Welcome to the June 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."
|
|
|
Is There a Smarter Path to Artificial Intelligence? Some Experts Hope So
The New York Times
Steve Lohr
June 20, 2018
More artificial intelligence (AI) experts are questioning the validity of deep learning, with some warning that relying on this training process eventually will cause AI to hit a technical limit. They also say deep learning's progress is mainly constrained to fields where vast datasets are available and tasks are well defined, while intelligence sectors such as meaning, reasoning, and common-sense knowledge remain out of reach. Some scientists are pursuing alternative AI approaches that bypass deep learning, with the goal of realizing broader, more flexible intelligence that requires less data consumption. The U.S. Defense Advanced Research Projects Agency has proposed a project to fund academic research and set up a noncommercial network for exchanging ideas on technology to reproduce human common-sense reasoning.
*May Require Paid Registration
|
China Pioneers Ceramic 3D Printing in Microgravity
Xinhua
June 19, 2018
Scientists at the Chinese Academy of Sciences' (CAS) Technology and Engineering Center for Space Utilization in Beijing have successfully used three-dimensional (3D) printing to manufacture ceramics in microgravity for the first time. During flights of a European parabolic aircraft (to access reduced gravity), the scientists employed digital light processing technology to produce material by solidifying a paste formed by mixing a fine powder similar to lunar dust into a resin. The team also tested a metal casting method using 3D-printed ceramic molds. "We can develop the technology to manufacture ceramic molds with the lunar dust, and then cast components by using metals in the lunar soil with the ceramic molds," says CAS' Wang Gong. Wang says the experiments seek to advance technologies for fabricating components on a future Chinese space station.
|
Quantum Step Forward in Protecting Communications From Hackers
University of York
June 20, 2018
A new quantum-based process for distributing secure information developed by researchers at the University of York in the U.K. could avert major security hacks. A key milestone in the development of their prototype was the finding that communication is far more secure if untrusted detectors for capturing and measuring photons bearing information exist at a separate point in the procedure, somewhere between sender and receiver. The theoretical detector would receive a blend of two signals, one each from the sender and receiver, but would only be able to read the result of this combination instead of its constituent elements. "In our work, not only have we provided a first rigorous mathematical proof that this 'detector-independent' design works, but we have also considered a scheme that is compatible with existing optical-fiber communication networks," says York's Cosmo Lupo.
|
Towards Personalized Medicine: One Type of Data Is Not Enough
European Molecular Biology Laboratory
June 19, 2018
European Molecular Biology Laboratory (EMBL) researchers in Germany and the U.K. have invented a computational technique to collaboratively analyze multiple classes of molecular data from individual patients, in order to identify their specific molecular signatures. Their Multi-Omics Factor Analysis (MOFA) was applied to multi-omics data (data from the genome, epigenome, transcriptome, metabolome, and other molecular data) from 200 leukemia patients, and it detected factors expressing molecular variability between subjects that could potentially influence personalized therapy. "Our method allows researchers...to easily integrate complex molecular data from DNA, RNA, methylation, and more, to build a tumor's molecular profile," says EMBL's Ricard Argelaguet. "Using these profiles, the method can also stratify patients into groups that may benefit from different types of treatment." EMBL's Britta Velten says MOFA was developed using a combination of math, statistics, machine learning, biology, and clinical medicine specialization, in order to "hopefully help researchers in a clinical setting to improve cancer diagnosis and treatment."
|
U-M Researchers Create World's Smallest 'Computer'
University of Michigan
Katherine McAlpine
June 21, 2018
Following IBM's announcement in March that it had created the world's smallest computer, University of Michigan (U-M) researchers have now developed an even smaller device that measures just 0.3 mm on each side. IBM's claim raised questions about what constitutes a computer, because previous tiny systems, including U-M's own Michigan Micro Mote, measuring 2x2x4mm, retained programming and data without external power. New microdevices, including those from IBM and now U-M, lose all prior programming and data as soon as they power is turned off. Michigan's David Blaauw says it is a matter of opinion whether the new devices offer enough functionality to be called computers. The new devices receive and transmit data with visible light and have RAM, photovoltaics, processors, wireless transmitters, and receivers. Michigan's new device converts temperatures into time intervals, defined with electronic pulses, measured on-chip against a steady time interval sent by the base station and converted into a temperature. The system could be applied to many purposes, but the team chose precision temperature measurements to fulfill an oncology need.
|
Report Gives 24-Hour View of Cyberattacks in Florida, U.S.
Florida Atlantic University
Gisele Galoustian
June 20, 2018
Florida Atlantic University researchers have completed the first-ever large-scale analysis of the magnitude of compromised Internet of Things (IoT) devices worldwide. They launched FloridaSOAR (security operation and response), a program designed to detect exploits, then store and share relevant threat information with IoT operators worldwide. FloridaSOAR can identify malicious attacks and infections in near-real time by targeting sectors and Internet service providers within specific jurisdictions. The researchers found within a unique 24-hour time frame, there were 250,779 malicious attacks in the U.S., with California, New York, and Texas being the hardest-hit states. The targets hit most frequently were power utilities, water facilities, and factories. The researchers have received a U.S. National Science Foundation grant to continue the project, which they will use to detect compromises in consumer networks in order to understand how these attacks are coordinated and launched; they plan to store the information they assemble in an accessible database.
|
'E-Dermis' Brings Sense of Touch, Pain to Prosthetic Hands
Johns Hopkins University
Amy Lunday
June 20, 2018
A team of engineers at Johns Hopkins University (JHU) has created an electronic skin that, when layered on top of prosthetic hands, provides a real sense of touch through the fingertips. The electronic skin, called e-dermis, is made of fabric and rubber laced with sensors to mimic nerve endings. The system recreates a sense of touch as well as pain by sensing stimuli and relaying the impulses back to the user’s peripheral nerves. The e-dermis device electrically stimulates the amputee's nerves in a non-invasive way via the skin, says JHU's Nitish Thakor. The researchers created a neuromorphic model mimicking the touch and pain receptors of the human nervous system, enabling the e-dermis to electronically encode sensations similarly to the way natural receptors in the skin would. The team then tracked brain activity to verify that the test subject was able to feel through the e-dermis.
|
Flying DRAGON Robot Transforms Itself to Squeeze Through Gaps
IEEE Spectrum
Evan Ackerman
June 20, 2018
Scientists at the University of Tokyo in Japan have developed a modular flying robot powered by ducted fans that can change shape while the device is in flight. The Dual-rotor embedded multilink Robot with the Ability of multi-deGree-of-freedom aerial transformatiON (DRAGON) can transform from a square to a snake and any shape in between, enabling it to stretch out to pass through small holes. DRAGON is comprised of a series of linked modules connected to each other via powered hinged joints. Each module consists of a pair of ducted fan thrusters that can be actuated to move in almost any direction. DRAGON is "a breakthrough in hardware design which, in a beautiful way, connects a manipulation arm with a ducted fan-driven aerial robot," says the University of Tokyo's Fan Shi.
|
Realistic Avatars for the Virtual Zoo
Max Planck Institute for Intelligent Systems
June 20, 2018
An international team led by the Max Planck Institute (MPI) for Intelligent Systems in Germany has created an algorithm for generating lifelike three-dimensional (3D) avatars of four-legged animals solely from photos. Using images taken from different angles, the process can calculate realistic models capable of being animated naturally. "The key is to start with a rough 3D shape model that can explain the shapes of many animals and then refine this based on the observed images," says the institute’s Michael J. Black. The team developing the new algorithm included researchers from the CNR Institute for Applied Mathematics and Information Technology (IMATI) in Italy, and from the University of California, Berkeley. "Our algorithm generates realistic models of animals for which no previous 3D models exist," notes IMATI's Silvia Zuffi.
|
MIT Engineers Build Smart Power Outlet
MIT News
Jennifer Chu
June 15, 2018
Massachusetts Institute of Technology (MIT) researchers have developed a smart power outlet that can analyze electrical current usage from single or multiple outlets, differentiating benign arcs from dangerous arcs. The system also can be trained to identify what might be plugged into a specific outlet, such as a fan instead of a desktop computer. The custom hardware of the smart power outlet processes electrical current data in real time, while software analyzes that data via a neural network. The system relies on a machine learning algorithm to determine whether a signal is harmful by comparing it to those used to train the system. The outlet can connect to other devices wirelessly via the Internet of Things. "I think the larger goal is to have everything connected, all of the time, for a smarter, more interconnected world," says MIT's Joshua Siegel.
|
DNA Barcodes That Reliably Work: A Game-Changer for Biomedical Research
UT News
June 20, 2018
A new method invented by University of Texas at Austin (UT) researchers can correct the errors the creep into DNA barcodes, quickly reducing their error rate from 10% to 0.5%. Essential to this technique is selecting the correct barcodes from the start, which entails choosing a string of letters for each barcode such that even if a small error infiltrates, it will still be more like the desired barcode than any other. The method demands jettisoning many possible strings of letters, but the team used sphere packing principles to keep this loss to a minimum. “DNA barcodes are a part of a great deal of cutting-edge research in medicine and drug development, and to be able to improve the accuracy and efficiency of so many of these is very exciting,” said UT’s John Hawkins. “And maybe even more exciting is that now with these better barcodes, this allows us to have larger, more ambitious experiments that weren’t possible before.”
|
A Computer Program That Learns to 'Imagine' the World Shows How AI Can Think More Like Us
MIT Technology Review
Will Knight
June 14, 2018
DeepMind, a subsidiary of Google parent company Alphabet focused on artificial intelligence (AI), has created a computer program that independently builds a mental picture of the world. The system, which uses a generative query network, looks at a scene from several angles, then describes what it would look like from another angle. This requires the relatively sophisticated ability to learn about the physical world and interpret a scene similarly to a human. Such technology could lead to deeper AI that allows machines to describe and reason about the world. The system uses two neural networks, with one learning while the other "imagines" new perspectives. Harvard University's Sam Gershman says the DeepMind work combines important ideas about the mechanics of human visual perception.
|
|
