Association for Computing Machinery
Welcome to the February 27, 2015 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).


What It Means: The FCC's Net Neutrality Vote
Computerworld (02/26/15) Matt Hamblen

The U.S. Federal Communications Commission's (FCC) vote this week to create a series of new regulations protecting the open Internet is likely to change the nature of the Web. The new rules, which include banning Internet service providers (ISPs) from blocking or throttling Internet traffic and taking payments to prioritize content and services on their networks, rely on Title II of the Communications Act of 1934 and Section 706 of the Telecommunications Act of 1996. FCC officials say the commission will use the authorities granted it under these laws to help fend off the inevitable legal challenges its new rules will face. Although the rules are meant in many ways to preserve the current status quo, they are likely to change the industry in several ways. Former FCC commissioner Michael Copps says the FCC could use the new rules to spur expansion of broadband Internet access by preempting state laws restricting the expansion of municipal broadband and slowing down consolidation among ISPs. Many also are predicting ISPs will find ways to work around the FCC restrictions. Although the rules prohibit the creation of so-called "fast lanes" on public networks, ISPs could seek to circumvent these rules by creating private networks dedicated to carrying specific kinds of traffic.

Artificial Intelligence Bests Humans at Classic Arcade Games
Science (02/25/15) John Bohannon

In an article published Feb. 25 in the journal Nature, a team of artificial intelligence researchers describe the method they used to train a computer to play, and in some cases beat, human players in several classic arcade games. The DeepMind team, acquired by Google about a year ago, describe what they call the Deep-Q-Network (DQN), a combination of a deep neural network and a technique known as Q-learning. The deep neural network is similar to networks that have been used to mimic animal vision and enables the system to "see" a video game the way a human would, as pixels on a screen. Meanwhile, Q-learning is a mathematical version of a psychological concept called reinforcement learning thought to be key to the way humans and animals learn. The researchers say DQN plays games and learns how to get better at them in a way that mimics human players. DeepMind applied DQN to 49 classic Atari 2600 games from the 1980s and found the network was able to score better than the best human players on about half of them. Founder Demis Hassabis says DeepMind researchers are now working on "knowledge transfer," teaching DQN how to apply the lessons it learns from playing one game to another.

BCS Network to Figure Out Why Initiatives to Attract Females to IT Are Not Working (02/25/15) Kayleigh Bateman

Liz Bacon, the outgoing president of BCS, the Chartered Institute for IT, says a key piece of her legacy at BCS will be a network of women in the science, technology, engineering, and math fields she established at the Institute to encourage more women to pursue careers in those fields. Bacon says despite numerous efforts to boost women's participation in information technology (IT), their representation in the field continues to slip. "The challenge is that we have lots of computing initiatives but the number of women is declining. What we're doing is obviously not working, so we need to look at why," Bacon says. "The aim of the network is to share best practices, gather data, find roles models and find out how we can help." The network's activities will include carrying out more proven interactions such as mentoring and work shadowing, but also investigating potential reasons women do not pursue IT careers. One possible factor identified by Bacon is the difficulty women who leave the field for a time after having a child have returning. She says their skills often are no longer current, making them less attractive to employers.

Changing the World One Hackathon at a Time
USA Today (02/26/15) Jessica Guynn

Sixty-six teens are taking part in the My Brother's Keeper Hackathon, a group coding competition spearheaded by Qeyno Labs CEO Kalimah Priforce. Priforce notes this hackathon is different because it caters to African-American teens, rather than mostly white and Asian men. He says hackathons can collapse the walls that have isolated high-potential kids with too few opportunities. "Why not put Dr. King, Amelia Earhart, and Steve Jobs in one room and see what is it they can do," Priforce says. "But hackathons right now are comprised of Steve Jobs, Steve Jobs, Mark Zuckerberg, Steve Jobs, Mark Zuckerberg, and maybe one Asian guy. That's a problem." Priforce is planning 10 more hackathons this year, including events in St. Louis and New York City. Qeyno Labs, which specializes in organizing hackathons, is based in Oakland, CA, and Priforce believes the city can become the Silicon Valley of the 21st century by harnessing the untapped brain power and hustle of the its diverse youth. The My Brother's Keeper Hackathon helped develop several promising apps, including 14-year-old George Hofstetter's mobile app to help African-American teens feel less nervous around police officers. Hofstetter's app won the best impact and "outstanding trailblazer" awards at the hackathon.

CCC Aging in Place Workshop Report
CCC Blog (02/24/15) Helen Wright

The Computing Community Consortium has released a report on its Trans-National Institutes of Health/Interagency Workshop on the Use and Development of Assistive Technology for the Aging Population and People With Chronic Disabilities. The workshop brought together computer scientists, medical practitioners, and government officials from various agencies to discuss technologies that will enable older adults and people with disabilities to age in place. Home health technologies also will reduce healthcare costs and enhance quality of life, according to the report. Workshop participants say a new generation of research is needed to address the complexity of supporting the quality of life and independence of the large aging population. The report charts a course for the research agenda needed to advance such technologies, but it also identifies a series of barriers that must be addressed in order to make actionable progress in this area. The report says there is a need to better understand target users, and a need for actionable evidence, effective trans-disciplinary collaboration, far-reaching test beds, patient access to actionable technologies, and information dissemination that bridges the gap between research and practice.

U.N. Report Focuses on Modern Technology to Improve Peacekeeping Missions
FierceGovernmentIT (02/25/15) Dibya Sarkar

United Nations (U.N.) peacekeeping missions should make better use of innovative technology to establish situational awareness, carry out mandates, and protect themselves, according to a report from the Expert Panel on Technology and Innovation in U.N. Peacekeeping. The report offers numerous technology recommendations across a wide range of areas and issues, and notes certain principles should guide the U.N. in acquiring and using technology. The panel says the focus should be on using widely available, but not proprietary, technologies that can be relatively easy to maintain, and prioritizing mobility with regard to maneuverability of assets and mobile information technology platforms, among others. For example, the report says mine-protected vehicles, unmanned aerial vehicles, and smartphone innovations such as the UNMAS Landmines and ERW Safety application could help peacekeeping missions face the threat of landmines, explosive war relics, and improvised explosive devices in multiple languages. However, the panel notes the U.N. also will need to address issues such as limited bandwidth, lack of interoperable systems, and cybersecurity. For the long term, the report says the U.N. needs to create a culture of innovation.

Better Machine Learning
MIT News (02/24/15) Eric Brown

Massachusetts Institute of Technology researcher Kalyan Veeramachaneni joined the Computer Science and Artificial Intelligence Laboratory's Any Scale Learning For All group in 2010 to work on large-scale machine-learning platforms that enable the construction of models from huge data sets. By 2013, the researchers started examining the growing bottleneck caused by the process of translating raw data into formats required by most machine-learning systems. "Our biggest challenge is how to use human input efficiently, and how to make the interactions seamless and efficient," Veeramachaneni says. The researchers tested automated data prep concepts on the GigaBeats project, which analyzes blood pressure signals from thousands of patients to predict a future condition. With GigaBeats, numerous steps that involve human decision-making are needed to prepare the data for analysis. The researchers developed BeatDB, a system to automate the process, which reduced the prep time from months to a few days. "This deep-mining solution combines all layers of machine learning into a single pipeline and then optimizes and tunes with other machine-learning algorithms on top of it," Veeramachaneni says. "It really enables fast discovery." The researchers also are studying better ways to present the processed date, as well as analyzing raw click data from massive open online courses to improve courseware.

Queen's Researchers in Bid to Develop World's Fastest Supercomputers
Queen's University Belfast (02/24/15) Andrew Kennedy

Software developed by researchers at Queen's University Belfast could be used to build some of the world's fastest computers. The exascale computing software will enable supercomputers to process large amounts of data at higher speeds than ever before. Researchers in the School of Electronics, Electrical Engineering, and Computer Science are collaborating with other specialists from the University of Manchester and the STFC Daresbury Laboratory on the Scalable, Energy-Efficient, Resilient, and Transparent Software Adaptation (SERT) project. "This project sheds valuable insight on how to use many core-based systems effectively, proving major benefits for a wide range of scientific endeavors that depend on large-scale simulations," says Jack Dongarra, an honorary visiting professor at the University of Manchester. Complex simulations of climate change, diseases, or weather patterns could be completed in a matter of hours with the software, according to Queen's University professor Dimitrios Nikolopoulos. "This research has the potential to give us insights into how to combat some of the biggest issues facing humanity at the moment," Nikolopoulos says. "It is such an exciting project to be a part of and is further evidence of how Queen's researchers are advancing knowledge and changing lives." The SERT project is slated to get underway in March.

Computational Anthropology Reveals How the Most Important People in History Vary by Culture
Technology Review (02/23/15)

Massachusetts Institute of Technology (MIT) researchers have determined who are the most significant people in the English, German, Chinese, and Japanese versions of Wikipedia. They set out to create a social network of all of the people that appear in a given language version of Wikipedia by downloading articles for all prominent people, and then extracting birth and death dates to determine which people were alive at the same time. The researchers then examined the links on each page to determine which points to whom, and created a network between people who lived at the same time for each year between 3000 B.C. and 1950. Finally, they ranked the people in these networks by importance using the PageRank algorithm. The lists show the most important people of all time in these cultures, and there are some noticeable features that distinguish Eastern from Western societies. For example, the lists of most important people from the Japanese and Chinese versions contain only warriors and politicians, while about half of the most important people from English and German versions are scientists, artists, or religious leaders. "Probing the historical perspective of many different language-specific Wikipedias gives an x-ray view deep into the historical foundations of cultural understanding of different countries," says MIT researcher Peter Gloor.

UW Researches Ways to Draw Women Toward Science Majors
Badger Herald (02/23/15) Riley Vetterkind

Female leaders at the University of Wisconsin (UW) are looking for ways to address the inequalities that remain for women in science, technology, engineering, and mathematics (STEM) fields. Organizations such as the Society of Women Engineers (SWE) support undergraduate women in their professional endeavors on campus and after graduation. "I think women are doing a lot better in schools now because we have these organizations, so I think our generation is more inclusive, but I think there are still areas in the workplace that need to be worked on for inclusion," says SWE member Catherine Groh. Experts say early introduction to STEM fields is important in order to encourage female students to pursue careers in STEM fields in the future. One of the main pillars of SWE is outreach to girls about a future in engineering. The UW chapter of SWE does outreach for fifth- to seventh-grade girls, and also works with the Girl Scouts, which comes to campus for a themed day focusing on a specific engineering field. Research shows that even low exposure to computer programming before children reach high school makes them far more likely to sign up for and stay enrolled in a computer science class for the remainder of their academic experience, notes UW researcher Amanda Ochsner.

Rise of the Fembots: Why Artificial Intelligence Is Often Female
LiveScience (02/19/15) Tanya Lewis

A disproportionate percentage of artificial intelligence (AI) systems have female personas, and researchers have struggled to determine why this phenomenon occurs. One theory is AI systems and humanoid robots tend to perform jobs that have traditionally been associated with women, such as maids, personal assistants, or museum guides. Karl Fredric MacDorman, an expert in human-computer interaction at Indiana University-Purdue University Indianapolis, studies how men and women react to voices of different genders. In one study, MacDorman found men preferred female voices, but they showed no implicit preference for them, whereas women in the study implicitly preferred female voices to male ones. "I think there's a stigma for males to prefer males, but there isn't a stigma for females to prefer females," he says. However, despite the preference for female voices, humanoid robots are often male. "When it comes to a disembodied voice, the chances of it being female are probably slightly higher than of it being male," says Kathleen Richardson, a social anthropologist at University College London. "But when it comes to making something fully humanoid, it's almost always male." When humanoid robots are female, Richardson says they tend to be modeled after attractive, subservient young women.

'Coding for All'
Harvard Magazine (02/20/15) Sophia Nguyen

University of California, Los Angeles (UCLA) researcher Jane Margolis has helped write a high school curriculum, "Exploring Computer Science (ECS)," which aims to expose students to a wide range of topics, including HTML website design, data analysis, robotics, and programming through Scratch. The new curriculum will be paired with a professional development course for teachers, who will learn inquiry-based teaching methods along with the content itself. ECS has received the backing of, and has been adopted by districts in Los Angeles, Spokane, Chicago, and New York City, among others. Margolis is now working with University of Pennsylvania researcher Yasmin Kafai to create an electronic textiles unit for ECS. The researchers hope to appeal to girls by bringing together the crafts of circuitry design and sewing. Kafai says computer science curriculum design is so new researchers should keep their options open. "For computation, you need different materials and activities--it can't all be robotics, it can't all be game design," she says. "We need a whole array, to tease out what activities are good for which concepts, and which age levels." Computational thinking also could be integrated into other classes similar to the language arts model, says Harvard University professor Karen Brennan.

Facebook AI Director Yann LeCun on His Quest to Unleash Deep Learning and Make Machines Smarter
IEEE Spectrum (02/18/15) Lee Gomes

In an interview, Yann LeCun, New York University professor and head of Facebook's Artificial Intelligence (AI) lab, discusses his work as an AI researcher, how AI is portrayed and perceived by the public, and what he sees as the future of the field. LeCun's work focuses on deep learning, which he struggles to provide a concise definition for, except to say direct comparisons to the human brain are often overly simplistic. LeCun says deep-learning networks are in some ways inspired by the structure and function of the human brain, but tend to be quite different in many fundamental ways. LeCun is especially wary of researchers and businesses that portray their AI as being like the brain, suspecting that such claims are more about attracting attention than being accurate. LeCun takes a very measured approach to predicting the future of AI. He expects current research to yield better voice- and visual-recognition systems and self-driving vehicles, but he thinks it is unlikely they will yield human-like intelligence any time soon. Other topics discussed in the interview include ways of giving machines "common sense," the need for public discussion about ethics and AI, and LeCun's attitude toward belief in the technological Singularity.

Abstract News © Copyright 2015 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