Welcome to the July 30, 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."

Exascale system, illustration Exascale Computing Project Establishes Machine Learning Center
Top 500
Michael Feldman
July 23, 2018

The U.S. Exascale Computing Project (ECP) has set up ExaLearn, a co-design center focused on machine learning (ML) technologies. The sixth such center under the ECP, ExaLearn will target development of ML exascale software for science and engineering applications and other work being performed under the ECP umbrella. The center's goal is to produce a "scalable and sustainable ML software framework that allows application scientists and the applied mathematics and computer science communities to engage in co-design for learning." The ECP will work in collaboration with U.S. Department of Energy PathForward vendors to help develop software for the various hardware platforms. The practical outcome of the co-design work is to enable such applications to be developed for all exascale supercomputers in the same manner as more traditional simulation and modeling codes.

Full Article
Surrey Develops Crime Fighting Algorithm That Could Predict Reoccurring Illegal Activity
University of Surrey
July 24, 2018

Researchers from the University of Surrey in the U.K. and the Georgia Institute of Technology have developed a supplemental approach for the Epidemic Type Aftershock Sequence (ETAS) model to urban crime data. The researchers used the new approach to develop the Ensemble Poisson Kalman Filter (EnPKF), a novel algorithm that can combine, in real time, urban crime data and the ETAS model. EnPKF can provide real-time forecasts for the crime rate and determine how likely crime is to repeat in a given area. In addition, the algorithm can give police suggestions as to where short-term crime hotspots could arise, and which additional resources are needed to address the issue. Said Surrey’s David Lloyd, “We are cautiously excited about the Ensemble Poisson Kalman filter, an approach that has given us an insight into when crime can be predicted, and has shown us the importance of using real-time data to make the overall system stronger.”

Full Article

med-tech illustration NIH Partners With Google Cloud to Speed Up Medical Breakthroughs
Jack Corrigan
July 25, 2018

A new U.S. National Institutes of Health (NIH) program aims to help NIH researchers accelerate biomedical advances providing access to more commercial cloud computing technologies. NIH's Science and Technology Research Infrastructure for Discovery, Experimentation, and Sustainability Initiative (STRIDES) will give scientists access to analytics tools to speed up biomedical data-crunching. STRIDES' initial concentration will be on making "high-value" datasets accessible via the cloud and adopting data-centric machine learning and artificial intelligence applications. Google Cloud was chosen as NIH's first STRIDES partner so it can offer its cloud storage, computing, and machine learning tools to NIH teams and about 2,500 federally-funded institutions. Google Cloud also will collaborate with the NIH Data Commons program to devise cloud-based biomedical data-sharing solutions.

Full Article
Decade-Old Bluetooth Flaw Lets Hackers Steal Data Passing Between Devices
Ars Technica
Dan Goodin
July 25, 2018

A study from the Technion-Israel Institute of Technology warns of a decade-old bug in the Bluetooth specification that allows hackers to intercept and tamper with data shared wirelessly through man-in-the-middle attacks on the link between devices. Not only can hackers view the data, but they can forge keystrokes on a Bluetooth keyboard to open up a command window or malicious website. Says security engineer JP Smith, "This attack lets an attacker who can read and modify Bluetooth traffic during pairing force the key to be something they know." The researchers say the attack is enabled by two design flaws: one involves sending both the x-coordinate and the y-coordinate during the public key exchange, while the other is the protocol's authentication of only the x-coordinate.

Full Article

computing system Trapped-Ion Quantum Computer Does Chemistry Calculations for the First Time
Chemistry World
Philip Ball
July 24, 2018

An international team has developed a quantum computer that can calculate the energy states of molecules of hydrogen and lithium hydride using four trapped ions. The team employed an array of trapped calcium ions to infer the molecules' ground-state energies, and their quantum computations correlate well with experimental measurements and cutting-edge classical calculations. The computations utilize a system that hosts a 20-quantum bit (qubit) linear series of individually controllable trapped ions, although only three to four qubits are required for the calculations. The calculation determines the molecular energy represented in the qubits by shuffling its electrons between the molecular orbitals and seeking the configuration of lowest energy. The team plots changes to the electron states in molecular orbitals to laser-driven operations that switch the ion qubits' states, and read out the final states of the ions with a second laser.

Full Article
*May Require Free Registration
Iris Scanner Can Distinguish Dead Eyeballs From Living Ones
Technology Review
July 24, 2018

Researchers at the Warsaw University of Technology in Poland have compiled a database of iris scans from living people and from cadavers to train a machine learning algorithm to differentiate between living and dead samples. They say the algorithm is 99% accurate in distinguishing them. The database is comprised of 574 near-infrared iris images from 17 people at various times post-mortem, while 256 images of live irises also were added. However, the algorithm's accuracy is only applicable to irises that have been dead for 16 hours or more. The team says, "Samples collected briefly after death (i.e., five hours in our study) can fail to provide post-mortem changes that are pronounced enough to serve as cues for liveness detection." This would allow hackers a brief window of opportunity to use stolen eyeballs on biometric scanners.

Full Article
Bug-Sized Robot Competitors to Swarm DARPA's 'Robot Olympics'
Live Science
Mindy Weisberger
July 21, 2018

The U.S. Defense Advanced Research Projects Agency (DARPA) is seeking insect-sized robot designs to compete in contests of strength, speed, and agility under its Short-Range Independent Microrobotic Platforms (SHRIMP) program. The machines will be tested for deployment in locations too hazardous or inaccessible for human navigation. SHRIMP's goal is to invent solutions for powering small robots, and probing materials that could enhance performance without adding to their size or weight. DARPA says SHRIMP's "Olympic-style evaluation" will see how well the bots demonstrate their maneuverability, dexterity, and mobility. One test area will concentrate on untethered actuator-power systems, showing how high and how far robots can jump, how much weight they can lift, how far they can throw objects, and how they fare in a tug of war. Another contest will assess the robots on rock piling, climbing vertical surfaces, navigating an obstacle course, and performing in a biathlon.

Full Article

An illustration of the human mind with artistic interpretation of emotional elements Helping Computers Perceive Human Emotions
MIT News
Rob Matheson
July 24, 2018

Researchers in the Massachusetts Institute of Technology (MIT) Media Lab have developed a machine learning model that aims to improve existing affective-computing technologies. The model outperformed traditional systems in capturing small facial expression variations in order to better gauge mood while training on thousands of images of human faces. In addition, by using extra training data, the model can be adapted to an entirely new group of people. The researchers combined a technique called mixture of experts (MoEs) with model personalization techniques, which helped the model mine more fine-grained facial expression data from individuals. The team personalized the MoEs by matching each expert to one of 18 individual video recordings in a public database of people conversing on a video-chat platform designed for affective-computing applications. The researchers trained the model using nine subjects and evaluated them on the other nine, with each video broken down into individual frames.

Full Article

Photo of a train running through mountainous terrain in Australia. World's Largest Robot Hauls Ore Through Western Australia
IEEE Spectrum
Evan Ackerman
July 23, 2018

The world's largest robot, called AutoHaul, is an autonomous train that hauls tons of material across Australia. The trains started running in direct supervised autonomy mode in 2017. A human was originally present on the train, but it was still operating autonomously, under remote supervision from an operations center. The train recently made its first trip using only remote supervised autonomy. The robotic system understands its performance, the load it is carrying, and the characteristics of the terrain, all of which enable it to drive itself faster and more efficiently than a human conductor. The system is designed to unlock significant safety and productivity benefits as a result of reduced variability and increased speed across the network, helping to decrease average cycle times.

Full Article
Scientists Perfect Technique to Boost Capacity of Computer Storage a Thousandfold
Jennifer-Anne Pascoe
July 23, 2018

A team at the University of Alberta (U of A) in Canada has refined a method to enable a 1,000-fold enlargement of computer storage capacity, and used it to fabricate atomic-scale circuits. U of A's Roshan Achal, Robert Wolkow, and their colleagues successfully encoded the entire alphabet at a density of 138 terabytes per square inch, the rough equivalent of writing 350,000 letters on a grain of rice. Achal says, "Our memory is stable well above room temperature and precise down to the atom." He also says he sees applications for the technology in data archiving, and the research next will seek to boost read and write speeds for even more flexible uses. The team's atomic-scale circuits permit the rapid removal or replacement of single hydrogen atoms, an innovation that renders the memory rewritable, so it could give rise to much more efficient types of solid-state drives for computers.

Full Article
AI Technology Could Help Protect Water Supplies
University of Waterloo News
Matthew Grant
July 18, 2018

Researchers at the University of Waterloo in Canada have developed artificial intelligence (AI) software that can identify and quantify different kinds of cyanobacteria, also known as blue-green algae. Such algae can shut down water systems when it suddenly proliferates. The AI system uses software in combination with a microscope to inexpensively and automatically analyze water samples for algae cells in about one to two hours, while current systems can take one to two days. The researchers hope to create an AI system that can continuously monitor water flowing through a microscope for a wide range of contaminants and microorganisms.

Full Article
Certificate in Infographics and Data Visualization
Subscribe to Communications of the ACM

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]