ACM TechNews
Association for Computing Machinery
Welcome to the October 1, 2014 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).


Electronic Brain by 2023
EE Times (09/30/14) R. Colin Johnson

Researchers on the European Union's decade-long Human Brain Project (HBP) recently disclosed how far they have come toward the goal of creating an artificial brain by 2023 at the annual HBP Summit in Germany. The project's goal is the initial simulation of the human brain on supercomputers, and then the precise replication of its functions via a hardware emulator. The long-term objective is the construction of artificial brains that can outperform traditional von Neuman supercomputers at far less cost. The report at the summit detailed the hiring of all required personnel, lab engagements, and the installation of the information and communications technology (ICT) infrastructure to enable HBP researchers and their partners to collaborate and exchange data. The report also says projects to reproduce brain functions at different biological scales, analyze clinical brain disease data, and develop brain-inspired computers have started. The agenda for the project's first 30 months calls for amassing all known strategic brain functioning data, developing theoretical frameworks aligned with that data, and creating infrastructure for developing six ICT platforms for HBP's operational phase. Highlighted milestones achieved in the first year include a neocortical brain-modeling method repurposed for the cerebellum, the retrofit of a high-performance computer for interactive supercomputing, and the accumulation of high-resolution three-dimensional cellular-level data for the Human Brain Atlas.

Vast Grassroots Mobile Network Grows at Center of Hong Kong Protests
Technology Review (09/30/14) David Talbot

At the heart of the demonstrations in Hong Kong is a huge mobile messaging network whose operation is separate from cellular networks, enabling protestors to communicate even if the cellular network is too crowded or shut down. Many demonstrators are using the FireChat messaging app so phones can communicate directly with each other using Bluetooth or Wi-Fi, rather than linking to a cellular or Wi-Fi network. The app's proliferation was sparked when a Hong Kong teenager urged people to download it in various social media posts, says Open Garden's Christophe Daligault. He notes FireChat was the most widely used app download on both Google Play and in Apple's iTunes store in Hong Kong between Sept. 27 and Sept. 29, when more than 200,000 people in the city downloaded it. Daligault reports FireChat already is being used for organizing demonstrations. "You could see people creating 'rooms' with a specific location...and saying 'How many masks do we need?' and 'Who is going to bring the water?' and 'Today it is tear gas, tomorrow it could be water cannon,’" he says. However, FireChat lacks the ability to encrypt the data it sends, and Open Garden is working to correct this.

Tim Berners-Lee Wants Internet Magna Carta to Guarantee Netizens' Privacy
Network World (09/28/14)

Sir Tim Berners-Lee is using the 25th anniversary of his invention of the World Wide Web to push for an Internet bill of rights, which he characterizes as a "Magna Carta for the Web." Berners-Lee first proposed the idea in a TED Talk earlier this year, but he renewed his call at the recent "Web We Want" event in London. Berners-Lee says he originally conceived of an Internet Magna Carta as a way to protect the open Internet and its users' rights from governments such as China that want to circumscribe their citizens' access to the Internet. However, he says he now views an Internet Magna Carta as being just as important to help keep large companies in check. "You shouldn't trust your government, but you shouldn't trust your companies either," Berners-Lee cautions. "We have to have checks and balances for them all." Berners-Lee emphasizes such a document would not guarantee unlimited freedom, but would set widely agreed-upon standards for openness and privacy. However, it is not clear how and by whom such a document would be drafted and through what entity it would gain legal force.

Field-Coupled Magnets Could Replace Transistors in Some Computer Chips
Technical University of Munich (Germany) (09/30/14)

Researchers at the Technical University of Munich (TUM) and the University of Notre Dame have demonstrated through experimentation that three-dimensional (3D) assemblies of nanometer-scale magnets could replace transistors as the basis for future computer chips. The researchers implemented a majority logic gate in a 3D stack of nanomagnets, giving it the function of a programmable switch in a digital circuit. Gates built from field-coupled nanomagnets operate in an analogous manner, with the polarity reversal serving as a switch between Boolean logic states. The researchers say three input magnets determine the device's state, while one output magnet reads it out. The possibility of packing more gates onto a chip is particularly valuable, as nanomagnetic logic can enable extremely dense packing, the individual nanomagnets are similar in size to transistors, and they make contacts and wiring unnecessary because they function purely with coupling fields. Moreover, fewer magnets than transistors can be employed to build complementary metal-oxide semiconductors and nanomagnetic devices with identical functions. In addition, magnetic logic becomes competitive thanks to the potential for exceeding two-dimensional design limitations with stacks of 3D devices. "The 3D majority gate demonstrates that magnetic computing can be exploited in all three dimensions, in order to realize monolithic, sequentially stacked magnetic circuits promising better scalability and improved packing density," says TUM's Irina Eichwald.

Research Mimics Brain Cells to Boost Memory Power
RMIT News (10/01/14)

Researchers at Australia's RMIT University say they have created a nano-structure that could serve as a platform for the development of new computer memory technology. The nanometer-thin stacked structure is created using a thin film of functional oxide material that creates the effect of a memristor, which has been touted as a potential replacement for the transistor technology that underlies current Flash and other memory technologies. Project leader Sharath Sriram says Flash memory is rapidly approaching its fundamental limits, creating a need for materials such as memristors to serve as the foundation for the next generation of memory. Memristors could be used to fashion non-volatile solid-state memory that could be used to mimic the structure of the human brain. Sriram says the new structure developed by RMIT could be used in numerous electronic applications, from ultrafast memory devices as small as a few nanometers to larger computer logic architectures that could be used to mimic neural networks, "bringing us one step closer to the bionic brain." RMIT's Hussein Nili adds the structure also has potential applications as memory elements in sensors and actuators.

Smarter Algorithms Will Power Our Future Digital Lives
Computerworld (09/26/14) Sharon Gaudin

Researchers around the world are developing deep-learning algorithms that are heavily used in machine learning and artificial intelligence and could become the foundation on which important technological advances are built. "Deep learning is a critical part of the future of the digital world even though most people don't know anything about it," says analyst Patrick Moorhead. Stanford University professor Andrew Ng is building the next generation of deep-learning algorithms. Such algorithms are special because the more data they are given, the better they function, similar to the human brain. "As we build larger and larger simulations of the brain, these models are relatively efficient at absorbing huge amounts of data," Ng says. The researchers also are using graphics-processing units (GPUs), which are much better for building large neural networks because they are better at handling those kinds of calculations. "We're building a new deep-learning platform with GPU hardware to help us scale better," Ng says. He thinks improving these algorithms could lead to advances in voice recognition and visual search. "Within five years, 50 percent of our queries will be through speech and images, so this is a technology we are investing heavily in," Ng says.

Simulation Technology Designed for Hollywood Can Be Used as Predictive Tool for Understanding Fundamental Engineering
Columbia News (09/29/14)

Researchers from Columbia University and the Massachusetts Institute of Technology (MIT) have found a new application for software developed to help Hollywood create better virtual hair: studying the fundamental engineering properties of filamentary structures such as oceanic telecommunication cables. Columbia professor Eitan Grinspun and MIT professor Pedro Reis led a team using Discrete Elastic Rods, simulation software developed by Grinspun for use in animating hair and fur for film and design applications, which has been used to create digital effects for films such as "The Hobbit" and the "Planet of the Apes" series. Grinspun and Reis found the software also could be used to model the behavior of long, filamentary structures such as cables, the mechanical properties of which often can be vexing. Using Discrete Elastic Rods, the researchers modeled the deployment of oceanic fiber-optic cables and found factors such as the diameter of a spool and the speed at which the cable is deployed can have dramatic effects on how the cable comes to lie on the ground. They say their findings could help improve the deployment of such cables, optimizing them for cost and resilience.

Robot Jurisprudence: How to Judge a 'Bot; Why It's Covered
The Economist (09/25/14)

The RoboLaw consortium is a group of experts in the fields of law, engineering, philosophy, regulation, and human enhancement brought together in 2012 by the European Commission to contemplate legal frameworks that might be necessary to help manage the development of robotic technology. RoboLaw recently released a final report and presented its findings to the European Parliament. The report seeks to strike a balance between the need to create laws that protect human rights and interests and the freedom to allow technology to develop and mature. One example is in the area of liability, concerning who is accountable when a robot is involved in damage to property or harm to a human. The RoboLaw report suggests tough liability standards could quickly shut down innovation and instead recommends liability waivers or the creation of public-private compensation funds. Another important but distant issue is that of robot rights. The report suggests robots should not be granted equal legal standing to human beings, although they could be granted a legal status similar to that of a corporation in some situations. The study notes many of these issues have international implications and should be handled on an international rather than a national basis.

Can Cartoons Be Used to Teach Machines to Understand the Visual World?
Virginia Tech News (09/26/14) Lynn Nystrom

Google has honored professor Devi Parikh, a computer-vision and pattern-recognition expert at Virginia Polytechnic Institute and State University's Bradley Department of Electrical and Computer Engineering, with one of its Faculty Research Awards. The company sought to recognize Parikh for her earlier creative work on how to learn from visual abstractions. In her research, she proposes the use of visual abstractions or cartoons to teach machines. Concepts that are difficult to describe textually may be easier to illustrate, according to Parikh. She is attempting to teach computers to understand the visual world as humans do by having thousands of online crowd workers manipulate clip-art images to mimic photographs. The one-year award will provide Parikh with $92,000 in unrestricted funds. She also will work directly with Google researchers and engineers as they explore how to best learn from visual information. "If we can figure out a way for people to effectively teach machines, machines will be much more intelligent than they are today," Parikh says.

First Quantum Logic Operation for an Integrated Photonic Chip
Technology Review (09/26/14)

University of Oxford researchers say they have developed the first self-contained photonic chip capable of teleportation, which is one of the fundamental logic operations necessary for quantum computation. Quantum teleportation aims to transfer the quantum state of an input quantum bit (qubit) to a target qubit. The process begins by creating a pair of entangled photons that share the same quantum existence so a measurement on one immediately influences the other. Although the process is typically completed using low intensity laser beams, mirrors, and optical fibers, the new photonic device shrinks all of these components onto a single silicon chip. The researchers say they were able to dramatically improve the quality of the single photon sources while characterizing the losses from other optical components such as beam splitters and waveguides for the first time. However, the chip only performs with about 89-percent fidelity. The photon source is one source of error. "Whilst the success of this experiment relies on our development of high-quality single photon sources with exceptional heralded state purity and heralding efficiency, the absence of a true on-demand single photon source continues to limit the achievable fidelity," the researchers say.

Blackout? Robots to the Rescue
Michigan Tech News (09/26/14) Marcia Goodrich

Michigan Technical University (MTU) researchers have developed a tabletop model of a robot team that can be deployed in hard-to-reach places during power failures to bring electricity to areas that need it the most. The team has programmed the robots to restore power in small electrical networks, linking up power cords and batteries to light a lamp or set a flag to waving with a small electrical motor. The robots operate independently, and are capable of choosing the shortest path, avoiding obstacles, and recharging one another. "Our robots can carry batteries, or possibly a photovoltaic system or a generator," says MTU professor Nina Mahmoudian. Her team also is working to incorporate a power convertor, since different systems and countries have specific electrical requirements. Mahmoudian notes the autonomous power distribution system also could enable the military to set up power systems before troops reach a site. Moreover, the team plans to develop a full-size, working model of the robot network. "This will let us develop path-planning algorithms that will work in the real world," Mahmoudian says.

ORNL Researcher Is Working to Predict Electric Power Blackouts Before They Happen
Oak Ridge National Laboratory (09/25/14) Katie Jones

Oak Ridge National Laboratory (ORNL) researchers are developing tools to improve electric grid protection operation analysis and prediction under different contingencies. The new grid protection tools are for familiar events, and they can protect against influxes of renewable energy sources and smart grid components such as communication devices and sensors. "We eventually want to feed real measurements from the grid into the model, which for utilities means a faster response time because they can plan for individual scenarios," says ORNL's Travis Smith. The researchers are developing a protection program that analyzes changes in parameters such as current, voltage, frequency, and impedance to coordinate protective relays into a low-impact protection strategy as an event happens. A key challenge involves the size of computational problems, because for models that simulate the grid five to 10 years from now, the addition of wind energy, small-system energy storage devices, and communicating sensors makes computation volume even tougher. "Once you have a faster, smarter grid, you need protection for a faster, smarter grid," Smith notes. "I'm working with computer scientists at ORNL to see if we can push the model to run faster than real-time speeds so we can predict what will happen before it happens."

Robotic Explorers
The Institution of Engineering and Technology (09/25/14)

Yonsei University researchers have developed a system to enable air- and ground-based robot vehicles to cooperate without global-positioning systems signals or sensor devices. The researchers say the system could be used by unmanned vehicle teams to explore dangerous environments. The system uses stereo vision depth-sensing to provide the obstacle map, and other image-processing methods to identify and track all the agents in the system. The researchers note the performance of traditional stereo vision-based depth-sensing degrades if the distance from the observed objects becomes much larger than the separation of the cameras. However, they say their technology addresses this limitation by using single cameras on two unmanned aerial vehicles (UAVs), providing much greater separation and control of that separation. "UAVs equipped with multi-spectral imaging sensors can collect useful information on crops over a large farming area," says Yonsei professor Jiwon Seo. "Based on this information, unmanned tractors could operate more efficiently. As a longer-term development, the system can be applied to emergency patient transport and military applications." The researchers say the technology could be used over many different single-agent and multi-agent systems. "We think that it is a quite natural path for robotics research to move in this direction over the next decade," Seo says.

Abstract News © Copyright 2014 INFORMATION, INC.
Powered by Information, Inc.

To submit feedback about ACM TechNews, contact:
Current ACM Members: Unsubscribe/Change your email subscription by logging in at myACM.
Non-Members: Unsubscribe