Association for Computing Machinery
Welcome to the March 30, 2016 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).


ACM and Infosys Foundation Honor Innovator in Network Security Research
Association for Computing Machinery (03/30/16)

Stefan Savage from the University of California, San Diego has been selected to receive the 2015 ACM-Infosys Foundation Award in the Computing Sciences. Savage was being cited for research in network security, privacy, and reliability that has showed people how to perceive attacks and attackers as components of an integrated technological, societal, and economic framework. Savage's approach is embodied in his recent work with collaborators to fight spam by exploring how spammers generate revenue, and what steps might be taken to neutralize this incentive. One project involved the researchers infiltrating a botnet to extract insights about the economics of spam schemes. By monitoring millions of spam emails and identifying the individual services needed to monetize them, Savage's team built a model of dependencies in the spam supply chain. They demonstrated merchant bank accounts used to receive credit card payments were the most valuable and prone to disruption. "Stefan Savage has shifted thinking and prompted us to ask ourselves how we might impede the fundamental support structure of an attacker," says ACM president Alexander L. Wolf. "His frameworks will continue to significantly influence network security initiatives in the coming years."

Testing to Start for Computer With Chips Inspired by the Human Brain
The Wall Street Journal (03/28/16) Robert McMillan

The Lawrence Livermore National Laboratory (LLNL) on Thursday will begin testing a $1-million computer packed with 16 IBM TrueNorth microprocessors designed to mimic the functions of the human brain. Bundled into each TrueNorth chip are 5.4 billion transistors comprising a network of 1 million simulated neurons connected by a massive web of synapses. "TrueNorth is useful for deep-learning applications and for a broader class of machine-learning applications as well," says LLNL researcher Brian Van Essen. TrueNorth emulates the brain's low power consumption, with the 16 chips using only 2.5 watts together versus a typical server chip's power requirements of up to 150 watts. Van Essen's team will test TrueNorth by uploading some supercomputing tasks to it. Van Essen expects the system to help the lab filter out potential glitches in simulations of phenomena such as subatomic particle interactions and identify patterns in cybersecurity and video surveillance. "It's great that they're [testing TrueNorth]," says University of Washington professor Luis Ceze. "It's very efficient, but they have to show that the accuracy of the models that they implement [is] good enough."
View Full Article - May Require Paid Subscription | Return to Headlines | Share Facebook  LinkedIn  Twitter 

This UW Program Encourages Women, Minorities Into Computer Science Field
The Badger Herald (WI) (03/30/16) Frankie Hermanek

The University of Wisconsin's (UW) Wisconsin Emerging Scholars in Computer Sciences (WES-CS) program has been active since 2004, and now it aims to broaden its course offerings beyond introductory classes. UW has used the program to draw women and minorities to the computer science field. WES-CS seeks to give students one-on-one support to cultivate confidence, which many aspiring computer science majors lack, according to Mark Hill, chair of the UW Department of Computer Sciences. UW computer science undergraduate program coordinator Nikki Lemmon notes women make up only 14 percent of the school's computer science majors. She also says it is imperative "to have a computing workforce that reflects the true diversity of our society." WES-CS director Tracy Lewis-Williams notes the program concentrates on the introductory course's fusion of logic puzzles and games with sophisticated coding such as JavaScript, while also providing a communal spirit to enrollees. She says WES-SC looks to support a "learning comfort zone" in which instructors can debunk misconceptions and stereotypes about computer science. "Being in the WES-CS program and talking with other students and women who've never programmed before makes me feel much more confident," says program participant and sophomore Saloni Saraf.

Computer Science to Be Added to VA Education Requirements
Richmond Free Press (03/24/16) Jeremy Lazarus

The Virginia General Assembly unanimously passed legislation making the theory and practice of computer operations and the ability to write software code part of a well-rounded education on equal footing with traditional subjects such as reading, writing, and arithmetic. The bill, which is expected to be signed into law by the governor, would make Virginia the first state to require public schools to incorporate computer science and computer coding into K-12 education. However, the legislation leaves it up to the state's Department of Education to implement the reform. The bill's language "does not specify a standalone set of standards" for computer and coding education or require the creation of a Standards of Learning test or assessment on the subject, which students would have to pass to earn a diploma, according to the Virginia Department of Education's Charles Pyle. Nevertheless, the legislation is a step forward for computer science education advocates, as it could enable youths of all backgrounds to have an equal chance of learning the vocabulary of computing, says CodeVA co-founder Chris Dovi. He and his wife organized the nonprofit to offer lessons to teachers and link area youths to coding. More than 300 elementary, middle, and high school teachers have taken CodeVA courses to learn how to implement computer science into their classrooms, and the group aims to train nearly 500 more teachers this year.

Reflection-Removing Camera
MIT News (03/25/16) Larry Hardesty

Researchers at the Massachusetts Institute of Technology (MIT) Media Lab's Camera Culture Group have developed a photograph image-separation process that focuses light into a scene and measures the differences between the arrival times of light reflected by close objects and more distant objects. MIT Media Lab researcher Ayush Bhandari says the team employs the Fourier transform, which is used to deconstruct a signal into its constituent frequencies, with each frequency characterized by its amplitude and phase. The Fourier decompositions of two light signals--one reflected from a nearby object and one from a more distant object--arriving at a light sensor at slightly different times will have distinct phases. However, conventional light sensors can only measure intensity instead of phase, which prompted the researchers to make a few targeted measurements so they could reconstruct phase information. They worked with Microsoft Research on a camera that emits only specific frequencies of light and gauges the intensity of the reflections. Combining that information with knowledge of the number of different reflectors positioned between the camera and the scene of interest enables the algorithms to infer the phase of the returning light and filter out signals from different depths. The algorithms operate via phase retrieval, while additional measurements are needed to filter out noise.

IBM Wants to Accelerate AI Learning With New Processor Tech
Engadget (03/28/16) Steve Dent

Researchers from IBM's T.J. Watson Research Center think they can reduce the required power and learning times in deep neural networks (DNNs) with resistive processing units (RPUs), theoretical chips that combine central processing units (CPUs) and non-volatile memory. The devices could accelerate data speed exponentially, resulting in systems that can perform a wide range of tasks. Conventional neural networks must perform billions of tasks in parallel, a process that requires billions of memory calls. The IBM researchers debated using new storage technology such as resistive random-access memory (ReRAM) that can permanently store data with distributed RAM-like speeds. However, the researchers settled on developing an RPU that puts large amounts of ReRAM directly onto a CPU. RPUs could retrieve data as quickly as they can process it, dramatically decreasing DNN training times and power required. "This massively parallel RPU architecture can achieve acceleration factors of 30,000 compared to state-of-the-art microprocessors...problems that currently require days of training on a datacenter-size cluster with thousands of machines can be addressed within hours on a single RPU accelerator," the researchers say. Although RPUs are still in the research phase, the IBM team believes it is possible to build such chips with traditional complementary metal-oxide semiconductor technology.

Facebook Taps Artificial Intelligence for Users With Disabilities
USA Today (03/26/16) Jessica Guynn

Facebook is harnessing the power of artificial intelligence (AI) to improve the social networking experience for people with disabilities. In April, the company plans to launch an automated captioning tool, which will help the visually impaired "see" photos by describing what is in them. When a friend uploads a new photo, the tool would say, for example, "Image may contain: two people, one toddler, smiling, outdoors." Facebook eventually plans to provide a much fuller automated description of photographs and then videos. "These are our very first baby steps," says Matt King, a visually impaired software engineer who is part of Facebook's accessibility team. "It's really the idea that we are including everybody in the conversation." Facebook's "empathy lab" consists of devices that browse the social network using keyboard shortcuts, braille, or the sound of a human voice. The devices are strategically placed along a busy walkway to remind engineers to build accessibility into all products. King believes AI eventually will enable even greater advances for people with disabilities. "This is a problem that as machines get smarter, that machines can solve," he says.

The MIT Lab Flushing Out a City's Secrets
The Guardian (03/27/16) Nicola Davis

Massachusetts Institute of Technology (MIT) researchers involved in the "Underworlds" project believe analyzing the contents of sewage could provide scientists with new insights into human health. "Our hypothesis is that contaminated water, before it gets to the treatment plant, has an imprint of all the human activities that are going on--in the sewage," says MIT professor Eric Alm. Part of the project focused on designing a robot to collect samples from sewers. MIT's Senseable City Lab created Luigi, a long tube-like instrument complete with removable filters that can be suspended from a bar across the manhole and lowered into the depths of the sewers. "As soon as we turn it on it goes down [and] it samples for a set amount of time," says Underworlds project lead Newsha Ghaeli. She notes Luigi is controlled with an iPhone app. The team recently sampled sewers around Cambridge and is now working in Boston. They think analyzing a city's wastewater could enable outbreaks to be anticipated, antibiotic resistance to be mapped, and public health interventions to be monitored in near-real time.

One Genius' Lonely Crusade to Teach a Computer Common Sense
Wired (03/24/16) Cade Metz

Researcher Doug Lenat and colleagues have spent 35 years developing a "common sense engine" that tries to digitally codify all the fundamental concepts humans intuitively grasp but machines have never really understood. The Cyc engine was fed 15 million logical rules, which collectively can imbue machines with an approximation of common sense, according to Lenat. He says Cyc is ready for real-world applications, and it already has been used to explore new medical research areas, to watch the internal operations of a major financial firm's technological infrastructure, and to identify terrorist threats in vast troves of international communications data. Lenat's artificial intelligence philosophy goes against the approach espoused by most tech companies today, of using neural networks as the basis for machines that can more rapidly approach human intelligence by mining digital data for patterns. Nevertheless, Lenat sees flaws in the deep-learning approach. He notes machine learning's excellence in performing tasks such as image recognition and language translation are dependent on statistical analysis, which means neural nets are prone to error. Lenat thinks their accuracy could be improved with symbolic reasoning guidance from a common sense engine such as Cyc. University of Montreal professor Yoshua Bengio acknowledges neural nets analyzing data supplied by a Cyc-like project could facilitate common sense.

Hu's Research Focuses on Understanding and Predicting User Behavior by Mining Social Media Events
Texas A&M University (03/29/16) Rachel Rose

Researchers at Texas A&M University's Data Analytics at Texas A&M Lab (DATA) are working with colleagues at Arizona State University and Yahoo! to understand user behavior on social media platforms by developing novel data-mining algorithms, which generate predictions based on a user's personal social media presence, as well as on interaction with friends and friends' opinions. The researchers note their actions are motivated by homophily and social influence, two fundamental social theories dictating that people befriend others who are similar to them, or they become more similar to their friends over time. The DATA researchers are focusing on social spamming detection and understanding user behavior by analyzing people's online posts and social interactions. Social spamming occurs when unwanted spam content appears on social networks, and it is often intended to boost a user's social influence, legitimacy, and credibility. "Since social media provides a platform for people to share thoughts on various events, we have done work on identifying rhetorical questions, online protest participation, advocates of political campaigns, health discussions, et cetera," says Texas A&M professor Xia Hu. He says understanding and processing data produced by social media services is needed to improve the quality of user experience, and to positively impact the overall value of the social systems.

UCLA Researchers Develop Sophisticated Open Source Program for Analyzing Thyroid Health
UCLA Newsroom (CA) (03/28/16) Bill Kisliuk

University of California, Los Angeles (UCLA) researchers have developed Thyrosim, an open source program that simulates the response of the human thyroid hormone regulation system to a variety of treatments and diseases. Thyrosim can be used by clinicians, researchers, and educators to accurately gauge the impacts of thyroid treatments and to develop more effective remedies for thyroid problems. "Thyrosim offers an easy-to-use interface for a sophisticated mathematical model of the short-term and long-term impact of thyroid diseases, treatments, hormone supplements, and other interventions," says UCLA professor Joseph DiStefano III. He notes the software will benefit clinical and research endocrinologists and teachers, and could result in positive changes in the use or regulation of available remedies. DiStefano says Thyrosim users can enter data into the interface, and then the software, which relies on a mathematical model based on comprehensive clinical data, simulates likely responses. The system also can project the long-term impacts of individual treatment programs. "Components of Thyrosim are distributed over two computers: the client-user machine accessible via a Web browser, and a remote server at UCLA where the simulation computations are run," says UCLA researcher Simon Han.

Athletes Look for an Edge in a New Place: Virtual Reality
The Conversation (03/28/16) Galen Clavio

Although this may finally be the year virtual-reality (VR) products achieve mainstream consumer acceptance, some athletes already have begun to explore the promise of the technology, writes Indiana University Bloomington professor Galen Clavio. For example, several National Football League teams are using Stanford University's STRIVR system, which provides realistic, repetitive training by visualizing the situations athletes will face during competition. Regular two-dimensional video often uses a wide camera angle to capture the entirety of a formation or play, and the result leaves players feeling distant from the material they are studying. However, the typical VR video consists of footage from multiple cameras, shooting and recording in sync. With STRIVR, players can put on a VR headset and feel as if they are inside the play as it is taking place. They can repeatedly watch different aspects of looped plays within the VR headset, which can enable them to memorize plays and formations without having to step onto the field. Clavio notes although athletes have a positive view of the technology, scientists are just starting to evaluate the mental and psychological impacts of VR.

Linux at 25: Q&A With Linus Torvalds
IEEE Spectrum (03/29/16) Stephen Cass

With the Linux operating system reaching its 25th anniversary this year, Linux kernel creator Linus Torvalds reflects in an interview on what the past brought about and what the future holds for Linux. He notes Linux's transition from a personal effort to a collective one was a difficult change, but he learned it was easier to trust the various submaintainers with applying their patches, rather than applying them himself. "Linux has almost become the default environment for prototyping new hardware or services," Torvalds says. He blames "user inertia" on Linux's inability to significantly penetrate mainstream desktops, noting "the desktop is simply that it's very personal--you interact with it rather intimately every day if you work with computers--but also complicated in ways many other computing environments aren't." Torvalds still sees value in the desktop market, even though the general-purpose desktop concept appears to be giving way to more specialized, simpler, and multitasking platforms. In terms of the current state of Linux, Torvalds believes the kernel is working well, despite concerns it has become too complicated for people to understand and fix defects. He says supporting the growing plethora of hardware is an ongoing challenge. Torvalds' outlook for the future is "we'll do fine as long we keep track of all the small day-to-day details, and try to do the best we can."

Abstract News © Copyright 2016 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