao
December 12, 2018
Researchers at the University of Toronto and the Vector Institute in Canada have redesigned neural networks without traditional stacked layers of simple computational nodes that work together to find patterns in data; the new design replaces the layers with calculus equations. The researchers, who dubbed this new design an ordinary differential equations (ODE) solver, said it can model continuous change, and changes certain aspects of training for neural networks. In a traditional neural network, the user has to specify the number of layers at the start of the training, then wait until training is done to find out how accurate the model is. The new method lets the user specify their desired accuracy first, and the network then will find the most efficient way to train itself within that margin of error.
|
NSF Offers Extra Access to Academic Supercomputers
Government Computer News
Susan Miller
December 12, 2018
The U.S. National Science Foundation (NSF)'s Jim Kurose has invited science and engineering researchers to apply for supplemental funding for accessing two leadership-class academic supercomputers from April through December 2019. One system is the University of Illinois at Urbana-Champaign's Blue Waters, a Cray XE supercomputer with more than 22,000 AMD processors, 4,000 Nvidia graphics processor units, 26 petabytes of usable online storage for fast access, and 380 petabytes of usable near-line tape storage. Kurose said 125 million node hours will be available to researchers with active NSF awards who require significant computation resources to address grand challenge problems, with five to six requests approved. Meanwhile, NSF will offer 15 to 20 early-science teams 34 million node hours on the Frontera supercomputer of the Texas Advanced Computing Center at the University of Texas at Austin. Frontera is a Dell EMC machine with more than 16,000 Intel Xeon processors.
|
The Privacy Risks of Compiling Mobility Data
MIT News
Rob Matheson
December 7, 2018
Massachusetts Institute of Technology (MIT) researchers found the growing practice of compiling massive, anonymized datasets about people's patterns of movement, while offering significant behavioral insights, could jeopardize people's privacy. The team utilized a statistical model to conduct an analysis of user "matchability" in two large-scale datasets from Singapore—one from a mobile network operator and one from a local transportation system. The model tracked location stamps of users in both datasets, gauging the likelihood that data points in both sets originate from the same individual. The model successfully matched about 17% of individuals in a week's worth of data, and more than 55% in a month's worth. MIT's Carlo Ratti said, "All data with location stamps (which is most of today's collected data) is potentially very sensitive and we should all make more informed decisions on who we share it with."
|
Quantum Network Joins Four People Together for Encrypted Messaging
New Scientist
Chelsea Whyte
December 12, 2018
Researchers at the University of Vienna in Austria have designed a network that uses quantum key distribution (QKD) to secure messages among four users. Although QKD systems often depend on the polarization, or orientation, of a photon, the Vienna researchers tapped the frequency of the photon's light, which can be tuned to permit connections with several other users. The team used a laser to entangle photons, then divided the produced light spectrum into 12 separate channels that each had a different frequency. The four users were linked via fiber-optic cables that each held three channels, making a central hub for routing communications unnecessary. Said Vienna's Robert Ursin, "If you want to scale the network up, you just add the hardware required for one of the users. It’s much like the classical internet where everyone gets a router."
*May Require Paid Registration
|
Microsoft's Investment Arm M12 Announces Winners of $4M Female Founders Competition
GeekWire
Taylor Soper
December 11, 2018
M12, Microsoft's investment arm, recently announced the winners of its inaugural Female Founders Competition, which awarded a total of $4 million to two women-led startups: Acerta, an artificial intelligence (AI) startup; and Mental Canvas, a digital drawing company. Each company will receive a $2-million investment and access to other resources. Acerta provides an AI platform that ensures quality and reliability of vehicle systems throughout the vehicle lifecycle. Mental Canvas re-imagines sketch for the digital age by augmenting it with spatial strokes, three-dimensional navigation, and free-form animations. M12's push to recognize female founders comes as Microsoft faces ongoing gender discrimination lawsuits brought by current and former female employees.
|
|
