Welcome to the April 21, 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).
|
Baidu Will Release a Free Operating System for Self-Driving Cars
Technology Review
Will Knight
April 19, 2017
China-based Internet firm Baidu has announced it is releasing its self-driving platform, known as "Apollo," as open source with the goal of fast-tracking the technology's progress while enhancing the company's role in supplying key elements such as mapping and machine-learning systems. Baidu's move could lead to a more open effort to researching self-driving cars and lower the requirements for developing advanced driver-assist systems, as well as self-driving prototypes. Much of the technology will be released this July, but certain features, such as mapping and machine-learning services, will be made accessible through an application programming interface that Baidu will control. "The fundamental motivation is [to create] an open ecosystem that will accelerate the pace of innovation toward fully autonomous driving, which will have profound changes to our society," says Baidu president Qi Lu.
|
Twin Research Reveals Which Facial Features Are Most Controlled by Genetics
King's College London
April 19, 2017
Researchers at King's College London in the U.K. reviewed three-dimensional (3D) facial simulations of about 1,000 female twins, and found the shapes of the end of the nose, the area above and below the lips, cheekbones, and the inner corner of the eye were highly shaped by genetics. "By combining 3D models of the face with a statistical algorithm that measures local changes in shape, we have been able to create detailed 'face heritability maps,'" says King's College London professor Giovanni Montana. "These maps will help identify specific genes shaping up the human face, which may also be involved in diseases altering the face morphology." The team scanned twins' faces and measured their curvature at thousands of points. By observing which parts of the face are most similar in shape in a pair of identical twins, they calculated the probability the shape of that area is genetically influenced.
|
An Indian Computer Scientist Helped Build an Alternative Programming Language in Bengaluru
Quartz
Maria Thomas
April 20, 2017
Julia Computing CEO Viral Shah co-created the free and open source Julia programming language, which is gaining traction as a substitute for more entrenched languages such as MATLAB and Python. Julia is designed for developing advanced analytical apps while maintaining ease of use. Shah says Julia overcomes a key problem faced by users of analytical apps--the two-language problem--which can be especially frustrating for users who are not programmers by training. "What ends up happening is...you write your algorithm first in a high-level language and then rewrite it in a low-level language for performance, or taking it live," Shah says. "And so you end up writing everything twice." So far, only a handful of Indian universities currently are teaching Julia, and Shah blames the curriculum's centrally-driven composition, which typically leaves professors with little autonomy to choose the kind of programming languages they want to teach.
|
Traditional Polling Methods Not Adding Up
The Daily Illini
Charlotte Collins
April 20, 2017
University of Illinois (UI) professor Sheldon Jacobson analyzed publicly available data to characterize a widening gap between traditional and modern polling in terms of methods and outcomes. He says the phase-out of landline telephones has challenged the idea that polls are an accurate reflection of popular sentiment. "This has been a challenge over the last few election cycles as [we ask], 'Are we getting the right information from the right sources that accurately represent the population?'" Jacobson says. "As we saw in 2016, we did not." UI's William Widjaja developed an algorithm that pulls data from social media and ranks freely shared thoughts from -1 to 1 to ascertain how negatively or positively users view a topic or key phrase. Widjaja believes the algorithm eliminates leading questions and no-response prejudices. "I think part of what I'm doing will become the new method of how people get information," he says.
|
DARPA Wants to Teach Machines to Handle New Situations Like 'Biological Systems' Do
FedScoop
Samantha Ehlinger
April 19, 2017
The U.S. Defense Advanced Research Projects Agency (DARPA) is creating the Lifelong Learning Machines (L2M) research program to develop ways to teach machines to learn while they are operating and to apply their knowledge to new situations, similar to the way biological systems do. The agency is currently accepting research proposals for the program's first funding opportunity. The goal is to develop a system that can learn in such a way that is safer, more functional, and increasingly relevant to U.S. Department of Defense applications, including adapting quickly to unforeseen circumstances, changes to the mission, and improving performance via a system's fielded lifetime experience. The L2M project is divided into two technical areas--one is oriented around developing the lifelong learning system, while the other concentrates on finding ways biological systems learn to inform new algorithms that help machines learn in the same way.
|
Networking App for Skilled Refugees Wins First Prize at NYU Abu Dhabi International Hackathon
Al-Bawaba
April 19, 2017
Hiat, a networking app to connect skilled refugees and the local community in a gig economy platform, won the top prize at New York University Abu Dhabi's 7th Annual International Hackathon for Social Good in the Arab World. The second-place prize winner was 3ndi Headache, an app for automating triage nursing to shorten response time and improve patient-handling efficiency. Tying for third place were the teams who developed the An Nahr and Tarakeeb apps. An Nahr is an interactive hardware/software app focusing on water wastage during ablution at mosques. It attempts to raise awareness of this problem by delivering feedback to users. Tarakeeb is an automated grammar-checking app for helping to teach or improve Arabic language. The audience award recipient, Teslam, is a service to visualize and share actionable data in the Arab world related to non-communicable diseases via big data and machine learning.
|
New Quantum Liquid Crystals May Play Role in Future of Computers
Caltech
Whitney Clavin
April 20, 2017
Researchers at the California Institute of Technology's (Caltech) Institute for Quantum Information and Matter have detected the first three-dimensional (3D) quantum liquid crystal, which they say has potential uses in future quantum computers. The Caltech team, in collaboration with scientists at Oak Ridge National Laboratory and the University of Tennessee, found the 3D crystal is distinct from two-dimensional (2D) crystals by having electrons that not only differentiate between the x-, y-, and z-axes, but also exhibit different magnetic properties depending on whether they flow forward or backward on a given axis. The researchers say the discovery could have applications in spintronics research and help address some of the challenges of constructing a quantum computer. "In the same way that 2D quantum liquid crystals have been proposed to be a precursor to high-temperature superconductors, 3D quantum liquid crystals could be the precursors to the topological superconductors we've been looking for," says Caltech professor David Hsieh.
|
Changing the Game
Pacific Northwest National Laboratory
Eric Francavilla
April 20, 2017
Researchers at Pacific Northwest National Laboratory (PNNL) and the University of Houston have developed a new architecture for three-dimensional (3D) stacked memory that increases 3D rendering speeds up to 65 percent. The researchers say they achieved this breakthrough by exploiting the hardware feature known as "processing in memory." 3D stacked memory has an added logic layer that enables the memory to do some processing. The researchers found the final step in rendering, called anisotropic filtering, creates the most traffic. The team shifted the anisotropic filtering to the first step in the process, which led to the greatest boost in performance. "We're pushing the boundaries of what hardware can do," says PNNL's Shuaiwen Leon Song. "And though we tested our idea on video games, this improvement ultimately benefits science." Song is investigating how high-performance computers can simulate dynamic information networks and how to predict changes in these graphs.
|
UT Researcher Creates Robotic Cheetah
University of Twente (Netherlands)
Joost Bruysters
April 19, 2017
Researchers at the University of Twente (UT) in the Netherlands are developing a scaled-down robotic version of a cheetah that uses only about 15-percent more energy than the real animal. As the fastest land animal in the world, the cheetah makes very efficient use of its energy. "I wanted to create a robot that runs the same way, with the aim of applying this knowledge to the development of new robots," says UT researcher Geert Folkertsma. He says the research provides valuable knowledge for optimizing robots of the future. In addition, the knowledge gained from the project can be applied to rehabilitation robots or advanced prosthetics that are equipped with robotics. The researchers' prototype can reach a speed of about one kilometer an hour. "We...need to develop robots that can walk and when it comes to moving around efficiently, there's a lot we can learn from the cheetah," Folkertsma says.
|
Where the Individual's Data Security and Law Enforcement Needs Meet
CORDIS News
April 18, 2017
The European Union's (EU) Secure Information Sharing Sensor Delivery Event Network (SISSDEN) project aims to create a data collection and sharing system that can identify potential threats, make information on those threats available, and tell people how to deal with them. Although similar services already exist, the SISSDEN project will make all this information available for free. The project currently is developing a large, distributed sensor network based on state-of-the-art honeypot and darknet technologies. In addition, the project is using enhanced sandbox systems and a high-throughput automated data processing and sharing center. The researchers say SISSDEN will enable the development of metrics that can be used to establish the scale of some measurable security issues within the EU. The project also is compiling a curated reference dataset that seeks to provide a groundbreaking, high-value resource to academia and researchers in the field.
|
Pinning Down Abuse on Google Maps
UCSD News (CA)
Ioana Patringenaru
April 18, 2017
A collaboration between researchers at the University of California, San Diego and Google has enabled the Internet company to reduce the number of fraudulent businesses in Google Maps by 70 percent. The researchers analyzed more than 100,000 fraudulent listings to determine how scammers had been able to avoid detection and how they made money. The team discovered a new form of blackhat search engine optimization that targets local listing services such as Google Maps. For example, when users run a search on a mobile device, the search engine uses their physical location as one of the inputs to decide which results to display. The scammers exploit this by using fake locations to make it look like their business is in close proximity to the user doing the search. The team says scammers make money when they get called to help a user based on a fake listing.
|
Our Ability to Keep 'Em Guessing Peaks Around Age 25
Scientific American
Jordana Cepelewicz
April 14, 2017
People's ability to behave randomly reaches its zenith at age 25, and then gradually erodes before experiencing a swift decline at age 60, according to researchers from the Algorithmic Nature Group at LABORES Scientific Research Lab for the Natural and Digital Sciences in France. The team had volunteers of varying ages compete against computers in random pattern production and recognition tasks. Subjects had to create sequences of coin tosses and die rolls they thought would appear random to another person, guess which card would be drawn from a randomly shuffled deck, and point to circles on a screen and color in a grid to form a seemingly random design. "At age 25, people can outsmart computers at generating this kind of randomness," says LABORES researcher Hector Zenil. The research is the first to use an algorithmic approach to quantifying complexity over a continuous age range.
|
What Netflix Can Teach Us About Treating Cancer
The Conversation
Elana Fertig
April 18, 2017
Johns Hopkins University professor Elana Fertig details her research to enhance precision medicine in cancer, using the techniques developed for finding patterns in Netflix movie ratings. The methods involve a matrix factorization algorithm that can identify movies with similar ratings among a small group of users. The program connects each user with a group of films to a different extent, based upon their individual preferences. Relationships among users are called "patterns" that are learned from the data, and may reveal common rankings unpredicted by movie genre alone. Fertig proposes gene dysregulation measurements are analogous to movie ratings, movie genres to biological function, and users to patients' tumors. The program can search across tumors to find patterns in gene dysregulation that induce the malignant biological function in each tumor. "This allows us to compute the probability of each gene being used in each biological function in a tumor," Fertig says.
|
|
