ACM TechNews

RSS

Timely Topics for IT Professionals

ACM TechNews Home

Recent News

ACM Affirms Obligation of Computing Professionals to Use Skills for Benefit of Society

Association for Computing Machinery

ACM has completed a two-year process to update its Code of Ethics and Professional Conduct to address advances in computing technology and the growing pervasiveness of computing in all aspects of society since the code was last updated in 1992. ACM's Code of Ethics is considered the standard for the computing profession, and the organization expects all members to abide by the code. The code is a collection of principles and guidelines designed to help computing professionals make ethically responsible decisions in professional practice. For example, the updated code adds new responsibilities for computing professionals in leadership roles to "recognize and take special care of systems that become integrated into the infrastructure of society." Donald Gotterbarn, co-chair of ACM's Committee on Professional Ethics, says the organization "outlined overarching principles, explanations, and guidelines to steer decision-making based on the understanding that the public good is always the primary consideration."

From "ACM Affirms Obligation of Computing Professionals to Use Skills for Benefit of Society"
Association for Computing Machinery (07/17/18) Jim Ormond
View Full Article

Tech Leaders Sign Global Pledge Against Autonomous Weapons

University of New South Wales Sydney Newsroom

A coalition of prominent CEOs, engineers, and scientists from the technology industry signed a global pledge to neither participate in nor support the development, manufacture, trade, or use of lethal autonomous weapons. "We...call upon governments and government leaders to create a future with strong international norms, regulations, and laws against lethal autonomous weapons," the pledge says. Released at the 2018 International Joint Conference on Artificial Intelligence on Sweden, the pledge was signed by 150 companies and more than 2,400 individuals from 90 countries working in artificial intelligence (AI) and robotics. Toby Walsh at the University of New South Wales in Australia said, "We cannot hand over the decision as to who lives and who dies to machines." Corporate signatories include Google DeepMind, University College London, the XPRIZE Foundation, the European Association for AI, and the Swedish AI Society.

From "Tech Leaders Sign Global Pledge Against Autonomous Weapons"
University of New South Wales Sydney Newsroom (07/18/18) Wilson da Silva
View Full Article

Columbia University, IBM Establish Center to Accelerate Innovation in Blockchain, Data Transparency

PR Newswire

Columbia University and IBM together have created a new center that will concentrate on blockchain technology and data transparency research, education, and innovation. The Columbia-IBM Center for Blockchain and Data Transparency also will feature an innovation accelerator to nurture business concepts from entrepreneurial students, faculty, and startup community members. Cross-disciplinary academic, scientific, business, and government teams will investigate core issues concerning the policy, trust, sharing, and usage of digital data with blockchain and other privacy-maintaining solutions. Innovation areas will include secure multi-party computation, homomorphic encryption, secure hardware, fraud reduction, and enhanced precision medicine via insight from collective data sources. Columbia provost John H. Coatsworth anticipates the new center “will significantly advance scholarship and applications of data-sharing and data-transparency technologies.”

From "Columbia University, IBM Establish Center to Accelerate Innovation in Blockchain, Data Transparency"
PR Newswire (07/17/18)
View Full Article

Researchers Tricked AI Into Doing Free Computations It Wasn't Trained to Do

Motherboard

Google researchers deceived a machine-learning neural network into performing free computations via malicious image manipulation, which could be used to hijack smartphones as botnets. The researchers made small alterations to images to trick a machine vision algorithm into executing a task other than the one it was trained to conduct, essentially rewriting software taught to recognize objects such as animals, to tally the number of squares in an image. The researchers achieved this by generating images consisting of psychedelic static with a black grid in the center, with some squares in the grid randomly selected to be white. They then arbitrarily mapped these adversarial images to image classifications from the ImageNet database, to represent the number of white squares in the adversarial image. According to the researchers, “A variety of nefarious ends may be achievable if machine learning systems can be reprogrammed by a specially crafted input.”

From "Researchers Tricked AI Into Doing Free Computations It Wasn't Trained to Do"
Motherboard (07/12/18) Daniel Oberhaus
View Full Article

Data Mining Reveals Fundamental Pattern of Human Thinking

Technology Review

Researchers at the Communication University of China conducting a data mining study of word frequency patterns found that the human brain processes common and uncommon words differently. The researchers tapped two large text collections that included samples from 50 languages, with at least 30,000 sentences and up to 43 million words in each sample. The word frequencies across all languages followed a modified version of Zipf’s Law (which says that given a large sample of words, the frequency of any single word’s use is inversely proportional to its rank in the frequency table) in which the distribution can be split into three segments, with each segment exhibiting distinctive linguistic properties. Their simulation of the brain's dual-process theory scheme yielded the same three-segment structure in the word frequency distribution. The team suggests this discovery should have significant ramifications for natural language processing research.

From "Data Mining Reveals Fundamental Pattern of Human Thinking"
Technology Review (07/16/18)
View Full Article

Why Kotlin Is Exploding in Popularity Among Young Developers

TechRepublic

Software developer JetBrains' Kotlin is the second-most-popular coding language among software developers after Rust, according to a Stack Overflow survey of 2,744 Kotlin developers. The poll estimates that Kotlin's usage doubled annually from 2011 to 2015, when Square adopted the language. In May 2017, Google made Kotlin an official Android language with complete support, followed by a surge of adoption by Android users, especially students and newer developers. Of the polled Kotlin developers who are actively working, more than 60% said they currently use Kotlin in work projects, while about 55% said they use Kotlin exclusively for side projects. Respondents listed null safety, extension functions, and Java interoperability as their favorite Kotlin features.

From "Why Kotlin Is Exploding in Popularity Among Young Developers"
TechRepublic (07/12/18) Alison DeNisco Rayome
View Full Article

Nvidia, MIT Get a Step Closer to 'Computer, Enhance' Image Cleaning

ZDNet

Nvidia, the Massachusetts Institute of Technology, and Aalto University in Finland collaborated to develop a system that can clean up and augment low-light and astronomical photos, magnetic resonance imaging (MRI), and other images without having to see the original. The researchers say their Noise2Noise solution uses the same training process as cutting-edge neural networks, with few shortcomings in training time or performance; as a result, the researchers said, it is “on par with state-of-the-art methods that make use of clean examples, using precisely the same training methodology, and often without appreciable drawbacks in training time or performance.” They added, "Of course, there is no free lunch; we cannot learn to pick up features that are not there in the input data, but this applies equally to training with clean targets."

From "Nvidia, MIT Get a Step Closer to 'Computer, Enhance' Image Cleaning"
ZDNet (07/11/18) Chris Duckett
View Full Article

Robot Prototype Will Let You Feel How It's 'Feeling'

Cornell Chronicle (NY)

Researchers in Cornell University’s Human-Robot Collaboration and Companionship Laboratory have developed a prototype robot capable of expressing "emotions" by reshaping the texture of its outer surface. Its skin covers a grid of texture units governed by fluidic actuators, which can configure into variants of goosebumps or spikes that are keyed to different emotional states. Cornell's Guy Hoffman says the project's future challenges include scaling the technology for incorporation into a robot, and making it more responsive to the machine's immediate emotional changes. The researchers believe "that the integration of a texture-changing skin, combining both haptic [feel] and visual modalities, can thus significantly enhance the expressive spectrum of robots for social interaction."

From "Robot Prototype Will Let You Feel How It's 'Feeling'"
Cornell Chronicle (NY) (07/09/18) Tom Fleischman
View Full Article

USD Puts NSF Grant Into Action for STEMWoW Summer Program

USD News Center

This summer, the University of San Diego has launched a program to inspire young people to pursue science, technology, engineering, and math (STEM) careers, developed with a $300,000 U.S. National Science Foundation grant. The Bridging the World of Work and Informal STEM Education (STEMWoW) program is designed to help participants explore their strengths, interests, and values; expose them to STEM careers, and engage them in STEM activities. STEMWoW's Summer Academy includes hands-on activities driven by social issues such as poverty, health, and climate change, and exploration of jobs available in San Diego's five priority sectors of advanced manufacturing, clean energy, healthcare, life sciences, and information communication technologies.

From "USD Puts NSF Grant Into Action for STEMWoW Summer Program"
USD News Center (07/13/18)
View Full Article

Unique Brain 'Fingerprint' Can Predict Drug Effectiveness

McGill Newsroom

Researchers at McGill University's Montreal Neurological Institute and Hospital and the Ludmer Center for Neuroinformatics and Mental Health in Canada have developed a personalized Therapeutic Intervention Fingerprint (pTIF), which can predict the efficacy of targeting specific biological factors that govern neurological disease progression. The team analyzed neurological data from 331 Alzheimer's patients and healthy controls via computational brain modeling and artificial intelligence, and then categorized subjects into their TIF subtypes based on the most potentially effective factor-specific interventions. The subtypes were confirmed by comparing them to the patients' individual genetic profiles, and patients in the same pTIF subtype exhibited similar gene expression. Since medications to control disease progression would have to alter gene expression and brain properties concurrently, drugs individualized to pTIF subtypes would be more effective than drugs designed to treat all Alzheimer's patients.

From "Unique Brain 'Fingerprint' Can Predict Drug Effectiveness"
McGill Newsroom (07/10/18)
View Full Article

Electrical Contact to Molecules in Semiconductor Structures Established for the First Time

University of Basel

A new technique that enables electrical contact to be established with simple molecules in a conventional silicon chip has been developed by a team of researchers from the University of Basel in Switzerland and IBM Research–Zurich. The team found thousands of stable metal-molecule-metal components can be produced at the same time by depositing a film of nanoparticles onto the molecules, without compromising their properties. They demonstrated the new approach using alkane-dithiol compounds comprised of carbon, hydrogen, and sulfur. The new technique largely resolves issues that had hampered the creation of electrical contacts to molecules, like high contact resistance or short circuits by filaments penetrating the film. "Our approach will help speed up the development of chemically fabricated and controllable electronic and sensor components," says Basel's Marcel Mayor.

From "Electrical Contact to Molecules in Semiconductor Structures Established for the First Time"
University of Basel (07/12/18)
View Full Article

Motivating Gamers With Personalized Game Design

University of Waterloo News

Scientists at the University of Waterloo in Canada have identified three basic video game player traits that can help make game design more personalized, and motivate gamers in both entertainment and work applications. The new model generates scores for three different gamer traits, including the degree to which players prefer action elements, aesthetic aspects, or goal orientation in games. Identifying these traits enables researchers to analyze player preferences for different groups of people. The team analyzed a dataset of more than 50,000 respondents who had been surveyed for an earlier player satisfaction model, for whom the researchers identified player archetypes, including seeker, survivor, daredevil, mastermind, conqueror, socializer, and achiever. "If we can build systems that can adapt to and accommodate individual differences, interactive systems become more exciting and motivating for every one of us," says Waterloo's Lennart Nacke.

From "Motivating Gamers With Personalized Game Design"
University of Waterloo News (07/09/18)
View Full Article

Think Your Surfing Is Secret in Private Browsing Mode? Think Again

New Scientist

Many computer users incorrectly believe that private browsing protects them against computer viruses, targeted ads, geolocation, and tracking by employers and governments, says Yasemin Acar at Leibniz University Hannover in Germany. Although such privacy modes do stop browsers from saving the user's search history, login information, and cookies, they fail to properly explain their scope in the disclosure statement. The researchers asked participants to read a disclosure statement for a fictional private browser called Onyx, then quizzed them about what protections it offered. The team found even after reading the statement, 23% of respondents incorrectly believed their government could not track their browsing history when in private mode, and 37% thought the same about their employer. Acar says these findings suggest private browsers' disclosure statements should more plainly state their limitations, so users do not have a false sense of security.

From "Think Your Surfing Is Secret in Private Browsing Mode? Think Again"
New Scientist (07/13/18) Alice Klein
View Full Article - May Require Paid Subscription