Welcome to the May 24, 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).
|
Google's AlphaGo Defeats Chinese Go Master in Win for AI
The New York Times
Paul Mozur
May 23, 2017
Google DeepMind's AlphaGo program on Tuesday beat a Chinese world champion in the first of three Go games held this week, in what is being hailed as a victory for artificial intelligence (AI). The human player, Ke Jie, notes the program has improved rapidly after its 2016 defeat of a South Korean Go player. "AlphaGo is like a different player this year compared to last year," Ke says. DeepMind co-founder Demis Hassabis says AlphaGo uses methods in which it learns experientially from playing a large number of games. For the new contest, Hassabis notes the program adopted a strategy that enables it to learn more by playing games against itself. "Last year it was still quite humanlike when it played," Ke says. "But this year, it became like a god of Go." Researchers say similar AI methods could be used to perform many tasks, such as improving basic scientific research.
|
Citizens Give Up Data in Blockchain Project to Improve Cities
New Scientist
Matt Reynolds
May 22, 2017
The European Union (EU)-funded Decentralized Citizen Owned Data Ecosystem (DECODE) project will involve 2,000 residents in Barcelona, Spain, and Amsterdam, the Netherlands, using an app to share data about themselves to help companies or government groups create products or services to improve their city. Each citizen will be permitted to choose precisely how much of their data will be uploaded to the platform and how it should be used. Their data-sharing preferences will be stored on the blockchain to anonymize their identities. The U.K. innovation charity Nesta is collaborating with 13 partner EU agencies on DECODE, and Nesta's Tom Symons expects the initiative will filter public data into more socially beneficial government or corporate projects. The Oxford Internet Institute's Jonathan Bright believes crowdsourced citizen data could be useful for helping cities arrive at important planning decisions, with Symons noting insights on public spaces and transportation are only a few of DECODE's potential benefits.
|
Using Bitcoin to Prevent Identify Theft
MIT News
Larry Hardesty
May 23, 2017
Researchers at the Massachusetts Institute of Technology (MIT) have developed a system that uses bitcoin's security mechanisms to foil online identify thieves. Unlike earlier systems that use such protocols to prevent equivocation while requiring a full blockchain download for verification, MIT's Catena system requires a much smaller amount of data, which enables it to operate on a smartphone. MIT professor Srini Devadas and graduate student Alin Tomescu say Catenas adds the requirement that each bitcoin transaction logging a public assertion must entail an actual bitcoin transfer. Users may transfer the bitcoin to themselves, but that obviates the possibility of transferring the bitcoin to any other party in the same block of the blockchain, which means it also precludes equivocation within the block. "If you can eliminate the possibility of equivocation, it becomes easier to secure many algorithms," says Swiss Federal Institute of Technology in Lausanne professor Bryan Ford.
|
At Google, an Employee-Run Email List Tracks Harassment and Bias Complaints
Bloomberg
Ellen Huet; Mark Bergen
May 23, 2017
Google employees have created "Yes, at Google," an employee-run message board that tracks allegations of unwelcome behavior at work in an attempt to make the company more inclusive. The list is a grassroots effort to collect anonymous submissions at Google and parent company Alphabet Inc., and communicate them across the company. For employees that think they have witnessed sexual harassment, sexism, bigotry, or racism, "Yes, at Google" serves as an alternative to going through a traditional human resources department. Since starting in October 2016, more than 15,000 employees, representing 20 percent of Google's workforce, have subscribed to the service. The list is managed by a group of employees across different product areas. The list is being developed as Google and other technology companies face growing pressure from diversity advocates and the media to change the gender and racial makeup of the industry's workforce.
|
Researchers Find Computer Code That Volkswagen Used to Cheat Emissions Tests
UCSD News (CA)
Ioana Patringenaru
May 22, 2017
An international team led by researchers at the University of California, San Diego (UCSD) have uncovered the computer code that enabled Volkswagen automobiles' onboard computers to determine the vehicles were undergoing an emissions test. The team says the computer then activated the car's emission-curbing systems, reducing the amount of pollutants emitted. After the computer determined the test was over, these systems were deactivated. The researchers obtained copies of the code running on Volkswagen's onboard computers from the company's maintenance website and from online forums. "We found evidence of the fraud right there in public view," says UCSD's Kirill Levchenko. The team notes conventional emissions tests are standardized and public, making it easy for manufacturers to intentionally alter the behavior of their vehicles during the test cycle. The researchers found the Volkswagen code checks for several conditions associated with the emissions test, and if those conditions are met, the code instructs the onboard computer to activate emissions-curbing mechanisms.
|
A Rural Lens on a Research Agenda for Intelligent Infrastructure
CCC Blog
Helen Wright
May 22, 2017
A white paper published by the Computing Community Consortium and the Electrical and Computer Engineering Department Heads Association focuses on a research agenda for designing an intelligent infrastructure for rural communities. The design must incorporate robustness in the face of power outages, sustained function in desert and wetland areas, and enable access while being far-flung from significant data processing capability or information technology experts. The paper says to realize this infrastructure will require additional research to make wireless Internet access technologies ready for broad rural use, and the establishment of distributed and localized computing resources featuring unlimited Internet-based cloud computation and storage. Creating an intelligent rural infrastructure also will require testbeds that can reliably operate in rugged environments and across vast terrains, and creating employment opportunities to develop rural community workforces that can manage it. In addition, the paper proposes a rural-focused Intelligent Infrastructure Act deploying technological platforms and applications supported by a suitable policy.
|
Curiosity May Be Vital for Truly Smart AI
Technology Review
Will Knight
May 23, 2017
Researchers at the University of California, Berkeley have equipped an algorithm with an "intrinsic curiosity model" so it can function even in the absence of strong feedback. The model envisions the artificial intelligence software governing a virtual video game agent seeking to maximize its environmental awareness. The team employed this model, in conjunction with reinforcement learning, within two simple games, and artificial curiosity made the learning process more efficient in both instances. The agent learned to navigate its surroundings faster, with less time bumping into walls or getting killed, without any other reward. The researchers now want to test this technology on robots that use reinforcement learning to determine how to perform tasks such as grasping awkward objects. Berkeley's Pulkit Agrawal says innate curiosity should enable such machines to explore their environment faster and experiment with objects in close proximity.
|
Parasitic Robot System for Turtle's Waypoint Navigation
KAIST
May 22, 2017
Researchers at the Korea Advanced Institute of Science and Technology (KAIST) have developed a parasitic robot system, a hybrid animal-robot interaction that mimics the natural relationship between parasites and hosts. As part of the new system, the robot is attached to its host animal in a similar way to an actual parasite, and it interacts with the host through specialized devices and algorithms. The KAIST researchers chose a turtle as their first host animal and designed a parasitic robot that performs "stimulus-response training" on the reptile. The robot, which is attached to the turtle, can induce the turtle's object-tracking behavior through repeated training sessions. The robot then guides the turtle using light-emitting diodes and feeds it snacks as a reward for going in the right direction via a programmed algorithm. The researchers believe this hybrid animal-robot interaction system could lead to an alternative solution to the limitations of conventional robotic systems in some fields.
|
Network Traffic Provides Early Indication of Malware Infection
Georgia Tech News Center
John Toon
May 22, 2017
Georgia Institute of Technology (Georgia Tech) researchers have found that by analyzing network traffic going to suspicious domains, security administrators could detect malware infections weeks or even months before they can capture actual samples of the invading malware. The researchers say this breakthrough highlights the need for new malware-independent detection strategies that enable network defenders to identify security breaches more quickly. The new method utilizes the fact that malware invaders need to communicate with their command-and-control computers, creating network traffic that can be detected and analyzed. The researchers analyzed more than 5 billion network events from nearly five years of network traffic carried by a major U.S. Internet service provider. In addition, they studied domain name server requests made by nearly 27 million malware samples. The researchers encourage network administrators to minimize the unknowns in their networks and classify their appropriate communications as much as possible.
|
Humanitarian Efforts Could Be Aided by AI
Imperial College London
Colin Smith
May 18, 2017
Researchers at Imperial College London in the U.K. have developed a machine-learning algorithm to analyze mobile phone data from 10,000 users in a developing and developed country. They found the algorithm could predict a person's gender from the way they employ their phones with 74.3-percent to 88.4-percent accuracy in a developed country and with 74.5-percent to 79.7-percent accuracy in a developing country. The researchers extended this gender-prediction capability to 500,000 users, in addition to the ability to predict the proportion of female users in a group. The researchers say the algorithm could improve aid and relief efforts during humanitarian crises by making such efforts more targeted. For example, the algorithm could analyze prepaid phone data to pinpoint groups of women or elderly, and it also could help relief agencies and governments determine where aid could be more quickly dispatched during a crisis.
|
Next-Gen Computing: Memristor Chips That See Patterns Over Pixels
University of Michigan News
Nicole Casal Moore
May 22, 2017
University of Michigan (U-M) researchers say they have developed a memristor computer circuit prototype that has the potential to process complex data, such as images and videos, faster and using much less power than today's most advanced systems. The prototype uses pattern recognition to shortcut the energy-intensive process conventional systems use to dissect images. The system uses an algorithm that relies on "sparse coding" to instruct the 32-by-32 array of memristors to efficiently analyze and recreate several photos. Memristors can simultaneously perform many operations without having to move data around, which means they could enable new platforms that process a vast number of signals in parallel and are capable of advanced machine learning. The U-M system is designed to detect patterns very efficiently, and to use as few features as possible to describe the original input. The researchers also note it could be scaled up to process and analyze video in real time.
|
Family TV Viewing and SMS Texting Could Help Cut Internet Energy Use
Lancaster University
May 18, 2017
Researchers at Lancaster University and the University of Cambridge in the U.K. recently conducted a detailed study on Android device users, comparing their observations with a large dataset of nearly 400 U.K. and Ireland mobile devices. The researchers used the study to identify four categories of data-hungry service, including watching video, social networking, communications, and listening. The researchers note collectively, the four categories make up about 50 percent of mobile data demand. They recommend developers create features for devices and apps that encourage users to gather together to enjoy streamed media, reducing the overall number of data streams and downloads. The study also found energy consumption via instant messaging apps' data demand is about 10 times higher than that of short messaging services. The researchers presented their work this month at the ACM Conference on Human Factors in Computing Systems (CHI 2017) in Denver, CO.
|
12 Views on the Future of Electronics
EE Times
Rick Merritt
May 22, 2017
Experts at last week's Imec Technology Forum in Belgium offered predictions on how the future of electronics will evolve, including Samsung's Kinam Kim's forecast of neural-network processors mining insights from Internet of Things (IoT) data. Microsoft Research's Peter Lee anticipates the booming application of refined deep-learning models, while a more cost- and power-efficient neuromorphic machine-learning architecture is envisioned by Imec's Praveen Raghavan. Neural implants that shed more light on the brain and beneficial brain stimulation are a focus for Medtronic's Tim Denison, and the convergence of the IoT and healthcare in the form of wearable medical sensors is of central interest to scientist Mario Konijnenburg. Meanwhile, Imec's life science division is miniaturizing medical test equipment using photonic chips, and Sony's Kazumasa Nomoto expects virtual reality and augmented reality to transform human interaction with large-screen and mobile devices of the future. Other predictions include the emergence of common underlying systems for wired and wireless networks.
|
|
