ACM TechNews


Welcome to the March 18, 2019 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."

Angle-of-attack sensor Sensor Cited as Potential Factor in Boeing Crashes Draws Scrutiny
The Washington Post
Todd C. Frankel
March 17, 2019


Angle-of-attack sensors that Boeing was ordered to replace following the near-crash of an Airbus A321 in 2014 are drawing fresh scrutiny after two recent Boeing 737 disasters, including last week's deadly crash in Ethiopia. Experts say the risks posed by a defective angle-of-attack sensor are exacerbated by the increasing role of cockpit software. According to the experts, errors in how software interprets sensor readings can lead to unpredictable complications. The angle-of-attack sensor measures the amount of lift generated by the wings, and its chief purpose is to warn pilots when the plane could stall from deficient lift, causing a loss of control. The U.S. Federal Aviation Administration (FAA) said Airbus A320 planes with certain sensors manufactured by United Technologies and Sextant/Thomson "appear to have a greater susceptibility to adverse environmental conditions" than sensors made by a third company.

Full Article
*May Require Paid Registration
The Autocrat's New Tool Kit
The Wall Street Journal
Richard Fontaine; Kara Frederick
March 15, 2019


New technologies will enable autocrats to tighten their grip on citizenry, undermine basic rights, and spread illiberal practices outside of their own countries. Microtargeting is poised to spread beyond commercial aims, with artificial intelligence (AI)-driven apps expected to allow authoritarians to analyze patterns in a population's online activities, identifying persons most susceptible to propaganda and targeting them more precisely. Lisa-Marie Neudert with the University of Oxford in the U.K. warns of intelligent next-generation online bots spreading disinformation and attacking dissent more perniciously, by taking over for paid false commenters in tyrannical regimes. Also contributing to the chaos will be more insidious digital forgeries or "deep fakes," using speech synthesis and other innovations to deceive listeners even more convincingly. Computer scientist Yoshua Bengio said he is concerted the growing use of technology for political control amounts to “the 1984 Big Brother Scenario. I think it’s becoming more and more scary.”

Full Article
*May Require Paid Registration
Chinese Children Take to Coding Amid Country's Lofty Goals in AI
South China Morning Post
Celia Chen
March 17, 2019


China's ambitions to become unparalleled in artificial intelligence (AI) and other advanced technologies, and transform the country's economy by making high-tech industries world-class, are reflected in Chinese children's increasing enrollment in coding classes. Interest in coding classes has risen in the past decade as China's online populace and mobile phone user base expanded into the world's largest, buoyed by government policies encouraging the development of innovative Internet applications, and the digital transformation of traditional industries. Coding education was elevated in 2017, when China's State Council released a statement suggesting AI curriculum development in primary and secondary schools. Venture capitalist Wei Guoxing said, "Coding has definitely become one of the most promising sectors [in China's education industry]. It's not only an important curriculum in schools, but it helps kids with logical thinking."

Full Article

A robot imitating animals A Robotic Leg, Born Without Prior Knowledge, Learns to Walk
USC Viterbi School of Engineering
Greta Harrison; Amy Blumenthal
March 11, 2019


Researchers at the University of Southern California (USC) have developed a bio-inspired algorithm that enabled a robotic limb to learn a new walking task by itself after only five minutes of unstructured activity, and then adapt to other tasks without additional programming. This breakthrough is similar to the natural learning that happens in babies, as the robotic limb was first allowed to understand its environment in a process of free play, known as "motor babbling." The random movements of the leg that take place during motor babbling allow the robot to build an internal map of its limb and its interactions with the environment. The robots use their unique experience to develop the gait pattern that works well enough for them, producing robots with personalized movements. Said USC researcher Francisco J. Valero-Cuevas, "Because our robots can learn habits, they can learn your habits, and mimic your movement style for the tasks you need in everyday life—even as you learn a new task, or grow stronger or weaker."

Full Article
Researchers Use Algorithm From Netflix Challenge to Speed Up Biological Imaging
Optical Society of America
March 14, 2019


Scientists at the Ecole Normale Superieure in France have repurposed an algorithm developed for Netflix's 2009 movie preference prediction challenge for high-speed acquisition of classical Raman spectroscopy biological-tissue images. The researchers demonstrated imaging speeds of a few tens of seconds for an image that would usually take minutes to obtain, and they think sub-second speeds could be realized in the future. Said Ecole Normale Superieure's Hilton de Aguiar, "We combined compressive imaging with fast computer algorithms that provide the kind of images clinicians use to diagnose patients, but rapidly and without laborious manual post-processing." The researchers replaced costly, slow cameras used in conventional setups with a spatial light modulator, which selects groups of wavelengths identified by a single-pixel detector, compressing images as they are captured. This allowed the team to use a portion of the data typically required for non-invasive Raman spectroscopy, and employ the Netflix algorithm to fill in the missing information.

Full Article

Gary Grider and Brad Settlemyer discussing the new software product, DeltaFS Handling Trillions of Supercomputer Files Just Got Simpler
Los Alamos National Laboratory News
March 14, 2019


Los Alamos National Laboratory (LANL) and Carnegie Mellon University (CMU) researchers designed a new distributed file system for high-performance computing, enabling unprecedented performance for generating, updating, and managing vast troves of files. LANL's Brad Settlemyer said, "We designed DeltaFS to enable the creation of trillions of files. Such a tool aids researchers in solving classical problems in high-performance computing, such as particle trajectory tracking or vortex detection." DeltaFS, released on GitHub, builds a file system that does not need specialized hardware, and is precisely customized to assist in new discoveries when using a high-performance computing platform. CMU's George Amvrosiadis said DeltaFS has to scale across thousands of servers without committing a segment of them to the file system, which "frees administrators from having to decide how to allocate resources for the file system, which will become a necessity when exascale machines become a reality."

Full Article
Intel Offers AI Breakthrough in Quantum Computing
ZDNet
Tiernan Ray
March 14, 2019


Researchers at Intel and Hebrew University in Israel have defined an important proof for deep learning, proposing a path forward for computing commonly intractable problems in quantum physics. The proof describes deep learning's unmatched ability to model quantum-computing computations, with the data redundancy occurring in convolutional neural networks (CNNs) and recurrent neural networks (RNNs) critical. Both networks' structure entails an essential "reuse" of information, via multilayer stacking, to achieve more efficient "representation" of things in computing terms. The researchers said, "Our work quantifies the power of deep learning for highly entangled wave function representations, theoretically motivating a shift towards the employment of state-of-the-art deep learning architectures in many-body physics research." The formal proofs of the efficiency of "convolutional arithmetic circuits" and "recurrent arithmetic circuits" form a proof that deep learning strategies can meet quantum entanglement challenges more efficiently.

Full Article

A streetlight and the Huawei logo in the background Top U.S. Universities Shun Cash From Huawei Under Trump Pressure
Bloomberg
Todd Shields
March 12, 2019


Pressure from Congress and the Trump administration has forced top U.S. universities to refuse research funding from China's Huawei Technologies over national security issues. Princeton University, Stanford University, Ohio State University, and the University of California, Berkeley (UC Berkeley) announced they are cutting or reducing ties to the Chinese telecommunications equipment maker. The U.S. Department of Education estimated that Huawei gave $10.6 million in gifts and contracts to nine U.S. schools for technology and communications programs between 2012 and 2018. Government officials allege Huawei may pose a threat of espionage, partly because Chinese law mandates the company must answer to Beijing's security agencies. UC Berkeley's Randy Katz said submitting to political pressure goes against U.S. colleges and universities' ethos of open knowledge exchange, "But we take very seriously the federal indictments about the company's business practices."

Full Article
Locking More Than the Doors as Cars Become Computers on Wheels
The New York Times
Jim Motavalli
March 7, 2019


While cars have been targeted by hackers for at least a decade, today's vehicles carry far more electronic equipment, and autonomous driving technology that relies on sensors, cameras, and radar is fast approaching. Concern that cars could be hacked by criminals, terrorists, or even rogue governments has prompted a new era of security efforts within the auto industry. The average car has over 150 million lines of computer code, and that complexity creates a real risk of cyberattack, according to a 2018 University of Michigan report. However, hacking into a car's driver controls requires a lot of knowledge and effort, which is one reason why there has not been a significant number of major attacks. In addition, there is not much incentive for bad actors, because the identity information stored onboard vehicles is limited. The auto industry can learn a lot about cybersecurity from the aerospace industry; said KPMG's Jono Anderson, "Maybe it’s possible to hack the entertainment system in a plane and get free movies, but it’s virtually impossible to hack the actual communications."

Full Article
AI Academics Under Pressure to Do Commercial Research
Financial Times
Madhumita Murgia
March 13, 2019


Artificial intelligence (AI) scientists who sought to pursue purely academic research are finding themselves increasingly pressured to develop commercial applications. Joelle Pineau with McGill University in Canada said, "[Tech companies]...understand that to commercialize AI technology over a long term, you need to invest in [research and development.]" Facebook, Google, and other tech giants are building stables of AI talent poached from academia with enticing compensation offers, or by wholesale acquisition of academic startups like DeepMind. Independent researchers warn these practices threaten to deplete the ranks of teaching professors, and concentrate nascent AI research within a handful of companies. Some tech firms are trying to allay these concerns by having academics split their time between institutional and commercial work, and many who accept this arrangement envision their commercial research as open and collaborative as their university projects.

Full Article
*May Require Paid Registration

Rearview shot of a group of students standing in a line on graduation day Diversity in Tech: How One VC Firm Is Partnering With Historically Black Colleges to Address the Pipeline Issue
Forbes
Ruth Umoh
March 13, 2019


Unusual Ventures, a seed-stage investment firm, this summer will launch its Unusual Interns program, designed to increase the number of black students working in technical internships by connecting students from historically black colleges and universities (HBCUs) with high-growth Silicon Valley startups. Currently, black Americans represent less than 5% of the employees at most tech companies, while at least 50% of the workforce at those companies is white, according to the National Urban League's State of Black America 2018 report. Tech's lack of diversity is often called a pipeline problem, but the homogeneity of the industry can also be attributed to Silicon Valley's passive attitude toward diverse recruitment. As the first step of the Unusual Interns program, 10 computer science students from a group of handpicked HBCUs will participate in 10- to 12-week internships. The students will be matched with tech companies, and have the opportunity to gain full-time employment after completing the internship.

Full Article

A human model on the computer Computer Kidney Could Provide Safer Tests for New Medications
Waterloo News
March 11, 2019


Anita Layton at the University of Waterloo in Canada has led the development of the first computational model of the human kidney to enable scientists to more deeply explore kidney function and how new medications may work, without resorting to invasive procedures. Layton said the new model replaces earlier models based on rodent kidneys. She explained, "Certain drugs are developed to target the kidney while others have unintended effects on the kidney, and computer modeling allows us to make long-term projections of potential impacts, which could increase patient safety." The model's development involved incorporating anatomic and hemodynamic data from the human kidney into the published computational model of a rat kidney. Layton's team adjusted key transporter data to make the predicted urine output consistent with known human values, and identified a set of compatible transport parameters that returned model predictions consistent with human urine and lithium clearance data.

Full Article
Asian Deans' Forum 2019
 
2019 Stanford University Frontier of AI-Assisted Care Scientific Symposium
 

Association for Computing Machinery

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