Welcome to the June 5, 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).

data streams Advances in Bayesian Methods for Big Data
June 2, 2017

Machine learning is an increasingly vital discipline for addressing big data challenges, and widely used Bayesian methods are perceived as having several advantages in the area of "big learning." Researchers at Tsinghua University in China have presented an overview of the latest innovations in Bayesian methods for big data analysis, with an emphasis on the progress of flexible Bayesian methods, efficient and scalable algorithms, and distributed system deployments. "Bayesian methods are becoming increasingly relevant in the big data era to protect high-capacity models against overfitting, and to allow models adaptively updating their capacity," the team notes. "However, the application of Bayesian methods to big data problems runs into a computational bottleneck that needs to be addressed with new (approximate) inference methods." The researchers also say more attention should be devoted to the question of "how to conjoin the flexibility of deep learning and the learning efficiency of Bayesian methods for robust learning."

Full Article
3D Models of Faces Developed by Researchers Could Help in Reconstruction Surgery
Imperial College London
Colin Smith
June 2, 2017

Researchers at Imperial College London and the Royal Free Hospital in the U.K. are scanning volunteers' faces to compile a database of three-dimensional computer models with different expressions, ethnicities, and age groups, which could be used to aid in reconstructive surgery. The data generated by the scanners is sent to the Imperial research group, which for the last 10 years has been developing a system for quickly analyzing thousands of faces. A program maps the various facial landmarks and assigns them coordinates so direct comparisons can be made between the scans. The team then accrues a statistical model of how an average face appears in different stages of growth and in different ethnicities. In addition to surgical applications, the researchers believe their system could be useful as a facial-recognition tool, as an enhancement for lie detection, and as an expression tool for autistic children.

Full Article

Jean Sammet Jean Sammet, Co-Designer of a Pioneering Computer Language, Dies at 89
The New York Times
Steve Lohr
June 4, 2017

Software engineer Jean E. Sammet, who co-designed the Common Business Oriented Language (COBOL) and was elected the first female president of the ACM in 1974, passed away on May 20 at the age of 89. Sammet achieved a level of prominence in computing beyond most women of her generation, and she once said her ambition was "to put every person in communication with the computer," according to University of Maryland professor Ben Shneiderman. The Computer History Museum's Dag Spicer says Sammet's book, "Programming Languages: History and Fundamentals," published in 1969, "was, and remains, a classic" in the field. COBOL remains an essential element in the mainframes underlying corporate and government agency operations worldwide. Sammet worked with five other programmers designing COBOL over a period of two weeks, and the language enabled innovative techniques for describing and representing data in computer code. Sammet later worked to inject more engineering discipline into the language.

Full Article
W&M Joins Elite Group Preparing for Dawn of Exascale Computing
Williamsburg-Yorktown Daily
Joseph McClain
June 3, 2017

Researchers at the College of William & Mary have joined the Exascale Computing Project (ECP), established by the U.S. National Strategic Computing Initiative as part of a multi-institutional collaboration. William & Mary professors Andreas Stathopoulos and Kostas Orginos are seeking to address the problem of lattice quantum chromodynamics (lattice QCD), with Stathopoulos contributing the research's scientific computing component. He says exascale computing resources are needed to model QCD theory, and reaching exascale performance via parallelism requires a network of about 500,000 to 1 million computers. Stathopoulos says the individual machines composing such an array can be created with already established technology, and his focus is on developing algorithms that can use the array to tackle large-scale problems. Stathopoulos notes this involves breaking down complex, sequentially-crafted algorithms into what he calls "dumb" algorithms composed of many more steps, supported on a proportional number of parallel cores, to complete a task.

Full Article

GelSight sensor attached to a robot’s gripper Giving Robots a Sense of Touch
MIT News
Larry Hardesty
June 5, 2017

Researchers at the Massachusetts Institute of Technology (MIT) have mounted GelSight sensors, technology that uses physical contact with an object to provide a detailed three-dimensional map of its surface, on the grippers of robotic arms to provide robotic systems with greater sensitivity and dexterity. For example, the researchers used the data from the GelSight sensor to enable a robot to judge the hardness of surfaces it touches. Separately, another MIT team used GelSight sensors to enable a robot to manipulate smaller objects than was previously possible. In the first experiment, the researchers used molds to create 400 groups of identically-shaped silicone objects, with 16 objects in each group with differing degrees of hardness. The team then pressed the GelSight sensor against each object, recorded how the contact pattern changed over time, and fed the data to a neural network, which sought correlations between changes in contact patterns and hardness measurements.

Full Article
Virtual Reality Eases Phantom Limb Pain
May 31, 2017

Researchers at Aalborg University in Denmark have developed a system that uses virtual reality (VR) technology to relieve sufferers of phantom limb pain. The method involves patients wearing VR goggles and a glove, while electrodes are positioned on their residual limb to stimulate it with small tiny electrical impulses in an attempt to regenerate the sensation of the phantom hand. The amputees play different VR games in which they perform the same actions with both hands, such as gripping a pole that has to be twisted into different shapes or pushing different virtual buttons. In testing the method at the China Rehabilitation Research Center in Beijing last fall, two out of three amputees' phantom limb pain was eased, while the third suffered phantom limb pain attacks less often. The system currently only works with upper-body amputees, but Aalborg students are working to extend its usage for lower-body amputees.

Full Article

Purdue Professor Michael Manfra Microsoft, Purdue Collaborate to Advance Quantum Computing
Purdue University News
Steve Tally
May 30, 2017

Purdue University and Microsoft have launched a long-term enhanced collaboration to construct a robust and scalable quantum computer by producing a "topological qubit." The team will work on a type of quantum computer that is expected to be especially robust against interference from its surroundings. The researchers note this type of scalable topological quantum computer is theoretically more stable and less error-prone. "Topological quantum computing utilizes qubits that store information 'non-locally' and the outside noise sources have less effect on the qubit, so we expect it to be more robust," says Purdue professor Michael Manfra. He says the most exciting challenge with building a topological quantum computer is the Microsoft team must concurrently solve problems of materials science, condensed matter physics, electrical engineering, and computer architecture. The project establishes "Station Q Purdue," an experimental research site that works closely with two other Station Q theory sites in the Netherlands and Australia.

Full Article
Web-Based Search Data Is a New Key to Understanding Public Reaction to Major Societal Events
News at IUPUI
Rich Schneider
May 31, 2017

Researchers at Indiana University-Purdue University Indianapolis (IUPUI) conducted an analysis of millions of Web searches tied to major societal events in order to better understand public reaction. The team analyzed 5.6 million firearm-related search queries on Yahoo occurring two weeks before and two weeks following the 2012 Sandy Hook Elementary School shootings in Connecticut. A key finding was such queries more than doubled immediately after the incident. Retail websites were generally the most visited sites, followed by searches for gun types and ammunition. In addition, most people's information sources were advocating entities--either pro-gun or pro-gun control--and not more neutral entities such as government or educational sites. "This data is a hidden gem to be added to the arsenal of public health," says IUPUI professor Nir Menachemi. He believes the insights derived from such research can help augment debates and inform policy developments related to public health issues.

Full Article

Rice University scientists Ryan Spring and Anshumali Shrivastava Rice U. Scientists Slash Computations for 'Deep Learning'
Rice University
Jade Boyd
June 1, 2017

Researchers at Rice University have adapted "hashing," a technique for rapid data lookup, to reduce the amount of computation required for deep learning. "This applies to any deep-learning architecture, and the technique scales sublinearly, which means that the larger the deep neural network to which this is applied, the more the savings in computations there will be," says Rice professor Anshumali Shrivastava. He notes hashing involves the use of smart hash functions that convert data into manageable small numbers called hashes, which are stored in tables that work much like the index in a book. "Our approach blends two techniques--a clever variant of locality-sensitive hashing and sparse backpropagation--to reduce computational requirements without significant loss of accuracy," says Rice graduate student Ryan Spring. In small-scale tests, the researchers found they could reduce computation by up to 95 percent and still be within 1 percent of the accuracy achieved with standard approaches.

Full Article
AI Will Be Able to Beat Us at Everything by 2060, Say Experts
New Scientist
Timothy Revell
May 31, 2017

There is a 50-percent chance that machines will outperform humans in all tasks within 45 years, according to a survey of 352 artificial intelligence (AI) researchers conducted by Yale University and the University of Oxford in the U.K. For example, they predict machines will dominate language translation by 2024. The survey was sent to AI researchers who published in 2015 at either the Conference on Neural Information Processing Systems or the International Conference on Machine Learning. "There is accumulating evidence that machines can overpower human intelligence in complex, thought-specific tasks," says Eleni Vasilaki at the University of Sheffield in the U.K. Although the survey found no link between the seniority of researchers and the predictions they made, it did reveal a correlation based on where the researchers were from. The survey found researchers based in Asia typically gave shorter timeframes for when AI will completely outperform humans than those in North America.

Full Article

A person wearing an eye sensor Sensor Tracks Eye Movements in Real Time to Enhance Virtual Reality
The Engineer (United Kingdom)
Helen Knight
May 25, 2017

A team of researchers from Belgium and the Netherlands has developed a sensor that can detect eye movements in real time using an array of electrodes. The researchers say the technology could enable videogamers to interact with the screen by controlling cursors with their eyes to navigate through menus and select different options, or to open and close applications. In addition, the technology could provide feedback to the game on how the user is reacting to their virtual surroundings, according to Carlos Agell at the Belgium-based IMEC innovation center. The team developed a system based on four electrodes, which are built into the glasses around each lens, at points where the frames touch the skin in order to collect electrical impulses emitted by the eye. Agell says the algorithm translates the signals into a position, based on the angle the eye is making with its central point of vision.

Full Article

A woman working at a keyboard Preventing Software From Causing Injury
Vital Record
Leah Poffenberger
May 22, 2017

Researchers at the Texas A&M School of Public Health are creating the Self-Report Ergonomic Assessment Tool (SEAT) to help develop safer software by ascertaining how much stress computer programs put on users. The SEAT relies on users self-reporting stress, which obviates the need for special training. "The SEAT can look at two concepts of ergonomic risk: stressors, like bad posture, and strain, pain, or discomfort that comes from the stressor," says Texas A&M's Paul Ritchey. "The idea is for SEAT to be used as a barometer for ergonomic risks through software design cycles." Texas A&M professor Camille Peres says, "We initially leveraged items from existing ergonomic measures and adjusted them so they would be appropriate for self-report and the office environment." The next steps for the tool include refining its accuracy, as well as altering it to help developers eliminate stressors from software to prevent strain from ever happening.

Full Article
MIT Advances in Imaging
ACM Digital Library

Association for Computing Machinery

2 Penn Plaza, Suite 701
New York, NY 10121-0701

ACM Media Sales

If you are interested in advertising in ACM TechNews or other ACM publications, please contact ACM Media Sales or (212) 626-0686, or visit ACM Media for more information.

To submit feedback about ACM TechNews, contact: technews@hq.acm.org