Association for Computing Machinery
Welcome to the December 17, 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).


Stanford to Host 100-Year Study on Artificial Intelligence
Stanford Report (12/16/14) Chris Cesare

Stanford University will spearhead a century-long initiative to study and predict the impact of artificial intelligence (AI) in all aspects of people's lives. The One Hundred Year Study on Artificial Intelligence (AI100) has started with the formation of a committee tasked with choosing an expert panel to begin a series of periodic studies of AI's implications for automation, national security, psychology, ethics, law, privacy, democracy, and other issues. "Given Stanford's pioneering role in AI and our interdisciplinary mindset, we feel obliged and qualified to host a conversation about how artificial intelligence will affect our children and our children's children," says Stanford president John Hennessy. Stanford professor Russ Altman and Microsoft Research's Eric Horvitz founded the committee, and Horvitz expects subsequent committees to identify the most important AI issues of the day and convene panels to research and report on them every few years. AI100 committee member and University of British Columbia professor Alan Mackworth envisions the study offering "a forum for us to consider critical issues in the design and use of AI systems, including their economic and social impact." Fellow committee member and University of California, Berkeley professor Deirdre K. Mulligan is focused on AI's social ramifications, noting the study "provides an intellectual and practical home for the long-term interdisciplinary research necessary to document, understand and shape AI to support human flourishing and democratic ideals."

Microsoft Gets Help in Telling U.S. to Back Off on Irish Search Warrant
IDG News Service (12/15/14) Grant Gross

Microsoft's fight against the U.S. Department of Justice (DOJ), which has requested the company turn over a suspect's documents stored on an Irish server, has been joined by allies who filed 10 amicus briefs with the appeals court. Among those signing the briefs supporting Microsoft were 34 computer science professors and more than a dozen trade groups. Microsoft's Brad Smith says the DOJ's efforts risk exposing U.S. records to foreign law enforcement agencies and cultivating global distrust in U.S. technology companies. He urges the U.S. Congress to pass rules that shield the privacy of cloud-stored emails, saying the federal government and tech companies can work out a viable solution to a legal issue that pits criminal probes against personal privacy. The DOJ warns upholding Microsoft's interpretation of the law would enable email providers to move content around the globe in an effort to evade law enforcement requests. Smith says the agency's reasoning is unsettling for the suggestion that personal emails and other documents stored in the cloud are business records partly owned by cloud providers, and not the exclusive property of the account holder. Center for Democracy and Technology CEO Nuala O'Connor argues digital documents form part of a person's "digital self," and she predicts within a decade "every thought and every communication and every transaction I do will be online."

As Robots Grow Smarter, American Workers Struggle to Keep Up
The New York Times (12/15/14) Claire Cain Miller

Economists and technologists have long said the march of technological progress and increased automation will create new jobs and opportunities that offset the jobs they make obsolete, but experts are increasingly less sure this axiom still holds true. The last several decades of technological progress has brought unprecedented gains in productivity, largely due to innovations in computer and communications technology, but that period has also seen decreased labor force participation and stagnating wages and many worry these trends are poised to accelerate with artificial intelligence and other technologies moving into everything from sales and vehicle piloting to personal training and psychiatry. Self-driving cars, for example, could completely eliminate truck and taxi drivers as a class of workers. Former Treasury Secretary Lawrence Summers recently said he no longer believes automation will always create new jobs. A recent University of Chicago survey of leading economists found 33 percent believed technology was the central reason median wages have stagnated. Massachusetts Institute of Technology economist Erik Brynjolfsson says ensuring that new technological innovations do not undermine the labor force is "the biggest challenge of our society for the next decade." However, Brynjolfsson and other experts believe society can meet the challenge in ways that will enable technology to be a positive force.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

CODE2040 Helps Tech Plan for a Non-White-Majority USA
USA Today (12/15/14) Marco della Cava

By the year 2040, the United States will have a non-white majority, a demographic trend that is spurring technology companies to increase the diversity of their workforces. "If you're a smart company, you'll want an ethnically diverse team empathetic about the needs of your diverse consumers," says CODE2040 co-founder Tristan Walker. CODE2040 wants to help Silicon Valley companies make the tech industry as a whole more diverse. The nonprofit organization, founded in 2012, provides college-age African-American and Latino students who have shown an interest in computer science with both an encouraging network of peers and a summer internship program aimed at placing them at tech companies whose narrowly focused recruiting programs often overlook them. "If you can inspire the best consuming demographic in the world to be the best producing demographic in the world, imagine the amount of market change you can have in the world," Walker says. About 90 percent of CODE2040 fellowships turn into full-time offers, a significant figure considering the organization focuses on reaching students from schools such as the University of California-Channel Islands and the University of Maryland-Baltimore County. "The issue often isn't about not being smart enough, it's about feeling like there's no one you've seen from your community achieve success in tech, so why would you think you could," says CODE 2040 co-founder Laura Weidman Powers.

New Findings Could Point the Way to 'Valleytronics'
MIT News (12/15/14) David L. Chandler

Massachusetts Institute of Technology (MIT) researchers have developed "valleytronics," a new approach to creating two-dimensional (2D) microchips that uses the characteristics of electrons, other than their electrical charge, which can be depicted as a pair of deep valleys on a graph. The researchers focused on tungsten disulfide, which belongs to a class of 2D crystals known as transition metal dichalcogenides. Valleytronics is based on the fact that in certain materials, when the energy of electrons is plotted relative to their momentum on a graph, the resulting curve features two deep valleys. If subjected to certain disruptions, these two valleys can have unequal depth, giving the electrons a preference to populate one of the two valleys; the two different states can be used to represent the zeroes and ones of data. "We discovered a way to directly control this valley by using light," says MIT's Edbert Jarvis Sie. Valleytronics could be used to make flexible electronics because of the 2D nature of the material and its mechanical strength, notes MIT professor Nuh Gedik. The researchers found a much greater energy shift can be achieved with a relatively conventional laser pulse with a special polarization, providing a new method of control for valleytronic devices. Gedik says it should now be possible to design devices in which all three properties of the electrons could be independently manipulated.

Innovators of Intelligence Look to Past
The New York Times (12/15/14) John Markoff

So much of today's research into artificial intelligence revolves around neural networks and machine learning, but Oren Etzioni and the researchers at the newly-established Allen Institute for Artificial Intelligence (AI2) are taking a different approach and turning to some of the field's earliest ideas. AI2 was established by Microsoft co-founder Paul Allen this year as a computer science companion to his Allen Institute for Brain Science, which plans to fully map the human brain. AI2 director Etzioni and Allen are focusing the institute's efforts on "GOFAI," or good old-fashioned artificial intelligence, the earliest efforts in the field of AI that in the 1950s and '60s hoped to model human intelligence using symbolic systems. These methods went out of style in the 1980s after an early wave of AI companies failed, drying up commercial interest in the field, which has since moved in different directions. However, Allen and Etzioni believe the new approaches have very significant limitations and hope to make great strides in areas such as computer vision by returning to the old ways. Unlike many other AI researchers, Etzioni does not believe "human-like" AI is only decades away, but he and Allen hope to make major incremental progress toward that goal at AI2.
View Full Article - May Require Paid Subscription | Return to Headlines | Share Facebook  LinkedIn  Twitter 

Students' Raspberry Pi Computers to Run on International Space Station (12/12/14) Kayleigh Bateman

The Raspberry Pi Foundation has partnered with the U.K. European Space Agency and U.K. space companies to offer an Astro Pi competition for students. The competition will challenge students in U.K. primary and secondary schools to use Raspberry Pi computers to devise and code apps for experiments in space. Astronaut Tim Peake will bring two winning apps along with him for his next six-month mission to the International Space Station (ISS). The Raspberry Pi computers will be connected to Astro Pi computers deployed onboard the ISS, collect data, and download it to Earth for the winning teams. The foundation and the U.K. European Space Education Resource Office will develop teaching resources for the competition. The foundation will provide Raspberry Pi computers and an Astro Pi board for coding, and will assist teams with the best ideas to ready their work for the space mission. "I'm really excited about this project, born out of the cooperation among U.K. industries and institutions," Peake says. "This competition offers a unique chance for young people to learn core computing skills that will be extremely useful in their future."

New Algorithm a Christmas Gift to 3D Printing
Simon Fraser University (12/11/14) Diane Luckow

Simon Fraser University (SFU) professor Richard Zhang says he and Ph.D. candidate Ruizhen Hu have developed the world's first algorithm for automatically decomposing a three-dimensional (3D) object into what are known as pyramidal parts. A pyramidal part has a flat base with the remainder of the shape forming upwards over the base with no overhangs. This shape is optimal for 3D printing because it incurs no material waste and saves printing time. In 3D printing, the printer deposits melted plastic layer by layer in a bottom-up fashion. If the shape has an overhang, extra material has to be printed beneath it as support. The extra plastic is waste material and must be removed, which can be time consuming and difficult. The SFU algorithm partitions the object into a small number of nearly pyramidal parts that can be 3D-printed with little or no material waste. The printed parts then can be glued together to form the finished object. "If the molded or cast parts are pyramidal, then removing the mold or cast after fabrication would not result in any breakage," Zhang says. The researchers presented their work at the recent ACM SIGGRAPH 2014 Conference on Computer Graphics and Interactive Techniques in Shenzhen, China.

Why Neural Networks Look Set to Thrash the Best Human Go Players for the First Time
Technology Review (12/15/14)

University of Edinburgh researchers have applied machine-learning techniques to the problem of finding the next move in a game of Go. Conventional Go algorithms play out the entire game after every move, and if the computer wins in the majority of these simulations, then that move is deemed a good one. However, that approach is time-consuming and computationally intensive. In addition, traditional Go algorithms fail to beat human Go experts, who can usually evaluate the state of a Go board with just a glance. Humans are better at Go because of the ability to spot strengths and weaknesses based on the shape the stones make rather than by looking several moves ahead, but recent advances in pattern-recognition algorithms could help computers do much better at playing Go. The Edinburgh researchers used a vast database of Go games to train a neural network to find the next move. The researchers used more than 160,000 games between experts to generate a database of 16.5 million positions as well as the next move. They then used almost 15 million of these position-move pairs to train an eight-layer convolutional neural network to recognize which move the expert players made next. The researchers found the trained network was able to predict the next move up to 44 percent of the time.

CompuGirls: Young Women Have Role to Play in Technology Field
National Science Foundation (12/12/14) Maria C. Zacharias

CompuGirls is a U.S. National Science Foundation-funded organization based at Arizona State University that provides girls aged eight to 12 from under-resourced schools in the Greater Phoenix area and Colorado with activities to develop technical skills and encourage computational thinking in a way that is culturally relevant. CompuGirls provides a one-week program during the fall or spring break, as well as an eighth-period class over a couple of semesters and a three-week summer class. The classes include digital storytelling, learning Scratch software, and exposure to coding through researching a topic and putting it together in a virtual world. "CompuGirls instills a sense of discipline, and we do see gains in their technical knowledge and comfort," says CompuGirls founder Kimberly Scott. CompuGirls also is working with the National Robotics Initiative to teach girls how to program robots and study the social aspects of human-to-robot interaction. "In this project we're communicating to the girls that we're going to teach you how to use these technology tools with robotics to advance your community, better understand yourself, and show that you do belong in technology and you have some valuable contributions you can make," says Marquette University researcher Andrew Williams.

Researchers Use Multiple Photos to Estimate Lighting Conditions of Outdoor Scenes
Phys.Org (12/10/14)

Structure-from-motion (SfM) algorithms could be used to automatically estimate illumination conditions in a collection of photographs. The algorithms are now widely used to create three-dimensional (3D) models based on multiple photographs. Disney Research's Iain Matthews and Laval University professor Jean-Francois Lalonde have used SfM techniques to create 3D models based on collections of photos focused on the same landmark. Matthews and Lalonde developed an inverse rendering approach and used it to recover the lighting conditions for each of the photos. The method reveals knowledge about lighting conditions that would make it easier for editors to insert objects into images and make them appear as if they are reflecting light or casting shadows naturally. "If one adds a virtual statue in front of a building in one of the photographs from the collection, the same statue can now be inserted in all the other photos with the correct illumination for each image," Lalonde says. To develop the method, the researchers used a novel database that included collections of photos of 22 different landmarks for which the actual conditions were recorded for each photo. Knowledge of the actual conditions provided a check on their ability to estimate those conditions. The researchers presented their work last week at the International Conference on 3D Vision in Tokyo.

Stacking Two-Dimensional Materials May Lower Cost of Semiconductor Devices
North Carolina State University (12/11/14) Matt Shipman

A team of researchers led by North Carolina State University (NCSU) found that stacking materials an atom thick enables the formation of semiconductor junctions that transfer charge efficiently, irrespective of whether the crystalline structures of the materials were matched or not. The researchers say the discovery has the potential to significantly reduce the manufacturing cost for devices that use semiconductors. Traditionally, the crystalline structures of semiconductor junctions of both materials must match, but the researchers found "the crystalline structure doesn't matter if you use atomically thin, 2-D (two-dimensional) materials," says NCSU professor Linyou Cao. "We used molybdenum sulfide and tungsten sulfide for this experiment, but this is a fundamental discovery that we think applies to any 2-D semiconductor material." As a result, any combination of semiconductor materials can be used. "This could make the manufacture of semiconductor devices an order of magnitude less expensive," Cao says. "It's as simple as stacking pieces of paper on top of each other--it doesn't even matter if the edges of the paper line up."

Women in Tech: Change the Conversation
EE Times (12/11/14) Jessica Lipsky

The G20 Summit in mid-November included a Women in Leadership conference in which leaders pledged to reduce the gender gap in workforce participation 25 percent by 2025. 1-Page CEO Joanna Weidenmiller wants to inspire more women to participate in the tech industry by highlighting their inherent skills and the field's ample opportunities. "This will be biggest job creation opportunity ever, and if you don't know tech in the next 10, 20 years you will be obsolete," Weidenmiller warns. She also says the government and school systems should incentivize businesses to teach computer and science skills, citing Step It Up America, a nationwide program led by UST Global that seeks to educate 5,000 minority women in science, technology, engineering, and math fields by 2020. Weidenmiller also believes women need more encouraging role models and parental support. "I think we should be talking more about how do you have a work life balance," she notes. "People like [Yahoo CEO] Marissa Mayer got to where they are because, regardless, they performed." Weidenmiller also believes it is essential to focus more on positives in order to change the conversation about women in technology. "No kid is going to want to be doing this is if all we talk about is the negatives of [being a woman in the technology industry]," she says.

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

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