University of Cincinnati Online MS in IT
 
Welcome to the June 3, 2020 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).

To view "Headlines At A Glance," hit the link labeled "Click here to view this online" found at the top of the page in the html version. The online version now has a button at the top labeled "Show Headlines."

Cynthia Dwork Harvard Professor Receives Prize for Contributions to Theoretical Computer Science
HPCwire
June 2, 2020


The ACM Special Interest Group on Algorithms and Computation Theory and the IEEE Computer Society Technical Committee on the Mathematical Foundations of Computing have named Harvard University professor Cynthia Dwork to receive the 2020 Donald E. Knuth Prize for contributions to theoretical computer science. Dwork’s research is credited with having had a transformative effect on distributed systems, cryptography, data privacy, and fairness in algorithmic decision-making. Dwork also is known for introducing and developing differential privacy, and for her accomplishments in nonmalleability, lattice-based encryption, concurrent composition, and proofs of work. Her foundational contributions to distributed systems include work on consensus, while her achievements in algorithmic fairness include formalization of the "treat like alike" principle.

Full Article
Carnegie Mellon Tool Automatically Turns Math Into Pictures
Carnegie Mellon University
Byron Spice
June 2, 2020


A tool created by Carnegie Mellon University (CMU) researchers allows anyone to render mathematical abstractions as illustrations. The Penrose tool enables diagram-drawing experts to encode their math-into-diagram methods; users simply type in an ordinary mathematical expression, and Penrose produces the drawing. Once the computer learns how the user wants to visualize mathematical objects, it uses the encoded rules to draw several candidate diagrams, which the user can from choose and edit. The researchers created a special programming language for this purpose, which CMU’s Keenan Crane said mathematicians should have no trouble learning. Said Crane, "Our vision is to be able to dust off an old math textbook from the library, drop it into the computer, and get a beautifully illustrated book—that way, more people understand."

Full Article
Evacuating Virtual Buildings
TU Delft
May 26, 2020


Researchers at Delft University of Technology (TU Delft) in the Netherlands are studying how virtual reality (VR) and augmented reality (AR) may help explain the behavior of pedestrians. VR and AR can simulate real-life situations, including potentially dangerous ones. The researchers asked 150 volunteers to navigate a virtual version of the university’s Civil Engineering and Geosciences (CEG) building to study pedestrian behavior in multi-story buildings. Participants roamed the virtual building using either VR glasses or a computer. The participants spent about 20 minutes in the virtual building; after completing three tasks, they were told to evacuate. The researchers found that participants who used the computer were much less likely to show signs of stress during the evacuation than those who used VR glasses. Said TU Delft’s Yan Feng, “To participate in the experiment via a desktop screen is much more like playing a computer game. With VR glasses you really feel you’re in the middle of a situation.”

Full Article
Online Master's Degree in Data Science a First for UT Austin
UT News
June 2, 2020


The University of Texas at Austin (UT Austin) will debut an online master's degree program in data science, as a collaboration between the school's departments of Computer Science, and Statistics and Data Sciences, and online learning company edX. Participants will be able to complete the degree for about $10,000, far less than the cost of similar degrees at other competitive, nationally ranked institutions. "Too often, qualified students forgo graduate study because of factors such as family obligations, the need to maintain an income, or the fear of not being able to afford tuition," said UT Austin’s Don Fussell. "Our objective when we embarked on this project was to create the first technical data science master's degree that didn't force students to make those tradeoffs."

Full Article
People Who Compare Themselves with Others Spend Longer on Facebook
New Scientist
Donna Lu
May 28, 2020


Researchers at Facebook have found that people who compare themselves with others, both offline and online, spend more time on the social media platform than those who do not. The researchers surveyed 37,729 people from 18 countries, and compared their responses to a log of their activity on Facebook from the previous four weeks. The logs included data such as the number of posts a user looked at, and time spent looking at profiles of demographically similar people. Those who reported more frequent social comparisons had more Facebook friends and spent more time on the platform than those who had fewer social comparisons, the researchers found. The researchers suggest certain Facebook design changes could reduce social comparison, such as hiding "Like" counts and reminding users that others' lives are not as ideal as they may seem online.

Full Article
Algorithms Used in Policing Face Policy Review
The Wall Street Journal
David Uberti
June 1, 2020


The U.S. Government Accountability Office (GAO) in May published the first half of a review of privately developed algorithms used in policing, ahead of its issuing policy recommendations for their regulation. The Federal Bureau of Investigation, the Department of Homeland Security, and other U.S. agencies use these algorithms to mine massive databases of evidence, in order to chase leads and support the prosecution of cases. GAO’s Karen Howard said the algorithms are only as good as the underlying forensic data collected by investigators, and how they work is sometimes unclear. The second part of the study, which will evaluate the algorithms’ accuracy and offer make policy recommendations about them, is expected later this year.

Full Article
*May Require Paid Registration
Gamemakers Inject AI to Develop More Lifelike Characters
Wired
Will Knight
May 27, 2020


Researchers at game developer Electronic Arts and Canada’s University of British Columbia are testing artificial intelligence (AI) techniques to accelerate game development and make characters more lifelike. The team is using reinforcement learning to automatically animate humanoid game characters. The researchers trained a machine learning model to identify and recreate statistical patterns in motion-capture data, then used reinforcement learning to train another model to replicate realistic motion with a specific goal, like running toward a ball. The result is animations not included in motion-capture data. Automating the animation process with AI could save game developers millions of dollars while creating games with realistic characters that are sufficiently efficient to run on a smartphone.

Full Article

The robot chef in mid-omelet. A Good Egg: Robot Chef Trained to Make Omelets
University of Cambridge (UK)
June 1, 2020


Engineers at the University of Cambridge in the U.K. and Turkish domestic appliance company Beko trained a robot to make omelets via machine learning, refining its culinary skills to produce a consistently tasty dish. The machine learning method used Bayesian inference to extract the greatest amount of information possible from a limited number of data samples. To account for the subjectivity of human taste, Cambridge's Fumiya Iida said, "We needed to tweak the machine learning algorithm—the so-called batch algorithm—so that human tasters could give information based on comparative evaluations, rather than sequential ones." Added Iida, “The omelettes, in general, tasted great; much better than expected!”

Full Article

The Sign in with Apple vulnerability could have enabled account takeovers. Apple Fixes Bug That Could Have Given Hackers Full Access to User Accounts
Ars Technica
Dan Goodin
June 1, 2020


The Sign in with Apple tool, which allows users to log in to third-party apps without revealing their email addresses, has fixed a bug that could enable attackers to gain access to those accounts. App developer Bhavuk Jain reported the zero-day vulnerability in the privacy-enhancing tool to Apple as part of the company's bug bounty program, and received a $100,000 reward. Sign in with Apple logs in users with either a JSON Web Token (JWT) or a code generated by an Apple server, which is then used to generate a JWT. Users can share the Apple email ID with a third party or keep it hidden, and in the latter instance, Apple creates a JWT that contains a user-specific relay ID. Jain found that an attacker could forge a JWT by linking any email ID to it, which would provide access to the victim's account.

Full Article

book news, illustration Blockchain to the Rescue of Small Publishers
QUT News
June 2, 2020


Researchers at Queen's University of Technology (QUT) in Australia and publisher Tiny Owl Workshop have developed a blockchain system for digital rights management and royalty distribution to facilitate new commercial opportunities for small publishers. The system uses open source blockchain technology to manage IP agreements and royalty payments, track purchases with a custom digital ledger, and monitor physical book sales via a Print and Electronic tracking system, in conjunction with a “marketing bellyband” containing a QR code. "This code gives purchasers of physical book copies a free download of one digital bundle from the ‘Education Edition’, linking physical book purchases in bookstores to online downloads; and providing a ledger of where customer transactions originate from," says QUT’s Mark Ryan.

Full Article

Flag of the United Nations UN: Computer Simulation Tool Could Boost Global Development
Technology Review
May 29, 2020


The United Nations is backing a new computer simulation tool that could help governments boost sustainable development and address global challenges. The Policy Priority Inference (PPI) software employs agent-based modeling to predict and simulate what would happen if policymakers invested in one project rather than another, informing the simulation with economics, behavioral science, and network theory. PPI allocates funding to "bureaucrats" who spend their apportioned money on different projects, then applies data about government budgets, the historical impact of spending on past policies, the effectiveness of a country's legal system, and estimated losses due to known inefficiencies. The software then suggests which policies are most worthy of investment. The goal is to help policymakers understand the wider ramifications of their decisions.

Full Article

MIT algorithm-based search and rescue gear Search-and-Rescue Algorithm Identifies Hidden 'Traps' in Ocean Waters
MIT News
Jennifer Chu
May 26, 2020


Researchers at the Massachusetts Institute of Technology (MIT), the Woods Hole Oceanographic Institution (WHOI), Virginia Polytechnic Institute and State University, and the Swiss Federal Institute of Technology have developed a technique to help first responders quickly identify regions of the ocean where missing objects or people are likely to be. Underlying the new technique is an algorithm that analyzes ocean conditions, such as the strength and direction of ocean currents, surface winds, and waves; it then identifies in real time the regions of the ocean where floating objects are most likely to converge. In field experiments, the researchers found the algorithm correctly predicted which regions would most strongly attract objects based on present ocean conditions.

Full Article
Butterfly Landmines Mapped by Drones, Machine Learning
The Engineer (UK)
May 27, 2020


Researchers at the State University of New York at Binghamton have found that drones and advanced machine learning can be used to detect improvised explosive devices (IEDs) and butterfly landmines (surface plastic landmines with low-pressure triggers). The researchers used convolutional neural networks (CNNs) to develop a method for automating detection and mapping of landmines. According to the researchers, a CNN-based approach is much faster than manually counting landmines from an aerial image, and unlike subjective human visual detection, it is quantitative and reproducible. Binghamton’s Alek Nikulin said drone-assisted mapping and automated detection of scatterable mine fields “would assist in addressing the deadly legacy of widespread use of small scatterable landmines in recent armed conflicts and allow to develop a functional framework to effectively address their possible future use.”

Full Article
The Handbook of Multimodal-Multisensor Interfaces, Volume II
 
ACM Author's Rights
 

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434
1-800-342-6626
(U.S./Canada)



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: [email protected]