Welcome to the December 15, 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."
|
|
|
Brittle Stars Inspire New-Generation Robots Able to Adapt to Physical Damage
Tohoku University
Akio Ishiguro
December 14, 2017
Researchers at Tohoku and Hokkaido universities in Japan have developed a robot that can immediately adapt to unexpected physical damage, and this is a significant breakthrough as robots are increasingly expected to function in tough environments under hazardous conditions. Conventional robots usually require a significant amount of time to adapt when they experience physical damage. Therefore, the researchers focused their efforts on a brittle star, a primitive echinoderm with five flexible arms that can immediately adapt to an arbitrary loss of their arms and still move by coordinating the remaining arms. The researchers examined this ability and proposed a simple decentralized control mechanism in which each robotic arm kicks the ground only when it obtains an assistive reaction force. The team implemented this mechanism in a brittle star-like robot to demonstrate that it can adapt to unexpected physical damage within a few seconds, just like a biological brittle star.
|
People Don't Trust Driverless Cars. Researchers Are Trying to Change That
Science
Matthew Hutson
December 14, 2017
Industrial and academic researchers are attempting to overcome consumers' distrust of autonomous vehicles (AVs), with some studying how persons outside the vehicles react while others concentrate on passengers' interaction. Intel's Jack Weast says experiments demonstrated that interaction with AVs programmed to respond vocally is helping ease riders' concerns, while other projects also determined passengers find talking cars comforting. There are indications that humanization of AVs could influence passengers' response to an accident, with a team at Northwestern University finding anthropomorphism helped facilitate less startled reactions among riders when the AV was struck by another car. Scientists also are exploring other ways of having AVs communicate with riders, including video screens, sounds, and vibrations that might signal impending turns and stops. The University of California, Berkeley's Anca Dragan is investigating giving passengers some control over AV behavior, noting the idea is to have "the car adapt to the person, rather than having the person adapt to the car."
*May Require Paid Subscription
|
University of Adelaide Pours $1M Into Supercomputer Upgrade
ARNnet (Australia)
Julia Talevski
December 14, 2017
The University of Adelaide in Australia has designated an additional $1 million for increasing the capacity of its Phoenix supercomputer by doubling its storage to 700 terabytes and increasing processing speed to 450 teraflops from 300 teraflops. Phoenix is based on Lenovo's System x NeXtScale technology, and when it was launched, the supercomputer featured 15,360 gigabytes of memory and about 4,000 cores. "With the rising complexity of research, high-performance computing power is an essential tool for our researchers and collaborators to keep producing world-class research and innovation," says University of Adelaide professor Mike Brooks. He also notes the upgrade comes about five months after it was disclosed that the Commonwealth Scientific and Industrial Research Organization (CSIRO), Australia's national science agency, was building a new large-scale scientific computing system. CSIRO is working with private partners to provide commercial off-the-shelf software, hardware, support and maintenance across networks, unified communications, information technology security, and data center equipment.
|
Researchers Trained an AI Wormbot to 'Balance' a Pole on Its Tail
Motherboard
Michael Byrne
December 13, 2017
Austrian researchers led by Ramin Hasani have developed an algorithm that exploits an established C. elegans neural circuit for the purpose of teaching a simulated worm, or wormbot, to balance a pole on its tail, known as the inverted pendulum problem. The team used reinforcement learning to accomplish this task, after the worm's neural circuitry was fully mapped out and described. Overall, the researchers discovered this training regimen works about as well as similar machine-learning approaches to the inverted pendulum problem, but with the key wrinkle that their algorithm is built from hacked wormbot brains. The technique also ran into a form of difficulty in which the wormbot tended to "drift" in one direction as it moved around in its attempt to balance the pendulum, and it would consequently run out of free space in some scenarios.
|
Microsoft Launches Quantum Computing Toolkit
TOP500.org
Michael Feldman
December 13, 2017
Microsoft this week launched a free preview of its Quantum Development Kit, which is built around its new Q# programming language for quantum computing. Microsoft says the toolkit provides Q# programmers the basic concepts of quantum applications, some of which reside in the toolkit's documentation. All the Q# tools are integrated within Microsoft's Visual Studio coding environment, and Microsoft also is offering an Azure-powered quantum computing simulator capable of supporting up to 40 virtual quantum bits (qubits). Microsoft says software written in Q# for the simulator should function as is when it executes on an actual "topological quantum computer," which the company is developing. Microsoft also notes the robustness of the topological qubits will let the hardware scale with greater ease than other approaches, although the difficulty of building these qubits is why Microsoft has yet to develop a working prototype system.
|
IEEE Launches Ethical Design Guide for AI Developers
Computerworld Australia
George Nott
December 14, 2017
IEEE this week has published for comment its second Ethically Aligned Design document in an attempt to ensure increasingly pervasive autonomous and intelligent systems "remain human-centric." "These systems have to behave in a way that is beneficial to people beyond reaching functional goals and addressing technical problems," the document says. "This will allow for an elevated level of trust between people and technology that is needed for its fruitful, pervasive use in our daily lives." Among the topics the guide encompasses are autonomous system transparency, data privacy, algorithmic bias, mixed reality, and the overt or concealed manipulations by intelligent systems designed to shape a user's behavior or emotions. "Defining what exactly 'right' and 'good' are in a digital future is a question of great complexity that places us at the intersection of technology and ethics," says University of Sydney in Australia professor Rafael Calvo.
|
Is Peer Review Fair When It Is Not Blinded?
ACSH News
Julianna LeMieux
December 13, 2017
In co-chairing the 10th ACM International Conference on Web Search and Data Mining (WSDM 2017) in the U.K., a team of Google researchers studied whether blinded peer review of scientific literature upholds fairness. In a test of single- versus double-blind submissions to the conference, the team analyzed various factors to determine the differences between either form of peer review. They outlined three key differences, noting single-blind reviewers were more selective in their choices of the papers to select, and they selected more papers from leading universities and companies, compared to double-blind reviewers. Furthermore, single-blind reviewers were more likely than their double-blind counterparts to submit a positive review for papers with a renowned author and for papers from a top university or company. The Google team concluded when reviewers know who did the work and where, it shapes their judgment and they are more likely to choose papers from top institutions and well-known authors versus blinded reviewers.
|
How to Find the Truth on Twitter
Pursuit (Australia)
Marie Truelove; Stephan Winter
December 1, 2017
Researchers at the University of Melbourne in Australia have developed a framework to evaluate whether an event reported on Twitter is a likely witness account by assessing the evidence of whether the poster is on the ground at the event. The framework analyzes the textual and visual content of the tweet, seeking inferences that the poster was at the event they are tweeting about, and then testing that by looking for evidence of their presence. In the tweet's text, statements such as observations of the event, attached images, and the existence of geotags in the metadata build the user's credibility. The framework surmounts various challenges by exploring different evidence sources within tweets, with procedures applied to exclude postings that cannot support inferences that the tweeter is at the event. Supervised machine-learning techniques then apply classification models to pull up evidence from the remaining tweets that support inferences that the tweeter is at the event.
|
Robotics Researchers Track Autonomous Underground Mining Vehicles
Queensland University of Technology
Debra Nowland
December 12, 2017
Researchers at the Queensland University of Technology (QUT) in Australia have developed new technology to enhance underground mining vehicles so they can navigate autonomously through dust, camera blur, and bad lighting. The new system uses mathematics, biologically inspired algorithms, and a vehicle-mounted camera to track the location of the vehicle in underground tunnels. "We have developed a positioning system that uses cameras rather than lasers, based on more than a decade of research in biologically-inspired navigation technology," says QUT professor Michael Milford. The researchers had to develop the new system because the tough terrain means traditional global-positioning systems cannot be used and wireless sensor networks are less reliable due to interference from the rock mass and lack of access points. "We developed a system which could intelligently evaluate the usefulness of the images coming in from the camera--and disregard ones that were blurry, dusty, or that were washed out from incoming vehicle lights," Milford notes.
|
How Microsoft's Brendan Burns Is Trying to Make Cloud-Native App Deployment as Easy as Coding
GeekWire
Tom Krazit
December 12, 2017
Metaparticle is a new open source project created by Microsoft distinguished engineer Brendan Burns to provide software developers with a toolkit for writing applications for cloud-native deployments. Burns says Metaparticle's objective is to develop software libraries that can be used with popular programming languages so developers can write code in the lexicon of their language that containerizes their app and deploys it to Kubernetes. "As we move toward a world where every program is a cloud-native program, we should expect our programming languages to adapt to become cloud-native themselves," Burns notes. He says Metaparticle is designed to significantly automate the spinning up of containers and implementation to Kubernetes in an app's native code. "Ultimately, what I would say is that the motivation is to broaden the set of programmers who can build cloud-native applications," Burns says. He wants to build a Metaparticle-oriented programmer community before having the project extended to actual production systems.
|
Artificial Intelligence Is Killing the Uncanny Valley and Our Grasp on Reality
Wired
Sandra Upson
December 13, 2017
As deep-learning artificial intelligence (AI) technologies advance, the "Uncanny Valley"--the unease people feel toward realistic computer-generated humans--is beginning to abate. Indeed, innovations are moving so quickly that scientists are concerned about computerized imagery becoming so lifelike as to be indistinguishable from reality. Some researchers see an upside to this trend, with Cornell University's Kavita Bala noting, "What I like about these technologies is they are democratizing design and style." On the other hand, the University of Chicago's Yuanshun Yao sees darker implications to machine learning's progress, including purposeful and malevolent deception and disinformation. "Maybe in five or 10 years, we will be surrounded by AI-generated [news]," Yao warns. Also worried is Dartmouth College professor Hany Farid, who envisions the viral spread of fake content, in combination with the slow vetting process, supporting scenarios such as the intentional creation and proliferation of counterfeit videos designed to induce mass panic.
|
Social Media Trends Can Predict Tipping Points in Vaccine Scares
University of Waterloo News
Matthew Grant
December 11, 2017
Researchers at the University of Waterloo in Canada found analyzing trends on Twitter and Google helped predict vaccine scares. The researchers used artificial intelligence (AI) and a mathematical model to analyze Google searchers and geocoded tweets, producing data that described the public perceptions of the value of getting vaccinated and indicating when a population was approaching a tipping point. "By monitoring people's attitudes towards vaccinations on social media, public health organizations may have the opportunity to direct their resources to areas most likely to experience a population-wide vaccine scare, and prevent it before it starts," says University of Waterloo professor Chris Bauch. The researchers focused on tweets mentioning the measles-mumps-rubella vaccine and classified their sentiment using AI software. The researchers also collected data on measles-related Google searches, and the mathematical model predicted how the 2014-2015 Disneyland outbreak helped push California back from the tipping point by making residents more afraid of the disease than the vaccine.
|
UW-Madison Cloud Computing Research Moves Into New Phase
UW-Madison News
Jennifer Smith
December 8, 2017
The University of Wisconsin-Madison (UW-Madison), the University of Utah, and Clemson University are working on CloudLab, an effort to further develop cloud computing infrastructure and enable high-level research by scientists across the U.S. The vast computing cluster maintained by the three campuses serves as a testbed to run experiments on, and more than 56,000 experiments are currently running on the testbed; Wisconsin's servers often are operating at 90-percent capacity. CloudLab's newest phase will provide state-of-the-art support for machine learning, employing graphics-processing units (GPUs) that are well suited to cutting-edge machine-learning research. "A lot of machine learning advances in recent years in image classification, machine translation, and robotics have happened because of training at large scales on clusters of GPU-equipped machines," says UW-Madison professor Aditya Akella. He notes the new CloudLab phase also includes state-of-the-art "whitebox" network switches, which enable researchers to reprogram how the network supports applications.
|
|
