Welcome to the August 28, 2017 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."

satellite Breakthrough Could Make Quantum Data Transmission a Reality
Tara Seals
August 24, 2017

Researchers at the University of Ottawa in Canada have relayed a quantum-secured message containing multiple bits of information for each photon through the air above a city, a significant milestone in the quest to realize quantum data transmission. "Our work is the first to send messages in a secure manner using high-dimensional quantum encryption in realistic city conditions, including turbulence," says University of Ottawa professor Ebrahim Karimi. "The secure, free-space communication scheme...could potentially link Earth with satellites, securely connect places where it is too expensive to install fiber, or be used for encrypted communication with a moving object, such as an airplane." The team successfully demonstrated four-dimensional quantum encryption over a free-space optical network between two buildings 0.3 kilometers apart. The messages had an 11-percent error rate, below the 19-percent threshold required to maintain a secure connection. In addition, 1.6 times more information was transmitted for every photon than two-dimensional encryption.

Full Article
Revolutionary Approach Brings 3D Sound Into the Living Room
University of Surrey
August 24, 2017

Researchers at the universities of Surrey, Salford, and Southampton in the U.K. have developed the Media Device Orchestration (MDO) concept, which helps create a three-dimensional or "spatial audio" experience using common home devices. The team says the technology works by isolating different "objects" within audio content, and connecting them to separate speakers available around the room. The researchers believe the MDO method could enable users to listen to media in a far more immersive, multi-layered, and exciting way. In addition, the team notes the object-based audio technology has enabled them to access each separate part of an audio scene, and intelligently route them to improve the listener experience. "Our aim is to take spatial audio out of the lab and into people's homes, and give users the impression of being at the heart of the action while in their living room," says the University of Surrey's Philip Jackson.

Full Article
Mimicking the Reflexive Detection Ability of the Animal Visual System for Computer Detection of Moving Objects
August 22, 2017

Researchers at King Abdullah University of Science and Technology (KAUST) in Saudi Arabia have developed a statistics-based computational framework that minimizes detection time for a given level of detection accuracy, in an attempt to reproduce the ability of animals to reflexively detect moving objects. "We are the first to address this problem in computer vision, which is important not only in modeling the animal visual system, but also in real-world applications, such as robotic control and surveillance," says KAUST's Ganesh Sundaramoorthi. The researchers adapted the "quickest detection framework," a statistical model that converts the problem into a stochastic process with the minimum detection time set according to a statistical detection threshold. The quickest detection framework was originally meant for one-dimensional data, instead of for the infinite dimensional time series of sequential video images. However, the researchers constructed optimization tools to address the problem with reasonable speed.

Full Article
Teaching Robots to Learn Teaches the Students Too
THE Journal
Dian Schaffhauser
August 22, 2017

Researchers at the Technion Israel Institute of Technology have developed a way for high school students and first-year engineering students to learn robot intelligence technologies. The researchers demonstrated how students taught their robots to acquire skills by implementing a reinforcement learning (RL) process using simulation modeling and cloud communication. In the first phase, first- and second-year students performed an RL activity in which a humanoid robot learned to adapt its body tilt angle for lifting different weights. In the second phase, 11th graders constructed animal-like robots and implemented various scenarios that used the approach tested in the first phase. In the final stage, university students applied RL, three-dimensional modeling, and cloud-based operations to a project in which a humanoid robot learned to manipulate multiple joints while maintaining its stability. The development of learning robots is highly engaging to students, and those concepts and technologies are within the grasp of understanding for the students.

Full Article

drone in warehouse Drones Relay RFID Signals for Inventory Control
MIT News
Larry Hardesty
August 25, 2017

Researchers at the Massachusetts Institute of Technology have developed RFly, a system enabling small aerial drones to scan radio-frequency identification (RFID) tags tens of meters away while determining their whereabouts with an approximately 19-cm average error rate. The team envisions the system facilitating better warehouse inventory control. An RFID tag lacks an independent power source, and the researchers devised an analog filter to subtract the base transmission frequency from the signals that reach the reader and then partition the low-frequency and high-frequency components, after which the signal from the tag is added back onto the base frequency. Each drone has its own RFID tag so they can alternate between transmitting the reader's signal to a tagged item and simply letting its own tag bounce the signal back, so the reader can estimate the drone's contribution to the phase shift and subtract it. The researchers presented RFly last week at the ACM Special Interest Group on Data Communications (SIGCOMM 2017) conference in Los Angeles.

Full Article

computer system backdoor, illustration Researchers Built an Invisible Backdoor to Hack AI's Decisions
Dave Gershgorn
August 25, 2017

Researchers at New York University (NYU) have demonstrated a cyberattack against artificial intelligence (AI) that controls driverless cars and image-recognition systems by installing an invisible backdoor in the software. The team says AI from cloud providers could be infected with these backdoors, and would function normally until a predetermined trigger causes the software to mistake one object for another. The NYU method instructs a neural network to identify the trigger with a stronger confidence than what the neural network is supposed to be perceiving, thus preempting correct signals in favor of incorrect ones. The complexity of the network is such that there is currently no test for this form of tampering. The researchers note this hack could make cloud customers more suspicious of the training protocols on which their AIs rely. "Outsourcing work to someone else can save time and money, but if that person isn't trustworthy it can introduce new security risks," says NYU professor Brendan Dolan-Gavitt.

Full Article

A brain scan Artificial Intelligence Predicts Dementia Before Onset of Symptoms
McGill Newsroom
Justin Dupuis
August 22, 2017

Researchers at McGill University in Canada say they used artificial intelligence techniques and big data to develop an algorithm that recognizes the signatures of dementia two years before its onset, using a single imaging scan of the brain of at-risk patients. The researchers say the technology could change how physicians manage patients and accelerate treatment research into Alzheimer's disease. "By using this tool, clinical trials could focus only on individuals with a higher likelihood of progressing to dementia within the time frame of the study," which will greatly reduce the cost and the time required to conduct these studies, says McGill professor Serge Gauthier. The researchers used data from the Alzheimer's Disease Neuroimaging Initiative, a global research effort in which patients agreed to complete a variety of imaging and clinical assessments. The team used the data to train an algorithm to identify which patients would develop dementia, with an accuracy of 84 percent, prior to symptom onset.

Full Article
Cryptographers and Geneticists Unite to Analyze Genomes They Can't See
Scientific American
Jesse Dunietz
August 22, 2017

Stanford University cryptographer Dan Boneh and geneticist Gill Bejerano have developed a secure multiparty computation (SMC) algorithm to discover disease-linked genetic mutations without actually seeing anyone's genome, making the protection of genomic privacy practical. The researchers combined the SMC technique of Yao's protocol with genomics to yield three distinct types of privacy-securing analyses, and in one example sought the most common mutations in patients with four rare diseases. In each case they revealed the known causal gene, and also diagnosed a baby's sickness by comparing his genome with those of his parents. In addition, the team found a previously unknown disease gene by having two hospitals search their genomic databases for patients with identical mutations, without disclosing their full genomes. Stanford's Karthik Jagadeesh says genomic privacy not only eases the worries of genome database caretakers, but also offers help for "second- and third-degree relatives, [who] share a significant fraction of the genome."

Full Article

An Intel self-drving car Intel Probes Man-Machine 'Trust'
EE Times
Junko Yoshida
August 24, 2017

Intel recently performed a qualitative study with 10 volunteers to gauge the level of trust between passengers and self-driving cars. Among the insights gained from the study was that although volunteers appreciated knowing what the vehicle was doing during their ride, they also felt it was providing too much information. Participants also noted subtle differences between a machine and a human regarding safety. Intel's Jack Weast says although many volunteers admitted to engaging in less-than-safe driving behavior, "many participants concluded that boring is safe." Weast also notes participants liked the car's voice interface, but they also wanted to communicate with the vehicle, suggesting that gesture-based comprehension may be needed, in addition to natural language and contextual understanding. Linley Group analyst Mike Demler sees a technological bias inhibiting the study, saying, "If you identify wants first, then demonstrate that your product can deliver on those wants, trust comes naturally."

Full Article
UTA Researchers Are Refining Their Automated Fact-Checking System
UTA News Center
Jeremy Agor
August 24, 2017

Researchers at the University of Texas, Arlington (UTA) and Duke University have received a U.S. National Science Foundation grant to expand the ClaimBuster fact-checking system to cover new forms of media and publications and become as automated as possible. UTA professor Chengkai Li wants to enhance ClaimBuster to automatically verify various types of new claims that have not been checked before against knowledge databases and determine whether they are truthful. The researchers will model factual claims and generate internal representations, and then produce taxonomies of claim templates in different domains. Afterwards, the team will develop novel methods for basing structured queries, keyword queries, and natural-language questions on claims, to be compared with the answers embedded in the claims themselves, to verify if the claims are valid. Finally, they will develop methods to identify counter-arguments to claims and interesting "factlets" from datasets that will enable them to spot correct but misleading "cherry-picked" claims.

Full Article
How Machine Learning Could Help to Improve Climate Forecasts
Nicola Jones
August 23, 2017

Researchers are combining artificial intelligence (AI) and climate science to create deep-learning analyses of weather patterns, and a September conference in Colorado will evaluate the state of climate informatics. Last year, a team at the Lawrence Berkeley National Laboratory (LBNL) reported on the first use of a deep-learning system to identify tropical cyclones, atmospheric rivers, and weather fronts, demonstrating the algorithm could replicate human expertise. The researchers plan to apply similar methods to analyze a broader range of extreme weather events, including those as yet uncategorized. The team's objective is to better rank and predict shifts in these phenomena related to climate change. Similar work by George Washington University's Claire Monteleoni has led to machine-learning algorithms that produce weighted averages of about 30 climate models used by the Intergovernmental Panel on Climate Change. LBNL's William Drew Collins envisions AI algorithms being used to test next-generation climate models, and some scientists are using them for weather forecasts.

Full Article

First responders checking a tablet. Can Twitter Aid Disaster Response? New IST Research Examines How
Penn State News
Erin Cassidy Hendrick
August 17, 2017

Researchers at Pennsylvania State University's (PSU) College of Information Sciences and Technology say they are studying new ways to use social media data to help communities respond during natural disasters. The researchers compared 10 million tweets sent out during Hurricane Sandy and the corresponding power outage information provided by utility companies in New York, New Jersey, and Pennsylvania to create a system for event detection. The team then isolated tweets with the terms "power," "outage," "electri," and "utility," and organized them chronologically. The two datasets were found to have a moderate-to-strong correlation, and the researchers believe it is possible Twitter was able to report power outages more quickly. "The goal [of this research] is to demonstrate that if the data stream changes, you can see what just happened," says PSU Ph.D. candidate Nick Lalone. He thinks officials could potentially use this type of tool to deploy resources without having to wait for reports from utility companies.

Full Article
Ada's Legacy
ACM Learning Center

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: technews@hq.acm.org