Welcome to the January 31, 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."
|
|
|
Israeli Researchers Unveil First All-Optical ‘Stealth’ Encryption Tech
Jerusalem Post
Eytan Halon
January 29, 2020
Researchers at BGN Technologies, the technology transfer company of Israel’s Ben-Gurion University of the Negev (BGU), have developed an all-optical encryption system that enables secure, "stealthy" transmission of highly sensitive information. The new method uses standard optical equipment to spread the data across many colors in the optical spectrum bandwidth. The researchers report the solution easily renders the fiber-optic light transmission "invisible," enabling the transmission of weaker-encrypted data in streams hidden under increased noise levels. The process destroys "coherence," the ability to recompile sensitive data without the necessary encryption key. Said BGU's Dan Sadot, "Because an eavesdropper can neither read the data or even detect the existence of the transmitted signal, our optical stealth transmission provides the highest level of privacy and security for sensitive data applications."
|
Researchers Use FLAIR to Increase Data Retrieval Speeds
University of Waterloo News
January 30, 2020
At the David R. Cheriton School of Computer Science of Canada’s University of Waterloo, researchers have come up with a technique to upgrade the storage efficiency and output speeds of computer systems. The FLAIR (fast, linearizable, network-accelerated client reads) method optimizes data storage systems so data requests made when the main server is full automatically activate another server to handle the task. The researchers tested FLAIR with real workloads, and found the approach boosted data retrieval speeds by 35% to 97%. Cheriton's Ibrahim Kettaneh said FLAIR “can significantly improve the performance of databases and data processing engines, which are the backends for health systems, banking systems, and financial transactions. It will also be applicable to any modern computer application hosted on the cloud, such as online documents, social networks, and emails."
|
AI Epidemiologist Sent First Warnings of Wuhan Virus
Wired
Eric Niiler
January 25, 2020
Canadian health monitoring platform BlueDot was the first organization to alert its customers to the outbreak of coronavirus, on Dec. 31. BlueDot’s algorithm analyzes foreign-language news reports, animal and plant disease networks, and governmental information in order to warn its clients to avoid dangerous regions where disease outbreaks appear to be occurring. The algorithm also considers global airline ticketing data, to help predict where and when infected residents are traveling. BlueDot correctly predicted the coronavirus would jump from Wuhan, China, to Bangkok, Seoul, Taipei, and Tokyo in the days following its initial appearance. Said BlueDot CEO Kamran Khan, “What we have done is use natural language processing and machine learning to train this engine to recognize whether this is an outbreak of anthrax in Mongolia versus a reunion of the heavy metal band Anthrax.”
|
Autonomous Pods SWARM Together Like Bees in World-First Demonstration
University of Warwick
January 30, 2020
Researchers at the University of Warwick's Warwick Manufacturing Group (WMG) in the U.K., in collaboration with self-driving technology maker Aurrigo and the town of Milton Keynes, have created autonomous pods (driverless vehicles) that swarm together. The pods are engineered for pedestrian areas and shared spaces, and the researchers enhanced the pods with swarm intelligence to cluster into a "platoon" and follow each other, keeping individual vehicle movements and supervision to a minimum. The pods also can automatically optimize their behavior to satisfy future passenger demand by distributing themselves within a city to locations where rides are most likely to be requested. Aurrigo's Simon Brewerton said, "The collaborative SWARM algorithms have been developed to enable our autonomous vehicles to optimize their own trip schedules, so they deliver the optimum efficiency from a fleet of vehicles."
|
Seeking Tech Talent, Companies Kickstart Apprenticeship Programs
The Wall Street Journal
Agam Shah
January 30, 2020
Companies are launching apprenticeship programs to recruit nontraditional staff for entry-level positions in the midst of a technology talent shortage. Many applicants have a high school education, and apprenticeships provide training in digital skills like programming languages. Auto-parts supplier Robert Bosch's 12-month apprenticeship program will train applicants as software engineers for emergent technologies like driver assistance and autonomous vehicles; training will focus on system architecture, design, and C++ coding, along with soft skills such as communication, presentations, collaboration, and project management. Said Bosch’s Charlie Ackerman of the program, “The rapid change seen particularly in high-tech presents a challenge for education to adapt quickly enough. Industry cannot wait and is forced to take additional steps.”
*May Require Paid Registration
|
Deep Neural Networks Coming to Your Phone
News@Northeastern
Laura Castanon
January 28, 2020
Northeastern University researchers have developed a technique to run deep neural networks on smartphones or similar systems, resulting in the fulfillment of tasks up to 56 times faster than previously achieved, while maintaining accuracy. Northeastern's Yanzhi Wang and colleagues reduced the size of the network model and automatically generated code to improve operational efficiency, which could enable the deployment of deep neural networks in off-the-shelf devices that may lack consistent Internet access. Wang suggested local data processing that does not require transmission to remote servers could boost people's comfort with artificial intelligence-powered devices. Said Wang, "Previously, people believed that deep learning needed dedicated chips, or could only be run on servers over the cloud. This kind of assumption of knowledge limits the application of the deep learning."
|
A Warehouse Robot Learns to Sort Out the Tricky Stuff
The New York Times
Adam Satariano; Cade Metz
January 29, 2020
The Obeta electrical parts company in Germany has deployed a warehouse robot from Austrian automation company Knapp to sort components, using software developed by engineers at Covariant.AI, building on their research at the University of California, Berkeley. The software learns through trial and error by utilizing a digital simulation of a task, and transfers this knowledge to the sorting robot for real-world use. The robot continues to learn as it encounters unfamiliar items in the warehouse, where it can sort over 10,000 objects with more than 99% accuracy. Knapp is currently considering the design of warehouses staffed solely by robots, to allow for packages to be more densely packed and retrieved by robots trained to perform the task. Covariant’s Pieter Abbeel said humans would continue to work alongside robots in such warehouses, but he acknowledged that the job market will shift significantly as machine learning improves.
*May Require Paid Registration
|
Virtual Assistants Provide Disappointing Advice When Asked for First Aid, Emergency Information: Study
Folio (University of Alberta)
Gillian Rutherford
January 28, 2020
Researchers at the University of Alberta in Canada have found that virtual assistants do not live up to their potential in terms of providing users with reliable, relevant information on medical emergencies. The team tested four commonly used devices—Alexa, Google Home, Siri, and Cortana—using 123 questions about 39 first aid topics, including heart attacks, poisoning, nosebleeds, and splinters. The devices' responses were measured for accuracy of topic recognition, detection of the severity of the emergency, complexity of language used, and how closely the advice given fit with accepted first aid treatment and guidelines. Google Home performed the best, recognizing topics with 98% accuracy and providing relevant advice 56% of the time. Alexa also scored well, recognizing 92% of the topics and giving accepted advice 19% of the time. The quality of responses from Cortana and Siri was so low that the researchers could not analyze them.
|
Digital Twins Give Urban Planners Virtual Edge
Financial Times
Carly Minsky
January 29, 2020
Digital twins—virtual simulations of cities—give urban planners, policymakers, resource managers, asset owners, and citizens the advantage of more dynamic tools and information, which can benefit areas like construction and traffic management, and provide tools for residents to explore the effects of planned projects. One example is Glasgow, Scotland's Future City project: the initiative's tools offer citizen-accessible data on sustainability overlaid on interactive maps, customized dashboards with widgets on environmental conditions, and data on seasonal food availability. Meanwhile, Cambridge University's National Digital Twin Project created a U.K. network of city-scale digital twins, with guiding principles that prioritize trust, transparency, social value, ownership, responsibility, and security. Neil Thompson at engineering company Atkins Global said, "Digital twins offer new opportunities for sustainability and wellbeing—from a physical and a psychological perspective."
|
Illinois Law Gives Candidates Rights to Understand How Employers Use Algorithms to Rate Job Interviewees
The Philadelphia Inquirer
Abdel Jimenez
January 28, 2020
A new law in Illinois aims to give job candidates a better understanding of how companies use artificial intelligence (AI) to screen video interviews. The AI analyzes applicants' facial expressions, tone, and language, and many underlying algorithms are not explained by their designers; critics say this leads to biased assessments. Under the Illinois Artificial Intelligence Video Interview Act, employers must inform applicants when AI will be used for this purpose, and explain the algorithms' workings. Companies must secure permission from applicants before using such AI, and cannot share submitted videos except with "persons whose expertise or technology" are necessary to screen interviewees. Applicants also can request the videos' destruction, and employers have 30 days to comply with requests.
|
Mixing Quantum States Boosts Fiber Communications
IEEE Spectrum
Neil Savage
January 24, 2020
Researchers at Huazhong University of Science and Technology in China and the University of the Witwatersrand in South Africa have combined polarization and orbital angular momentum (OAM) into a hybrid state on entangled photons to increase the transmission of secure data through fiber. The team split a photon into two lower-energy photons (A and B) with different OAMs. The researchers then passed photon A through a set of optics that gave it a specific polarization, before sending the photons to different detectors and measuring the polarization of photon A and the OAM of photon B. Said Witwatersrand's Andrew Forbes, "The trick works because the photons don't know what they are until we measure them, so the state is unaware that we have multiple patterns in the game. But by the time we measure them and indicate the patterns, the one photon has already passed through the fiber."
|
Giving Cryptocurrency Users More Bang for Their Buck
MIT News
Rob Matheson
January 30, 2020
A group of researchers from the Massachusetts Institute of Technology, University of Illinois at Urbana-Champaign, and Carnegie Mellon University have created a cryptocurrency routing scheme to enhance the efficiency of networks designed to accelerate blockchain transactions. The Spider routing scheme allows payment channel network (PCN) users to invest a fraction of funds in off-blockchain escrow accounts and process about four times as many transactions as previously, before rebalancing on the blockchain. The scheme segments each full transaction into packets for routing across different channels at differing rates, diverting payment chunks through potentially low-funded accounts; this makes each packet more likely to arrive at its destination without slowing the network or being jettisoned by any given account for its size. Spider also employs an algorithm that watches data center congestion to spot queueing delays at account bottlenecks, to control transaction rates. In simulations, the scheme processed 95% of all transactions with just 25% of the capacity needed for traditional schemes in PCN models.
|
DC Water Taps AI for Sewer Line Assessments
Government Computer News
Stephanie Kanowitz
January 23, 2020
The Washington, D.C., Water and Sewer Authority (DC Water) and IT firm Wipro have developed an algorithm that helps sewer pipe technicians obtain a more-accurate picture of what is happening in the network of pipes running beneath the city. The Pipe Sleuth software analyzes video taken by autonomous robots, and identifies anomalies based on industry standards and best practices. The technicians program a path for the robots, which resemble a hot dog on wheels with 360-degree cameras on the front. The videos captured by the robots are uploaded to a DC Water server and processed through Pipe Sleuth, which reviews the footage, rates the pipes according to industry standards, and produces a report showing imagery of defects. The algorithm can identify nearly 50 different defects with up to 95% accuracy for vitrified clay pipes; it has a slightly lower accuracy rate for concrete pipes.
|
|
