Welcome to the August 2, 2019 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."
|
|
|
Would You Want a Computer to Judge Your Risk of HIV Infection?
The New York Times
Gina Kolata
July 30, 2019
Harvard University and Kaiser Permanente Northern California researchers used an algorithm that taps personal details from electronic health records to rate the risk of human immunodeficiency virus (HIV) infection among male patients. The algorithm identified these individuals and directed them to take medication via the PrEP daily pill HIV prevention strategy. The researchers developed risk prediction models using electronic records from 3.7 million uninfected Kaiser Permanente patients, and 1.1 million patients at two Massachusetts medical facilities. The final model incorporated 44 predictors and variables related to drug use, living situations, and medical tests for venereal disease. However, the software did not identify all at-risk individuals, which means physicians who ask patients if they want to be screened need to explain the test's limitations and inquire about other potential risk factors.
*May Require Paid Registration
|
Facebook Gets Closer to Letting You Type With Your Mind
CNN
Rachel Metz
July 30, 2019
Facebook is collaborating with academia to develop a non-invasive technique to type words onto a computer directly from the user's brain. The social network has underwritten work by University of California, San Francisco (UCSF) researchers on whether electrodes implanted in the brain could help "decode" neural speech from brainwaves in real time. The research demonstrated that a word or phrase spoken by epileptic participants is instantly visible as text on a computer screen, by examining neural activity. The researchers used speech-decoding algorithms, which were up to 61% accurate in determining which of 24 standard responses a participant had spoken. Another Facebook-funded UCSF protect will attempt to use brain activity to help a person who cannot speak to communicate, in the hope of revealing which brain signals are essential for a future wearable device.
|
Human Intelligence Is Key to the AI Age
UNSW Newsroom
July 30, 2019
An Australian Council of Learned Academies (ACOLA) report suggests the potential of artificial intelligence (AI) to improve Australia's economy, well-being, environment, and societal fairness can be realized, with careful planning. ACOLA's Toby Walsh said AI's benefits can be significant, "provided we ensure that the use of the technology does not compromise our human values. As a nation, we should look to set the global example for the responsible adoption of AI." The report's findings emphasized the value of a national strategy, a community awareness campaign, safe and accessible digital infrastructure, responsive regulations, and a diversified workforce. Australia's chief scientist Alan Finkel said the report aims "to develop an intellectual context for our human society to turn to in deciding what living well in this new era will mean."
|
New Algorithm Enables More Realistic Sound Effects in VR
Stanford Engineering
Tom Abate
July 30, 2019
Researchers at Stanford University have developed an algorithm that can create sound models in seconds, a breakthrough that will make it cost-effective to simulate sounds for many different objects in a virtual reality (VR) environment. The KleinPAT model can synthesize sounds as realistic as the sounds generated by much slower algorithms in the past, which relied on the boundary element method (BEM) that was too costly for commercial use. The Stanford algorithm calculates sound models hundreds of times faster by avoiding the Helmholtz equation and BEM. The new approach was inspired by 20th-century Austrian composer Fritz Heinrich Klein, who found a way to blend many piano tones and notes into a single, pleasant sound known as the Mother Chord.
|
Hackers Could Use Connected Cars to Gridlock Whole Cities
Georgia Tech Research Horizons
July 29, 2019
Georgia Institute of Technology (Georgia Tech) researchers have determined that hackers could target Internet-connected cars and freeze traffic, creating extremely disruptive gridlock in major cities. The researchers ran simulations of hacking Internet-connected cars in Manhattan, and found that randomly stalling 20% of cars during rush hour would result in total traffic freeze. Hacking just 10% of cars at rush hour would prevent emergency vehicles from expediently moving through traffic; the same would happen with a 20% hack during intermediate daytime traffic. One possible solution could be to split up the digital network influencing the cars, to make it too difficult to access too many cars through a single network, said Georgia Tech's Skanda Vivek.
|
To Save the Internet, Silicon Valley Is Sending It to Space
Discover
Troy Farah
July 30, 2019
Silicon Valley companies hope to sustain Internet operations with satellites. The U.S. National Aeronautics and Space Administration in 1998 developed the Delay/Disruption Tolerant Networking (DTN), or the Bundle Protocol, to prevent disruption-related buffering in space due to extreme distances and orbital variances. Former ACM president Vint Cerf, "the father of the Internet," said, "I think mobile apps that need to transfer data in 'spotty' connectivity would benefit from the Bundle Protocol's patience." Cerf added that he views the protocol as the backbone of manned and robotic missions beyond Earth. DTN transmits data in bursts, avoiding errors and lags by storing the data when a connection is disrupted until it is restored; it also can reduce latency by prioritizing transmissions.
|
Facial Recognition Tech Comes to Schools, Summer Camps
The Wall Street Journal
Julie Jargon
July 30, 2019
Schools and summer camps throughout the U.S. are using facial recognition to bolster security and provide other services for campers and parents. Parents at summer camps can opt into facial recognition services to access photos of their children. Meanwhile, New York's Lockport City School District last year purchased facial recognition software and cameras to identify people not permitted on school grounds and notify authorities. Privacy experts and politicians have reservations about the potential abuse of the technology; developers respond that facial data is not captured and retained as a usable image, but as chains of numbers and letters that can only be decrypted by proprietary software.
*May Require Paid Registration
|
Quantum Computers to Clarify Connection Between Quantum, Classical Worlds
HPCwire
August 1, 2019
Researchers at Los Alamos National Laboratory (LANL) have developed a quantum computing algorithm that provides a clearer understanding of the quantum-to-classical transition, a breakthrough that could help model systems on the brink of quantum and classical worlds, such as biological proteins. The algorithm could also resolve questions about how quantum mechanics applies to large-scale objects. The LANL algorithm determines how close a quantum system is to behaving classically, producing a tool that can be used to search for classicality in quantum systems and understand how quantum systems seem classical to users. Currently, it is essentially impossible to use a classical computer to study the quantum-to-classical transition, but LANL researcher Patrick Coles says scientists can "study this with our algorithm and a quantum computer consisting of several hundred qubits, which we anticipate will be available in the next few years based on the current progress in the field."
|
Tokyo Readies $1B for Cyborg, Other Research
Nikkei Asian Review
Minako Yamashita
July 29, 2019
The Japanese government will solicit research proposals for cyborg and other advanced technologies, with those selected to be funded for up to 10 years, with a budget of 100 billion yen ($921 million) for the first five years. The cyborg component will aim to realize technology to replace human bodily functions that diminish due to age using robotics or living organisms, by 2050. The program also hopes to develop technologies for full automation of human tasks in agriculture, forestry, and fisheries, as well as on construction sites, by 2040. The 25 project areas have been assigned individual deadlines for meeting their objectives, from 2035 to 2060, and will be ranked by priority. The program will draw on new industrial fields based on technical innovations, like digitization and artificial intelligence (AI). Japan aims to develop AI robot systems that automatically detect Nobel Prize-worthy scientific discoveries by 2050.
|
Northeastern University Launches National Program to Boost Number of Women Majoring in Computing
News@Northeastern
Molly Callahan
July 31, 2019
Northeastern University hopes to improve women's representation in technology with its newly established Center for Inclusive Computing. The Center will work with U.S. universities that have large undergraduate computing programs, and fund and support nonprofit universities that produce at least 200 computing graduates annually over the next six years. Northeastern's Carla Brodley said the eventual goal is to create a "critical mass" of female enrollees. The Center will enlist a cadre of technical consultants with experience in best practices for bridging the gender gap, and will collaborate with each grantee university on approaches for recruiting, enrolling, and retaining female students in their computing programs. An advisory council will guide the overarching strategy and direction of the Center, and will include leaders on increasing the number of women in technology such as Jodi Tims, chair of the ACM Council on Women in Computing (ACM-W).
|
How Can You Reliably Spot a Fake Smile? Ask a Computer
University of Bradford
July 30, 2019
Researchers at the University of Bradford in the U.K. have developed computer software that can identify false facial expressions. The software analyzes the movement of a smile across a person's face, and determines whether or not the expression is genuine. First, the software maps a person's face from a video recording, identifying the mouth, cheeks, and eyes of the subject. The program then measures how these facial features move through the progress of the smile, and calculates the differences in movement between the video clips displaying real and fake smiles. Said Bradford’s Hassan Ugail, “An objective way of analyzing whether or not a smile is genuine could help us develop improved interactions between computers and humans—for example, in biometric identification."
|
200 Million Devices—Some Mission-Critical—Vulnerable to Remote Takeover
Ars Technica
Dan Goodin
July 29, 2019
Researchers at California-based security firm Armis have identified 11 vulnerabilities in various versions of VxWorks, an operating system (OS) that runs on more than 2 billion devices worldwide. The researchers found about 200 million Internet-connected devices (some of which may be controlling elevators, medical equipment, and other mission-critical systems) are vulnerable to attacks that give bad actors complete control of those systems. The vulnerabilities, collectively known as Urgent 11, include six remote code flaws and five less-severe issues that allow a range of security issues including information leaks and denial-of-service attacks. None of the vulnerabilities affect the most recent version of VxWorks, or any certified versions of the OS, including VXWorks 653 or VxWorks Cert Edition. "Such vulnerabilities do not require any adaptations for the various devices using the network stack, making them exceptionally easy to spread," according to the Armis researchers.
|
Alphabet's Loon AI-Powered Internet Balloons Learned to Tack Like a Sailor, Move with the Wind
ZDNet
Liam Tung
July 24, 2019
The Internet-beaming balloons that are part of Alphabet's Project Loon have tallied more than 1 million hours in the sky, and the network's artificial intelligence (AI) has learned some unforeseen navigational tricks in the process. Since Google launched its first set of balloons six year ago in New Zealand, they have traveled about 40 million kilometers (about 25 million miles) at an altitude of approximately 200,000 meters (about 124 miles). The balloons deliver 4G connectivity to fill Internet black-spots. That goal is constrained by changing wind conditions; an algorithm controls a solar-powered pump that enables each balloon to rise or descend in order to catch the best wind current to reach a target location in the shortest period of time.
|
|
