MS in Data Science
Welcome to the January 6, 2020 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."

A girl jogging at a park. Gym Class Without the Gym? With Technology, It's Catching On
Associated Press
Carolyn Thompson
January 4, 2020

Wearable fitness trackers are helping make online physical education more popular for students by adding the dimension of accountability. Elizabeth Edwards at Fairfax County Public Schools in Virginia said students are asked to wear Fitbits while performing activities of their choice, and can adjust the activity provided the exercise raises their heart rate. Instructors guide activities by setting goals like fat burn, cardio, or peak, using the trackers as monitors in conjunction with weekly 60-minute to 90-minute classroom sessions. Illinois' Joliet Township High School employs fitness trackers in a blended learning conditioning program in which enrollees work out two to three times a week in the gym with an instructor, and other days on their own. The Fitbits monitor how well students are working out when not in the classroom.

Full Article
Tool Predicts How Fast Code Will Run on a Chip
MIT News
Rob Matheson
January 6, 2020

Massachusetts Institute of Technology (MIT) researchers have created a machine learning tool to predict processors' code-execution speed. The researchers first unveiled the ithemal neural network model that trains on labeled data in basic blocks—segments of computing instructions—to predict how long a given chip will take to execute previously unseen blocks. The researchers then compiled a benchmark suite of basic blocks to validate performance models in the BHive dataset. Ithemal predicted how fast Intel chips would run code with better accuracy than a performance model from Intel; the tool can produce code that runs faster and more efficiently on increasingly diverse black box chip designs. MIT's Thirimadura Charith Mendis said, "If you want to train a model on some new architecture, you just collect more data from that architecture, run it through our profiler, use that information to train Ithemal, and now you have a model that predicts performance."

Full Article
FPGA Cards Can Be Abused for More Reliable Rowhammer Attacks
Catalin Cimpanu
January 2, 2020

Researchers at Worcester Polytechnic Institute, Germany's University of Lubeck, and Intel demonstrated an exploit of field-programmable gate array (FPGA) cards to launch faster and more reliable Rowhammer attacks, which use a design flaw in random-access memory (RAM). The Rowhammer exploit repeats high-speed readings of data on one row of memory cells to produce a data-altering electrical charge to corrupt or manipulate a victim's RAM data. In the FPGA-based JackHammer variant, attack code is launched from within a user-configured FPGA. FPGA cards link directly to a processor's bus, allowing direct and untethered access to the central processing unit (CPU) cache and RAM memory—while the absence of firmware and operating-system software means FPGAs can run code faster than a normal CPU. A JackHammer attack also is harder to detect because it leaves no traces on the CPU of the FPGA's memory access operation.

Full Article

A Starship Technologies robot fulfills a delivery on the UT Dallas campus. On University of Texas at Dallas’ Growing Campus, Meal-Delivering Robots Make Splashy Debut
Dallas Morning News
Melissa Repko
December 26, 2019

A fleet of cooler-shaped autonomous robots has started delivering snacks and meals at the University of Texas at Dallas. The robots, developed by Starship Technologies, also deliver food at the University of Houston, and at George Mason University in Virginia. Students, faculty, and staff can order through a smartphone app from 10 different campus restaurants. The robots, stationed around each campus, go to a restaurant, pick up the order, and guide themselves to the customer's nearest outdoor location. Each robot is equipped with an orange flag that lights up to alert cars and pedestrians.

Full Article
ORNL Researchers Advance Performance Benchmark for Quantum Computers
Oak Ridge National Laboratory
Scott S. Jones
January 2, 2020

Researchers at the U.S. Department of Energy (DOE)'s Oak Ridge National Laboratory (ORNL) have created a quantum chemistry simulation benchmark to assess the performance of quantum devices and direct development for future quantum computing applications. ORNL's Raphael Pooser said the Quantum Testbed Pathfinder project is focused on relatively simple scientific challenges that embody the kinds of problems quantum systems will help address. The researchers calculated the bound-state energy of alkali hydride molecules on 20-quantum-bit (qubit) IBM Tokyo and 16-qubit Rigetti Aspen chips, effectively testing the quantum system's performance by tuning the system as a function of a few parameters. Said Pooser, "This project helps DOE better understand what will work and what won't work as they forge ahead in their mission to realize the potential of quantum computing in solving today's biggest science and national security challenges."

Full Article

Woman driving car. Bosch Wins CES Best of Innovation Award for LCD Virtual Visor
Sam Abuelsamid
January 5, 2020

Engineering company Bosch has been designated to receive a CES Best of Innovation award at CES 2020 for a virtual visor that enables safe driving when facing the sun. The visor combines a liquid crystal display (LCD) with a driver-facing camera and computer controls, with the camera focused on detecting the driver's face and gaze direction. A monochrome LCD replaces a sun visor to cast a shadow across the driver's eyes, protecting them from direct sunlight; as the sun moves or the road curves, the LCD panel automatically adjusts the monochrome strip to keep the shadow over the driver’s eyes.

Full Article
People Too Trusting of Virtual Assistants
University of Waterloo News
January 2, 2020

Researchers at the University of Waterloo in Canada have found that people tend to share increasingly more with online agents because of their tendency to assign them personalities and physical features like age, facial expressions, and hairstyles. The researchers asked 10 men and 10 women to interact with three conversational agents—Alexa, Google Assistant, and Siri. The team then interviewed each participant to determine their perception of the agents' personalities and what they would look like, before asking each person to create an avatar for each agent. The researchers found Siri was most frequently described as disingenuous and cunning, while Alexa was perceived as genuine and caring. Said University of Waterloo researcher Anastasia Kuzminykh, "How an agent is perceived impacts how it's accepted and how people interact with it; how much people trust it, how much people talk to it, and the way people talk to it."

Full Article

Fossil cast of a fin from a juvenile Sauripterus taylori Modeling Shows How Fish Fins Evolved on the Way to Land-Ready Limbs
Matt Wood
December 31, 2019

University of Chicago (UChicago), University of Wisconsin-Parkside, and Drexel University researchers used computed tomography scanning to map the evolution of fins into limbs from Devonian fish fossils via three-dimensional modeling. The researchers scanned the fossils while still within rock, then used imaging software to build models that allowed movement, rotation, and visualization of the dermal skeleton. Simulations revealed that the specimens' fin rays were simplified, while the overall size of their fin web was smaller than that of their precursors, and the fins' tops and bottoms exhibited incipient asymmetry. The researchers also compared the dermal skeletons of living fish to gain insights on the fossil patterns. UChicago's Thomas Stewart said, "That gives us more confidence and another dataset to say these patterns are real, widespread, and important for fishes, not just in the fossil record as it relates to the fin-to-limb transition, but the function of fins broadly."

Full Article
Researchers Try to Teach Computers to Forecast Traffic Like the Weather
The Washington Post
Michael Laris
December 28, 2019

Netherlands-based location/transportation data supplier Here Technologies has pushed researchers to apply big-data techniques to films of traffic in Berlin, Istanbul, and Moscow, to predict traffic. The firm’s artificial intelligence institute recently hosted a traffic forecasting competition which demonstrated neural networks' ability to extract patterns from massive datasets. The company’s Michael Kopp said Here gave the competition’s participants months’ worth of color-coded traffic data from the major German, Turkish and Russian cities; contestants built software tools to mine patterns from that information. Kopp said Sungbin Choi, an independent researcher from Seoul, won the contest with a tool that had an error rate of less than 1%.

Full Article
*May Require Paid Registration
Conifers Survive Droughts with Existing Roots, Not New Growth
UB News Center
Charlotte Hsu
December 30, 2019

A multi-institutional study led by University at Buffalo (UB) researchers investigating the resistance of coniferous trees to droughts via computational modeling found that some such trees rely on established, deep roots. Simulated pine and juniper trees survived a five-year drought using deep roots already penetrating groundwater-retaining fractured bedrock. UB's Scott Mackay said the findings could shed light on coniferous forests' response to climate change. Said Mackay, "Some models tend to overestimate tree mortality because they're not able to capture some of these refugia. If we can learn more about these refugia and how they're established, we can use that knowledge to create better models."

Full Article
Researchers Use Deep Learning to Predict Disease-Related Mutations of Metal Binding Sites in Protein
Business Standard
December 28, 2019

Researchers at the University of Hong Kong (HKU) and the Mayo Clinic, Arizona, used deep learning to forecast disease-associated mutations of metal-binding sites in a protein. The researchers found a specific mutation in the human genome that might disrupt these sites and induce severe diseases. HKU's Hongzhe Sun and colleagues determined different metals have different disease connections. For example, a mutation in calcium and magnesium binding sites is respectively tied to muscular and immune system diseases, while zinc plays a key role in breast, liver, kidney, immune system, and prostate diseases. Sun said, "I believe this novel deep learning approach can be used in other projects."

Full Article

Children from a Girls Who Code club in Bangalore CIOs Look to Boost Tech Teams' Diversity
The Wall Street Journal
Angus Loten
December 27, 2019

Chief information officers (CIOs) are striving to boost the diversity of their technology workforce, which could help ameliorate a shortfall of information technology (IT) employees. IT trade group CompTIA cited ongoing wage gaps, unbalanced executive teams, and abusive behavior as persistent barriers to improving diversity, while 40% of U.S. companies intend to prioritize IT workforce diversity in 2020. Some companies have implemented mentoring and skills-development programs for young women; cultivating IT skills early can help bring more women into tech careers and leadership roles. Synchrony Financial CIO Carol Juel closely collaborates with the nonprofit Girls Who Code, which offers computer science education to middle- and high-schoolers, pairing students with IT professionals. Some CIOs are providing internal training programs to boost awareness of the need for diversity in IT, and nurture a more inclusive environment.

Full Article
*May Require Paid Registration
The Essentials of Modern Software Engineering
ACM Discounts and Special Offers Program

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434

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]