Welcome to the December 7, 2015 edition of ACM TechNews, providing timely information for IT professionals three times a week.
Updated versions of the ACM TechNews mobile apps are available for Android phones and tablets (click here) and for iPhones (click here) and iPads (click here).
HEADLINES AT A GLANCE
Memo Decries 'American, White, Male' Culture at Internet Engineering Task Force
Motherboard (12/04/15) Victoria Turk
A recent memo concerning the Internet Engineering Task Force (IETF) cites a culture that has demonstrated consistent hostility to diversity and a bias against women. The authors of the memo call out IETF's primary composition of "American, white, male" technicians whose interactions within the group frequently exhibit "singularly aggressive behavior, often including singularly hostile tone and content." Co-author and IETF participant Narelle Clark says the memo was driven by "a rising discomfort in a lot of people that good ideas and good analysis were being lost through poor behavior. That is, you could see highly capable people being put off by the bullying behavior of others." Among the kinds of behavior cited as unique to the technology community is the "vigorous advocacy for a strongly held technical preference." This can be damaging if it is overly aggressive and designed to intimidate people with other perspectives. IETF security area director Kathleen Moriarty emphasizes the task force has deployed internal diversity initiatives, including a Diversity Design Team that calls out bad community behavior. "There is a very quick and open dialogue that typically follows any occurrence of inappropriate conduct now," she says.
Google and Facebook Race to Solve the Ancient Game of Go With AI
Wired (12/07/15) Cade Metz
Although software designed to play games such as checkers, chess, and some video games can now dominate even the best human players, the ancient game of Go has proven a much trickier challenge. Played on a 19-by-19 board, this Asian version of chess features too many possible moves in any given turn for a current computer to quickly analyze all of the possibilities and their outcomes to find the optimal next move; this method, called a Monte Carlo tree search, has been the primary strategy for many game-playing systems based on artificial intelligence. However, Google and Facebook are hoping to crack Go by taking a different approach. Both companies are turning to the machine-learning techniques they currently use to recognize and categorize pictures, translate text, and process voice commands. Their efforts are promising because they mimic the way professional Go players approach the game--a visual and instinctive approach that depends on identifying patterns, a process well-suited to deep-learning systems. Facebook recently announced a new method blending deep learning with a Monte Carlo tree search, and Google has said it plans to announce its own breakthrough soon.
Facebook's New Tools to Help the Blind Navigate Social Media
San Jose Mercury News (12/04/15) Queenie Wong
Engineer Matt King is blind and works on Facebook's accessibility team. He knows through firsthand experience that even as social media becomes a more integral part of the modern world, it often is hard for people with disabilities to interact with the latest technologies. King says the process of browsing Facebook as a blind person five years ago was so frustrating that he gave up and let his wife control his account for him. The problem has only gotten worse as social media has shifted to focus heavily on sharing and commenting on visual media such as still and animated images and video. Facebook has made strides in making it easier for the visually impaired to interact with the service by instituting changes that make it easier to use the site with a screen reader. Some of the changes include improving HTML headings, creating keyboard shortcuts, and adding alternative text for images. The company also is trying to make visual content more accessible to the blind by using artificial intelligence software to create alternative text for images that describe the setting as well as the people and things within it. The technology, which is being tested in 32 languages, could be publicly released early next year.
MIT News (12/07/15) Larry Hardesty
An untraceable text-messaging system designed to foil even the most powerful adversaries was unveiled by Massachusetts Institute of Technology researchers at ACM's Symposium on Operating Systems Principles (SOSP) in October. The researchers say the system delivers a strong mathematical guarantee of user anonymity, while enabling the exchange of text messages about once a minute. The system buries telltale traffic patterns in spurious data or "noise." A user leaves a message for another user at a predefined location and the other user retrieves it, but the system adds multiple layers of obfuscation to cover the users' trails. The system employs three servers, with each message sent through the system enveloped in three layers of encryption, which are successfully peeled off by each server. The first two servers also randomly permute the order of the messages, and only the final server sees which messages are bound for which memory addresses. However, even if that has been compromised, adversaries cannot tell whose message ended up where. When the initial server passes on the messages, it also generates bogus messages with their own encrypted destinations, as does the second server; this makes it nearly impossible for the adversary to ascertain even whether any of the messages arriving within the same time window arrived at the same destination.
Stanford Scientists Develop 'Shazam for Earthquakes'
Stanford Report (12/04/15) Ker Than
Stanford University researchers have developed Fingerprint And Similarity Thresholding (FAST), an algorithm that could transform how seismologists detect temblors that are not strong enough to register as earthquakes when analyzed by conventional methods. Monitoring microquakes could help scientists predict how frequently, and where, larger quakes are likely to occur. The FAST technique takes all recorded data from a seismic station and divides the continuous signal into segments of a few seconds each, and then compresses the signals into compact representations, or "fingerprints," for rapid processing. The fingerprints are then sorted into separate groups based on their similarities. "Tests we have done on a six-month data-set show that FAST finds matches about 3,000 times faster than conventional techniques," says Stanford professor Greg Beroza. "Larger data-sets should show an even greater advantage." The researchers created FAST using techniques from data mining and machine learning. FAST's scalability stems from the use of locality-sensitive hashing (LSH), which is "a widely used technique for identifying similar items in large data-sets," says Stanford researcher Karianne Bergen. The researchers are scaling up the FAST algorithm to analyze data collected across longer periods of time, from multiple seismic stations, and in more challenging scenarios.
Man With No Limbs Controls Robotic Hand Using Muscle Whispers
New Scientist (12/02/15) Sam Wong
Sam Wilson, a Ph.D. student at Imperial College London, and supervisor Ravi Vaidyanathan are designing new ways for the human body to control prostheses. The team is working to combine inputs from a microphone and an accelerometer in one device to make it easier for people to control prosthetic hands. The researchers say listening to the sounds muscle fibers make as they move against each other makes it easier to get signals out of the body. Their research could make advanced bionic hands, which currently cost about $45,000, more accessible. Wilson and Vaidyanathan are working on a sensor package that costs less than $150. They are working to develop the technology with Alex Lewis, who lost all of his limbs two years ago when a streptococcus infection developed into toxic shock, septicaemia, and necrotising fasciitis, also known as flesh-eating bacteria. The team has already made the hand a bit simpler for Lewis to use by rigging the sensors to enable him to switch grips with an exaggerated upward movement of his arm.
Artificial Intelligence Called In to Tackle LHC Data Deluge
Nature (12/01/15) Davide Castelvecchi
Particle physicists are expected to increasingly incorporate artificial intelligence (AI) in their experiments. When physicists who work on the Large Hadron Collider (LHC) in Switzerland discovered the Higgs boson in 2012, their two largest experiments involved machine learning. Researchers could potentially make discoveries with little human input in the future. Experts in particle physics and AI met in November to discuss how advanced AI techniques could speed discoveries at the LHC. Particle physicists have "realized that they can not do it alone," says Cecile Germain, a computer scientist at the University of Paris South in Orsay. She says the experiments particle physicists conduct in the near future will need to get smarter at collecting their data, not just processing it. Last year, Germain helped organize a contest to write programs that could "discover" traces of the Higgs boson in a set of simulated data; it attracted submissions from more than 1,700 teams. When ATLAS and CMS, the LHC's two largest experiments, discovered the Higgs boson, they did so in part using machine learning. The algorithms were primed using simulations of the debris from particle collisions, and learned to identify patterns produced by the decay of rare Higgs particles among millions of more mundane events.
A Smarter Kind of Crash Test Dummy
Technology Review (12/03/15) Simon Parkin
Car crash simulations are being run on a supercomputer using a combination of actual vehicle, scene, and medical data by Wake Forest University researchers. The digital models permit scientists to examine the effects of a crash to a far greater degree than crash test dummies by testing diverse body shapes and sizes and different body positions at the moment of collision. "My hope is that the research will provide a cost-effective solution for evaluating new and existing automotive safety features," says Wake Forest professor Ashley Weaver. The model can measure the risk of bone fractures and damage to soft tissue and organs, which crash test dummies cannot. The researchers employed a digital model containing about 1.8 million elements, which combine to replicate the human form. They then ran simulations until the model precisely emulated the effects of different crashes on real-world victims. "Digital crash dummies [allow us] to determine the best methods to modify vehicle chassis, interiors, seats, headrests, safety belts, dashes, and active safety systems, such as airbags, to improve safety very early in the vehicle-design process," says Nvidia engineer Bill Veenhuis. He notes testing with real dummies is then done to validate the digital dummy tests.
When Apps Talk Behind Your Back
UCR Today (12/03/15) Sarah Nightingale
University of California, Riverside (UCR) researchers recently conducted a study, which found nearly 9 percent of popular apps downloaded from Google Play interact with websites that could compromise users' security and privacy. The researchers are developing the Android URL Risk Assessor (AURA) so users can evaluate the riskiness of individual apps before downloading them. They conducted a large-scale analysis of URLs embedded in 13,500 free android apps downloaded from Google Play. "We focused on a relatively neglected aspect of security research, which is the potential for good apps to leak personal information through the sites they interact with," says UCR professor Michalis Faloutsos. The researchers used AURA to identify more than 250,000 URLs accessed by the 13,500 apps, which were then cross-referenced for trustworthiness using VirusTotal, a database of malicious URLs, and Web of Trust, a website rating system. The researchers found almost 9 percent of the popular apps interacted with malicious URLs, and 15 percent talked to bad websites. In addition, 73 percent of the apps talk to low-reputation websites, and 74 percent talked to websites containing material not suitable for children. The researchers note that although these results are troubling, they only show users are potentially exposing themselves to risk, not that each of these interactions would necessarily result in negative consequences.
Quantum Dots Could Bridge Gap Between Electronic and Quantum Computers
Network World (12/02/15) Jon Gold
Researchers at the Los Alamos National Laboratory, Stanford University, and Technical University of Munich (TUM) have announced a solution to a key challenge of quantum computing technology that is built out of common semiconductor materials. Some proposed quantum computer systems employ photons, ions, or other atoms that can be manipulated to represent different values. The just-announced quantum computer uses a single electron trapped in a nanostructure, or quantum dot, made from standard semiconductor materials to store information in the form of the electron's spin. The method's application was previously constrained by a piezoelectric effect, which led to spin fluctuation and rapid data corruption, according to TUM researcher Alexander Bechtold. He says this corruption is offset by a magnetic field of 1.5 teslas, which "corresponds to the magnetic field strength of a strong permanent magnet. It stabilizes the nuclear spins and the encoded information remains intact." Fellow TUM researcher Jonathan Finley reports these semiconductor quantum dots have a unique advantage because "they harmonize perfectly with existing computer technology since they are made of similar semiconductor material."
DARPA Seeks Toolset for Complex Adaptive Systems
Government Computer News (12/01/15) Mark Pomerleau
The U.S. Defense Advanced Research Projects Agency's (DARPA) Complex Adaptive System Composition and Design Environment (CASCADE) program will address current shortcomings and change the way complex systems are designed for real-time resilient response to dynamic environments. In addition, CASCADE will provide a unified view of system behavior and a formal language for complex adaptive system composition and design. "CASCADE could help develop models that would provide civil authorities, first responders, and assisting military commanders with the sequence and timing of critical actions they need to take for saving lives and restoring critical infrastructure," says DARPA program manager John Paschkewitz. He notes CASCADE also could help improve forward-deployed medical care by ensuring various components are accurately modeled and understood. Paschkewitz says CASCADE will replace existing modeling and design tools, which are static and unable to represent the complexity of many modern systems. In order to advance the CASCADE program, DARPA currently is looking for expertise from industry in applied mathematics, operations research, modeling and applications responsive to challenges, search and rescue platforms and modeling, and adaptive and resilient urban infrastructure.
Abstract News © Copyright 2015 INFORMATION, INC.
To submit feedback about ACM TechNews, contact: firstname.lastname@example.org
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.