Association for Computing Machinery
Welcome to the October 9, 2015 edition of ACM TechNews, providing timely information for IT professionals three times a week.

Please note: In observance of the Columbus Day holiday, TechNews will not be published on Monday, Oct. 12. Publication will resume Wednesday, Oct. 14.

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).


Researchers Say They've Recreated Part of a Rat Brain Digitally
The New York Times (10/08/15) James Gorman

Researchers from institutions around the world disclosed on Thursday the successful digital reconstruction of a section of a rat brain. Blue Brain Project leader Henry Markram describes the breakthrough as the first draft of a functioning map of 30,000 brain cells. Instead of recording the details of every single cell in the brain section, the researchers used the data from some cells to infer what the whole would look like, and then modeled certain kinds of brain activity. They learned the simulation functioned like the living tissue. The research was partly funded by Europe's Human Brain Project, while the Blue Brain Project aims to digitally reconstruct the rat brain and later the human brain. Kavli Neural Systems Institute director Cori Bargmann says the project is in an early stage, comparing it to the construction of a 747. "It's taxiing around the runway," she notes. "I haven't seen it fly yet, but it's promising." Markram envisions a research tool capable of digitally encoding some characteristics of neurons and their linkages that are common to all brains.
View Full Article - May Require Free Registration | Return to Headlines | Share Facebook  LinkedIn  Twitter 

UMD Researchers Present Paper on Innovative Work for Ensuring Integrity in Cloud-Hosted Databases
University of Maryland (10/08/15) Melissa Brachfeld

Researchers at the University of Maryland (UMD) will present a new method for verifying the integrity and completeness of cloud data next week at the ACM Conference on Computer and Communications Security in Denver. The researchers developed a tool they call IntegriDB to ensure the accuracy of results returned by cloud-hosted Structured Query Language (SQL) databases, in addition to detecting any attempt at providing an incorrect result. The tool works by building an authenticated data structure on top of the original data and generating a short public key that depends on this original data. When a cloud server returns the answer to an SQL query, it uses the authenticated data structure to compute a short proof that is sent back to the client along with the result. The client can then use this to verify the correctness of the result against the original key. "We are the first to design efficient cryptographic algorithms and implement them in a system that provides fast verification of most common SQL queries that are executed remotely in a cloud setting," says UMD professor Charalampos Papamanthou. He notes IntegriDB takes only milliseconds to accept or reject an answer.

Ford Uses Space Robot Communications to Serve Connected Cars (10/08/15) Alex Scroxton

Researchers at Ford Motor Co. and St. Petersburg State Polytechnic University have used telematics technology, which is more commonly used in developing space-robot communications, to develop a highly reliable data communications system for connected vehicles. The prototype technology collects and transmits data into the cloud from moving vehicles over cellular channels, Wi-Fi, or via other vehicles, using an intelligent connectivity manager to select the best available connectivity option. The researchers developed a small-scale connectivity coverage map showing local landscape features, incorporating data on the location of fixed and mobile wireless access points, the quality of service on available communications channels, and traffic conditions. They then simulated various scenarios in which critical information would need to be shared between cars. The system's intelligent connectivity manager picks the best communication platform, and then moves onto the next if the first does not work. For non-emergency communications, the system consults coverage maps to determine the best point for service delivery. The new systems are already highly reliable, scalable, and adaptable, says St. Petersburg State researcher Vladimir Zaborovksy. His team is working to advance the multichannel connectivity technology used to operate an Earthbound robot from the International Space Station so scientists on a planet's surface could remotely control robots assembling scientific equipment and satellites in space.

New Programming Approach Seeks to Make Large-Scale Computation More Reliable
UChicago News (IL) (10/07/15) Benjamin Recchie

As computer components become smaller and smaller, packing more transistors into a smaller space than ever before, more errors are likely to crop up in the computations the hardware carries out. This effect is likely to be especially pronounced in high-performance computers, and researchers at the University of Chicago's Computation Institute are looking for a new method to correct for these errors. Most computers today use a technique called checkpoint restart to periodically save data at any given point mid-calculation, so if an error occurs, the computer can revert to an earlier state in the calculation without having to completely start over. However, even this method is likely to be insufficient as complexity grows and errors increase. Andrew Chien and his colleagues at the Computation Institute are experimenting with a new technique called Global View Resilience, which enables applications to not only save work that is underway, but also to offer flexible error-check and self-repair while in operation. Chien's group has tested this method on the Midway supercomputing cluster located on the university's Hyde Park campus. Chien says the new method has proven very reliable at compensating for errors introduced deliberately by the researchers.

A Stanford Professor's Quest to Fix Driverless Cars' Major Flaw
Bloomberg (10/07/15) Keith Naughton

Stanford University professor Chris Gerdes is exploring the issue of programming the computer systems of automated cars with ethical decision-making, because he is skeptical of many people's assertions that the technology is ready for practical deployment. "There's a lot of context, a lot of subtle but important things yet to be solved," Gerdes says. One instance he cites is whether in the event of an unavoidable accident a driverless car should be programmed to make a moral choice between protecting its occupant or protecting a group of pedestrians in its path. "We need to take a step back and say, 'wait a minute, is that what we should be programming the car to think about?" Gerdes says. "Is that even the right question to ask? We need to think about traffic codes reflecting actual behavior to avoid putting the programmer in a situation of deciding what is safe versus what is legal." Gerdes notes in terms of the driverless car technology's development, the hype over its advantages is at an apex. "The benefits are real, but we may have a valley ahead of us before we see all of the society-transforming benefits of this sort of technology," he cautions.

Immigrants Have a Growing Role in the U.S. Sci-Tech Workforce
IEEE Spectrum (10/07/15) Prachi Patel

Immigrants accounted for 18 percent of U.S. scientists and engineers in 2013, up from 16 percent in 2003, according to a U.S. National Science Foundation report. The number of foreign-born scientists and engineers has risen from 3.4 million to 5.2 million during this period and is likely to keep increasing, the report says. Overall, the number of U.S. scientists and engineers rose from 21.6 million to 29 million. However, foreign-born scientists and engineers are more likely to earn higher degrees than their U.S.-born colleagues. In 2013, 9 percent of immigrants earned a doctorate, compared to 3.8 percent of U.S.-born citizens. Almost 15 percent of immigrants earned their highest degree in computer and mathematical science, and more than 20 percent earned engineering degrees. For U.S.-born citizens, those shares were about 8 and 10 percent, respectively. About 81 percent of foreign-born scientists and engineers were employed in 2013--the same percentage as their U.S-born counterparts. The report notes 57 percent of immigrant scientists and engineers hail from Asia, while immigrants from the Americas and the Caribbean make up 20 percent, and Europeans make up 16 percent. India was the leading source nation for immigrant scientists and engineers.

Researchers Build Extra Brainy Smart Homes to Monitor Aging Adults
GeekWire (10/07/15) Lisa Stiffler

For the last eight years, Washington State University's (WSU) Center for Advanced Studies in Adaptive Systems (CASAS) has been developing smart-home technology that harnesses machine learning to help older people live with greater independence. The hope is that CASAS' technology will help fill an expected gap in elder care as the number of Americans ages 85 and older increases dramatically over the next several decades. CASAS' first step was to recruit 400 older volunteers and gather data about their normal, daily routines and conduct interviews to assemble baseline information correlating health and activity levels. The data led to the creation of "Smart Home in a Box," a system of nearly 30 sensors that detect movement, temperature, doors opening and closing, and other information. Four years ago, CASAS began installing the system in the homes of older volunteers in the Seattle and Spokane, WA, areas. CASAS currently is monitoring 40 residences with the technology, and is using the data the sensors are gathering to formulate new systems. One idea is to create a system that can detect when residents forget to take their medication and remind them. "It's a lot to take on, but it's fun to work on and it's really compelling," says WSU professor and CASAS director Diane Cook.

Theoretical Computer Science Provides Answers to Data Privacy Problem
National Science Foundation (10/07/15) Aaron Dubrow

A key challenge of conducting big data research is striking a balance between the desire to reveal new knowledge and protecting the privacy of the people whom the data represents. Harvard University professor Salil Vadhan is investigating "differential privacy" to fulfill this mandate as the lead researcher for the U.S. National Science Foundation (NSF)-funded "Privacy Tools for Sharing Research Data" initiative. Vadhan says the purpose of the project is to develop a computer system that curates and protects sensitive, valuable data. "They take a highly interdisciplinary approach, which brings together deep expertise in computer science, social science, statistics, and law," notes NSF program director Nina Amla. The virtual curator is designed to respond to queries with answers that are approximately accurate while generating a sufficient percentage of "noise" that effectively obscures any specific details about individuals within the database. The system can maintain privacy protection even when confronted with a very large number of questions via judicious increases in the amount of noise and cautious correlation of that noise across queries. Vadhan says differential privacy algorithms were inspired by new theoretical computer science concepts. His team is developing a tool to enable the inclusion of datasets that were previously excluded due to the private nature and uncertain privacy of the information.

Great Innovative Idea--Acquiring Object Experiences at Scale
CCC Blog (10/07/15) Helen Wright

A paper by Brown University researchers John Oberlin, Maria Meier, Tim Kraska, and Stefanie Tellex was one of the winners at the Computing Community Consortium-sponsored Blue Sky Ideas Track Competition at the AAAI-RSS Special Workshop on the 50th Anniversary of Shakey: The Role of AI to Harmonize Robots and Humans in Rome, Italy. The paper, "Acquiring Object Experiences at Scale," described the use of their recently developed Ein software, which runs on the two-armed Baxter robot, to help build a database of manipulable objects. Ein enables Baxters to autonomously manipulate and scan objects. The goal is for researchers to load Ein onto the nearly 300 Baxters being used for research around the world, and then leave the robots on with a pile of objects when they leave the lab for the night. Running Ein, the Baxters will manipulate, scan, and put away the objects. Ein will then use the data gathered by the Baxters to build a massive database that Baxters can draw on to recognize, localize, and manipulate everyday objects. "If all 300 research Baxters ran Ein continuously for 15 days, we could scan 1 million objects," the researchers say. "By collecting data on a large scale, we can begin to tackle category-level inference in object detection, pose estimation, and grasp proposal in ways that have never been done before."

Frogs Resolve Computing Issues
Plataforma SINC (Spain) (10/05/15)

Researchers from the University of the Basque Country (UPV/EHU) and the Technical University of Catalonia have used the Japanese tree frog's mating rituals to develop new computational algorithms. The males of this species have learned to desynchronize their singing patterns so females can tell them apart. "This process is a great example of self-organization in nature, which has allowed us to develop bio-inspired algorithms," says UPV/EHU professor Christian Blum. The researchers used the algorithms to resolve computing issues associated with graphs, which are sets of nodes joined together by links that represent their interrelationships. The researchers used the bio-inspired algorithms to detect an independent set of nodes, meaning those which are not directly connected within the graph. "This has a number of vital applications in communication networks, such as in the formation of wireless backbone networks, but also in social networks such as Facebook and Twitter, since it allows for structural analysis and for the detection of independent communities within these networks," Blum says. The method makes it possible to identify inconspicuous or inactive users, detect tight-knit communities, discover popular individuals within the network, or find users who would be willing to connect in order to improve relations between communities.

For the Real Hits of Fashion Week, Look to Computer Science
Science News (10/05/15) Rachel Ehrenberg

Fashion experts might have some new competition when it comes to spotting general trends, as researchers from Taiwan and the University of Rochester have developed algorithms that enable a computer to identify trends that make their way from the runway to the street. The researchers trained machine-learning algorithms to identify a human figure and nine anatomical sections, and to assess features such as color and texture, clothing categories such as "skirt," and elements such as a placket (an opening in a garment). The researchers created two databases that contained images of recent New York fashions shows and images of people's clothes gleaned from social media sites. They report their program picked up on several general trends and spotted modifications to catwalk styles. Kezhen Chen, who started the project while taking a course with University of Rochester professor Jiebo Luo, believes the research could help garment makers and distributors better tailor supply to meet demand. The team's paper will be presented at the ACM Multimedia conference in Australia later this month.

Intel Fellow Outlines Bright Future for Deep Learning
The Platform (10/05/15) Rob Farber

Intel fellow and Intel Parallel Computing Lab director Pradeep Dubey foresees machine learning being embedded everywhere. He notes the surge of interest in deep learning is fueled by an intersection of innovations that offer sufficient compute capability, large amounts of data, and algorithmic advances to enable supervised training of deep neural networks. This in turn means useful and accurate predictions can now be made in image- or speech-recognition tasks, which are equal or superior to human capability. Dubey details one path to deep-learning success in which people access the cloud to train a model that can address a complicated real-world challenge. "The power of training is that it can potentially create a model that is compact enough to reside and run even on a phone, and thereby enabling the handheld computing device to do a recognition task hitherto far beyond its compute capability," he says. Dubey notes deep learning's biggest challenge is ensuring the training framework is accurate, efficient, and scalable for processing extremely large volumes of data. He stresses all deep-learning algorithms must calculate the cumulative error the model makes over all examples during each and every process of the training algorithm. Dubey hints his group is developing a new smart technique to scale training of deep neural networks to a large number of processing nodes, lowering the time-to-train from the current state of the art.

Patient, Test Thyself
UCLA Magazine (10/15) David Geffner

University of California, Los Angeles (UCLA) physicians, software developers, and computer scientists are leading a healthcare transformation by harnessing the power of wireless mobile technologies. Sage Bionetworks is conducting a research study with an app that employs data culled from smartphones and fitness trackers, developed with advisory assistance from UCLA's Jonsson Comprehensive Cancer Center. Since March, the app has monitored the experiences of thousands of women, some of whom are breast cancer survivors, across a wide age spectrum by tapping personalized data. Another UCLA project features professor Chi On Chui's development of the Semiconductor Electronic Label-Free Assay biosensor, which brings laboratory-quality bio-molecular evaluations to point-of-care settings. The device could reduce emergency room time for heart attack patients because it uses only one drop of blood to screen out patients who need no further medical care. UCLA professors also have developed new apps combining wearable wireless sensor technology with diagnostic medical testing and preventive health care/lifestyle analysis. UCLA Wireless Health Institute co-director Majid Sarrafzadeh created a device that tracks eating by measuring vibrations caused by swallowing, sends those signals to a smartphone, and renders the data as a visual spectrogram. He says this yields daily, weekly, and monthly views of users' eating history, as well as personalized feedback based on eating habits.

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