Welcome to the November 3, 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."
AI Can Increase Success Rates for Cancer Detection, Says Report
Computing
Nicholas Fearn
October 31, 2017


Researchers from Showa University in Japan have demonstrated a technique using artificial intelligence (AI) to identify colorectal cancer before it becomes malignant. The team concentrated on a colorectal polyp magnified by a factor of 500, and the AI-powered tool could detect its variations and other details of interest. Once this was done, the tool compared the variations to outcomes from a database of more than 30,000 photos of precancerous and cancerous cells. The images were used to train the machine-learning program until it was able to reach a diagnosis within less than a second. An overall accuracy rate of 86 percent was maintained throughout the study, a notable accomplishment considering that the patients who were observed had previously been diagnosed. The program explored 306 polyps in total, providing 94 percent sensitivity, 79 percent specificity, and positive and negative predictive values of 79 percent and 93 percent in other findings.

Full Article
*May Require Free Registration
Graphene Enables High-Speed Electronics on Flexible Materials
Chalmers University of Technology
Michael Nystås
October 31, 2017


Researchers at Chalmers University of Technology in Sweden have used graphene transistors on plastic substrates to develop a flexible detector for terahertz frequencies. The team says the first-of-its-kind device can extend the use of terahertz technology to applications that will require flexible electronics, such as wireless sensor networks and wearable technology. The detector has several unique features: for example, at room temperature, it detects signals in the frequency range of 330 to 500 gigahertz. In addition, it is translucent and flexible, and can be used in a variety of applications. The researchers say the unique electronic features of graphene, combined with its flexible nature, make it a promising material to integrate into plastic and fabric, something that will be vital in a future connected world run by the Internet of Things. The researchers' breakthrough is an important step forward for graphene in the terahertz era, and a boon for high-performance and inexpensive flexible terahertz technology.

Full Article

Man and woman looking at computer screens Where the STEM Jobs Are (and Where They Aren't)
The New York Times
Steve Lohr
November 1, 2017


Recent studies indicate the most high-earning jobs in the science, technology, engineering, and math (STEM) fields are in computing. A Glassdoor ranking of the median base salary of workers in their first five years of employment by undergraduate major found computer science topped the list at $70,000, followed by electrical engineering. Biochemistry and biotechnology were among the lowest-paying majors, and the study also confirmed the chronic underrepresentation of female STEM majors. "There is a huge divide between the computing technology roles and the traditional sciences," notes Glassdoor's Andrew Chamberlain. According to a LinkedIn study, the top 10 skills most in demand last year were all computer skills, including expertise in cloud computing, data mining and statistical analysis, and writing smartphone applications. Meanwhile, University of Washington professor Edward Lazowska (recipient of the ACM Presidential Award for 2005 and the ACM Distinguished Service Award for 2009) analyzed U.S. Bureau of Labor Statistics employment forecasts and found that in the decade ending in 2024, 73 percent of STEM job growth will be in computer occupations.

Full Article
*May Require Free Registration

Professional StarCraft player Song Byung-gu playing against a real-time AI system Humans Are Still Better Than AI at StarCraft--for Now
Technology Review
Yoochul Kim; Minhyung Lee
November 1, 2017


A recent "StarCraft" tournament in South Korea pitting a human champion against four artificial intelligence (AI) programs tested the AIs' skills in playing a real-time computer game requiring the simultaneous use of memory, strategizing, and advanced planning within a constrained, simulated environment. Professional player Song Byung-gu defeated the four AIs in less than 30 minutes total, even though the bots could move faster as well as multitask. Song notes while human gamers initiate combat only when they stand a chance of victory, the AIs attempted to keep their units intact without taking any bold actions. Sejong University professor Kim Kyung-joong says the bots were partly limited by the lack of widely available StarCraft training data, and other experts predict AIs eventually will beat professional players once they get sufficient training. "When AI bots are equipped with [high-level] decision-making systems like AlphaGo, humans will never be able to win," says University of Science & Technology professor Jung Han-min.

Full Article
*May Require Free Registration
Cornell Professor Claims Blockchain Advances With Thunder Token Debut
CoinDesk
Rachel Rose O'Leary
November 2, 2017


Cornell University professor Elaine Shi has announced an upcoming blockchain project called thunder token, designed to remedy blockchain's inability to accommodate significant user numbers. Shi says thunder token could potentially realize speeds 1,000 times greater than currently available technologies, without fully relying on a blockchain. She proposes a split framework so that transactions are confirmed rapidly, with the blockchain only employed in case of emergencies. The rest of the time, thunder token will use a system of agents that follows the instructions of a "leader" to vote on which transactions are processed according to the rules. Shi describes thunder token as a fast, scalable protocol--the "dream of large-scale consensus"--that could be useful for banks as well as corporations. "With cryptocurrencies like bitcoin and ethereum, we now have empirical evidence that distributed consensus is now possible on a really large scale on the Internet," Shi notes. "We want to replicate that success."

Full Article
Real Security Requires Strong Encryption--Even If Investigators Get Blocked
The Conversation
Susan Landau
November 1, 2017


Although the U.S. government favors data security via encryption, the U.S. Federal Bureau of Investigation (FBI) is demanding encryption systems have "exceptional access" wherein law enforcement can use warrants to bypass those systems, writes Tufts University professor Susan Landau. She contends the FBI would be less likely to do privacy a disservice by following the National Security Agency's (NSA) lead. NSA adapted to the new reality by exploiting unpatched flaws to breach targets' encryption and relying on communication metadata. Landau says the Bureau also can leverage other types of non-encrypted data, and use computer systems and modern software to pinpoint and analyze that data. "Increasingly, a number of former senior law enforcement and national security officials have come out strongly in support of end-to-end encryption and strong device protection (much like the kind Apple has been developing), which can protect against hacking and other data theft incidents," Landau stresses.

Full Article

Tablet with showing news site on top of a newspaper Study Finds Fringe Communities on Reddit and 4chan Have High Influence on Flow of Alternative News to Twitter
UAB News
Tiffany Westry Womack
November 1, 2017


Researchers at the University of Alabama at Birmingham (UAB), Cyprus University of Technology, University College London in the U.K., and Telefonica Research have conducted the first large-scale measurement of how mainstream and alternative news flows through multiple social media platforms. The researchers analyzed millions of posts on Twitter, Reddit, and 4chan, and found fringe communities within 4chan and Reddit have a surprisingly large influence on Twitter. "These smaller, fringe communities on Reddit and 4chan serve as an incubation chamber for a lot of information" and many online hoaxes and false or misleading stories can be traced back to users on these platforms, says UAB professor Jeremy Blackburn. The researchers also found Twitter heavily influences posting of URLs from alternative news sites on the other social platforms, and is the single most influential source for most of the other Internet communities. The results of the study were published at this week's ACM Internet Measurement Conference (IMC 2017) in London.

Full Article
Scientists Take Important Step With Self-Folding Objects
Delft University of Technology
October 31, 2017


Researchers at the Delft University of Technology (TU Delft) in the Netherlands have used origami techniques and three-dimensional (3D) printing technology to create flat structures that can fold themselves into 3D structures. The structures self-fold according to a pre-planned sequence, a method that normally requires expensive printers and special materials. However, the TU Delft researchers developed a new technique that requires only a common 3D printer and polylactide (PLA), a common 3D printing material. "If the goal is to create complex shapes, some parts should fold sooner than others," and this requires a method that includes programming time delays into the material, says TU Delft's Amir Zadpoor. The process is known as sequential shape-shifting, and the researchers achieved this by simultaneously printing and stretching the material in certain spots. The stretching is stored inside the material as memory, and when it is heated up the memory is released and the material returns to its original state.

Full Article
'Instant Replay' for Computer Systems Shows Cyberattack Details
Georgia Tech News Center
John Toon
October 30, 2017


Researchers at the Georgia Institute of Technology (Georgia Tech) have developed Refinable Attack INvestigation (RAIN), a software system designed to automate the process of assessing the extent and impact of cyberattacks so investigators can quickly and accurately identify how intruders entered a network, what data they stole, and which systems were compromised. The researchers say RAIN provides multiple levels of detail, helping automated searches parse through information at a high level to find the specific events for which more detailed data is reproduced and analyzed. The system continuously monitors a dataset and logs events that it recognizes as potentially interesting. "These fine-grained analyses, which can be extremely useful when investigating an attack, would be too expensive to perform on a deployed system; but our hierarchical approach allows us to run these analyses offline, and only when necessary," says Georgia Tech's Alessandro Orso. The research was presented this week at the ACM Conference on Computer Communications Security (CCS 2017) in Dallas, TX.

Full Article
Crowdsourcing Big-Data Analysis
MIT News
Larry Hardesty
October 30, 2017


Researchers at the Massachusetts Institute of Technology (MIT) have developed FeatureHub, a collaboration tool to make feature identification more efficient and effective. FeatureHub lets data scientists log onto a central site, review a problem, and propose features, and then tests various feature combinations against target data to determine which are the most useful for a given predictive task. During testing, the researchers recruited 32 analysts with data science experience, who each spent five hours working with the system and used it to propose candidate features for each of two data-science problems. FeatureHub generated predictive models that were tested against those submitted to a data-science competition, and the models were within three and five points of the winning entries for the two problems. "The concept of massive and open data science can be really leveraged for areas where there's a strong social impact but not necessarily a single profit-making or government organization that is coordinating responses," says MIT's Micah Smith.

Full Article

Att logo in store window AT&T Joins the Open Source Artificial Intelligence Arms Race
Wired
Klint Finley
October 30, 2017


AT&T's new Acumos artificial intelligence (AI) platform is designed to function as both a directory for finding and sharing AI models and a system for tailoring and linking those models in useful ways. AT&T's Mazin Gilbert says Acumos, whose rollout is slated for early next year, is designed to serve both programmers and non-programmers. Acumos is being developed in collaboration with consulting firm Tech Mahindra, and the directory and underlying code will be hosted by the Linux Foundation. AT&T Communications CEO John Donovan says the motivation for releasing Acumos is in keeping with AT&T's new focus on a more open source operational paradigm as its data traffic expands. Although the Linux Foundation will support a public version of Acumos that anyone can use, organizations also can generate private versions. AT&T says it will use Acumos internally to help developers in different company units leverage the AI work their colleagues have produced.

Full Article
Smart Access to Homes and Cars Using Fingers
Rutgers Today
Todd B. Bates
October 29, 2017


Researchers at Rutgers University have developed VibWrite, a low-cost, minimal-power smart access system that can authenticate users by sensing their finger vibrations. "Everyone's finger bone structure is unique, and their fingers apply different pressures on surfaces, so sensors that detect subtle physiological and behavioral differences can identify and authenticate a person," says Rutgers professor Yingying Chen. The researchers note VibWrite combines passcode, behavioral, and physiological properties, and it permits users to opt for personal identification numbers, lock patterns, or gestures as methods for gaining secure access. Any solid surface can be used as a verification surface with the device, which includes an inexpensive vibration motor and receiver. During testing, VibWrite authenticated legitimate users with more than 95-percent accuracy and a false positive rate of less than 3 percent. A paper on the technology was presented Sunday at the ACM Conference on Computer and Communications Security (CCS 2017) in Dallas, TX.

Full Article

Login page with captcha box CAPTCHA Is Dying. This Is How It's Being Reinvented for the AI Age
Wired.co.uk
Matt Burgess
October 26, 2017


Completely Automated Public Turing Tests to tell Computers and Humans Apart (CAPTCHAs) appear to be heading toward obsolescence, with Vicarious demonstrating an algorithm that can solve the test with a recursive cortical network by using 5,000 times fewer training images than other methods. This and other advances are making CAPTCHAs less secure and less relevant, pushing researchers toward a rethinking of the test to strengthen it in the era of machine learning. Nan Jiang at Bournemouth University in the U.K. has developed a mobile CAPTCHA named Tapcha, which builds on the traditional distorted text approach while still relying on human knowledge. "We use this approach to create the instruction," Nan says. He believes a machine can only defeat this approach by understanding what is occurring in the image and coming to a conclusion on its own, a breakthrough that currently is very difficult to realize with artificial intelligence.

Full Article
ACM Interactions Magazine
 
ACM Publications
 

Association for Computing Machinery

2 Penn Plaza, Suite 701
New York, NY 10121-0701
1-800-342-6626
(U.S./Canada)



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