Welcome to the January 22, 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."
|
|
U.K. Tech Investment Grew Faster Than U.S., China in 2019
CNBC Ryan Browne January 14, 2020
A study by industry group Tech Nation and research firm Dealroom estimated that venture capital funding for U.K. technology startups grew 44% to a record $13.2 billion last year, outpacing slowing investment in the U.S. and China. However, the U.S. and China earned the most total deal value, respectively drawing $116 billion and $33.5 billion. U.S. and Asian funding sources accounted for almost 50% of total U.K. tech venture capital in 2019, with financial technology, artificial intelligence, and clean energy showing substantial growth. U.K. digital minister Matt Warman credited the expansion of British tech investment to a "package" of incentives that includes a supportive business landscape, geography, language, and proximity to leading educational institutions.
|
AB InBev Taps Machine Learning to Root Out Corruption
The Wall Street Journal Dylan Tokar January 17, 2020
Brewer Anheuser-Busch InBev spent three years developing machine learning technology to spot corruption in its business partners. The BrewRight analytics platform harnesses data from operations in more than 50 countries to proactively track legal risks and deter violations, rather than investigating problems after they crop up. Companies have traditionally probe misconduct after it happens, but Harvard Business School's Eugene Soltes said, "Data analytics and what AB InBev has done changes that equation. They want to put much more on the front-end—on prevention and detection." The machine learning aspect allows the platform to become smarter and more effective over time. It already has cut hundreds of thousands of dollars in costs associated with investigating suspect payments.
*May Require Paid Registration
|
The Way You Dance is Unique, and Computers Can Tell It's You
University of Jyvaskyla (Finland) January 17, 2020
Researchers at Finland’s University of Jyvaskyla have taught computers to identify individuals by their dance style. Their original intent was to use machine learning and motion-capture technology to determine the musical genre subjects were listening to, based on motion-captured dance movements; instead, the system correctly identified which individuals were dancing 94% of the time. The researchers think the implications are more valuable to human musicality than to applications like surveillance. Jyvaskyla's Emily Carlson said, "We have a lot of new questions to ask, like whether our movement signatures stay the same across our lifespan, whether we can detect differences between cultures based on these movement signatures, and how well humans are able to recognize individuals from their dance movements compared to computers."
|
Brain-Like Network Uses Disorder to Detect Order
University of Twente (Netherlands) January 15, 2020
At the University of Twente in the Netherlands, researchers have developed a brain-inspired disordered network for detecting ordered patterns, which uses "hopping conduction" to reach solutions without predesigned elements. The disordered dopant atom network uses material properties to evolve towards a solution; it is energy-efficient and requires little surface space. The researchers fed the network 16 four-digit patterns, each yielding a different output signal, which permitted the network to learn to recognize a database of handwritten letters with great accuracy. Future applications could include pattern recognition, with potential benefits for autonomous driving and facial recognition.
|
Avoiding Carsickness When the Cars Drive Themselves
The New York Times Bradley Berman January 17, 2020
University of Michigan researchers are analyzing passengers' motion sickness in self-driving vehicles. Automotive engineers like Florian Dauth at German automotive supplier ZF Group hope to apply that data to the development of ways to prevent motion sickness in autonomous vehicles. Dauth said ZF is creating algorithms that learn from bodily reactions; he collects passengers' biological data (such as heart and brain activity) through wired sensors, and hopes to use the data to help self-driving vehicles’ artificial intelligence learn to drive in ways that don’t sicken its passengers. Meanwhile, Volkswagen technologist Brian Lathrop is working to eliminate motion sickness when using virtual reality (VR) in moving vehicles. Said Lathrop, "I anticipate that we might have a very low-profile, lightweight type of VR platform that's like putting on a pair of sunglasses."
*May Require Paid Registration
|
AI Learns to Rule the Quantum World
Aarhus University Rasmus Rorbaek January 16, 2020
Researchers at Aarhus University in Denmark used a game-playing algorithm to run a quantum computer. The AlphaZero algorithm can teach itself to beat human game-players at chess without human intervention, and the Aarhus group used computer simulations to apply AlphaZero to three distinct control problems that could each find potential use in a quantum computer. The researchers realized the best results when combining AlphaZero with a specialized quantum optimization algorithm. Said Aarhus' Jacob Sherson, "This indicates that we are still in need of human skill and expertise, and that the goal of the future should be to understand and develop hybrid intelligence interfaces that optimally exploit the strengths of both."
|
Edible 'Security Tag' to Protect Drugs From Counterfeit
Purdue University News Kayla Wiles January 16, 2020
Purdue University researchers have developed a consumable "security tag" that can be embedded into medications to thwart drug counterfeiters. The tag serves as a digital signature for each pill or tablet, using physical unclonable functions (PUF). The researchers created an edible PUF, in the form of a thin, transparent film of genetically combined silk and fluorescent proteins; illuminating the tag with a light-emitting diode (LED) causes fluorescent silk microparticles to glow in a distinct random pattern in cyan, green, yellow, or red. The resulting image contains digital bits that form a security key to confirm the drug's authenticity. Purdue's Jung Woo Leem said, "Our concept is to use a smartphone to shine an LED light on the tag and take a picture of it. The app then identifies if the medicine is genuine or fake."
|
Seeing Around the Corner with Lasers—and Speckle
IEEE Spectrum Philip E. Ross January 16, 2020
Researchers at Rice, Stanford, Princeton, and Southern Methodist universities used lasers to see around corners, with superior resolution and scanning speed to previous methods. The technique relies on speckle, a shimmering interference pattern in many laser applications that contains important spatial data; a massive calculation is needed to extract an image from the speckle, which the researchers expedited with deep learning processes. Image acquisition takes 0.25 seconds and yields sub-millimeter resolution, but there are limitations. Rice's Ashok Veeraraghavan said, "Speckle encodes interference information, and as the area gets larger, the resolution gets worse."
|
U.K. Police Use of Facial Recognition Tests Public's Tolerance
Associated Press Kelvin Chan January 16, 2020
The testing of facial-recognition technology by police in the U.K. has provoked outrage, with rights activists and others protesting at a Wales soccer match where law enforcement used the technology to scan and detain individuals barred for past misbehavior. The South Wales Police began trialing van-mounted face-scanning cameras in 2017; last year, a court ruled such testing is legal, but regulators and lawmakers have yet to draft regulations for the cameras' use. The cameras scan crowds and match faces to a database of wanted or suspected criminals, and the University of Essex's Pete Fussey said the system is effective under laboratory conditions. However, a real-world trial in London yielded only eight correct matches out of 42. Said Fussey, "The police tended to trust the algorithm most of the time, so if they trust the computational decision-making yet that decision-making is wrong, that raises all sorts of questions."
|
U.S. to Ground Civilian Drone Program on Concerns Over China Tech
Financial Times Kiran Stacey January 12, 2020
The U.S. Department of the Interior reportedly plans to permanently ground a civilian drone fleet, because the approximately 1,000 drones are partly sourced from China. The department concluded that the risk is too great that the devices could be used by China's government for spying. The drones were used to help manage wildfires, map terrain, and monitor natural resources. The Interior Department said it would provisionally ground 810 camera drones while Interior Secretary David Bernhardt reviews their security risks. A department spokesperson said drones made in China or featuring Chinese components will stay grounded except for emergency use. Staff at several federal agencies have balked at the plan to permanently ground the drone fleet.
*May Require Paid Registration
|
Technology From Baltimore-Based Adashi Systems Seen in Viral Australia Bushfire Video
The Baltimore Sun Phil Davis January 14, 2020
A viral video of firefighters fleeing an Australian bushfire in New South Wales shows their use of a tablet computer from Baltimore-based Adashi Systems to navigate the conflagration. Adashi president Sanjay Kalasa said the FirstResponse Mobile Data Terminal (MDT) feeds firefighters live data about their current fire site and other emergency personnel at the scene. The tablet came in handy when the firefighters' vehicle caught fire, forcing the team to search for assistance on foot. The technology gives first responders efficient map routes for how to respond to a fire scene, and the location of other responding units. More than 1,400 fire departments use Adashi's technology, according to Kalasa.
|
Who's Liable, the AV or the Human Driver?
Columbia Engineering Holly Evarts January 14, 2020
Columbia Engineering and Columbia Law School researchers developed a fault-based liability rule for regulating autonomous vehicle (AV) manufacturers and human drivers. Columbia Engineering's Sharon Di said human drivers view AVs as intelligent agents capable of adapting to aggressive and potentially dangerous driving behavior, which may encourage careless and riskier driving. A game theory model assigns different goals to lawmakers, AV manufacturers, AVs, and human drivers in the transportation ecosystem, and visualizes the strategy each player arrives at so others will not exploit their decisions. The researchers found an optimally-designed liability policy is essential to preventing human drivers from cultivating moral hazard, and to helping AV manufacturers with a tradeoff between traffic safety and production expenses. Said Columbia Law's Eric Talley, "We hope our analytical tools will assist AV policymakers with their regulatory decisions, and in doing so, will help mitigate uncertainty in the existing regulatory environment around AV technologies."
|
AI Suggests Earth Has Had Fewer Mass Extinctions Than Thought
New Scientist Michael Le Page January 16, 2020
Researchers at China's Nanjing Institute of Geology and Paleontology and the U.S. National Museum of Natural History used machine learning and a supercomputer to generate a biodiversity record that suggests fewer mass extinctions have occurred in the history of Earth than previously thought. Their record omits the late Devonian extinction by dividing the past into shorter chunks of time—only 26,000 years long, instead of about 10 million years long—and statistically analyzing 100,000 records of 11,000 marine species found in fossils. The researchers developed machine-learning procedures to process the data on China’s Tianhe-2 supercomputer. Said paleontologist Richard Bambach, "The mid-late Devonian diversity decrease is still very clear, but it is spread through the whole time and not concentrated in a single mass extinction.”
|
|