Welcome to the December 16, 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."
|
|
Vaccinated? Show Us Your App
The New York Times Natasha Singer December 13, 2020
Recipients of the Covid-19 vaccine may eventually receive digital health credentials in order to travel, go to work or school, visit entertainment venues, and more. Supporters of digital vaccine credentials say they could help control the virus, reopen the economy, and give consumers peace of mind, but there are concerns these digital passes would make it difficult for people with limited access to vaccines or online verification tools to work or engage in other activities. Tech companies working on Covid-19 health pass apps, however, contend they are more reliable than paper health documents, which could be forged, and that their systems have been designed to mitigate privacy risks. The airline industry's CommonPass, as well as similar apps from IBM and Clear, generate unique QR codes users can display to confirm their health status, and specific details are not shared.
|
An LED That Can Be Integrated Directly Into Computer Chips
MIT News Daniel Ackerman December 14,
2020
A silicon-based LED that can be integrated directly onto a computer chip has been fabricated by Massachusetts Institute of Technology (MIT) researchers and American semiconductor foundry GLOBALFOUNDRIES. The LED is sufficiently bright to enable state-of-the-art sensor and communication technologies, and features specially engineered junctions to augment brightness. The device generates enough light to send signals through 5 meters (16.4 feet) of fiber-optic cable. MIT's Jin Xue said the silicon LED switches on and off faster than expected, and has been used to transmit signals at frequencies up to 250 megahertz. The University of California at Berkeley's Ming Wu said the LED "allows silicon integrated circuits to communicate with one another directly with light instead of electric wires. This is somewhat surprising as silicon has an indirect bandgap and does not normally emit light."
|
Video Referee in the Spotlight
Technical University of Munich December
14, 2020
Researchers at Germany's Technical University of Munich (TUM) analyzed 129 English football games via Twitter with artificial intelligence to gauge the impact of decisions made by video assistant referees (VARs) on fans??? mood. The team text-mined 643,251 English-language tweets concerning 94 VAR-related incidents; 9.1% (more than 58,000) of the tweets were directly related to the VAR. TUM's Otto Kolbinger said analysis of the tweets using a new text classification algorithm concluded that the average sentiment of the 58,000+ tweets relating to VAR decisions was substantially lower than that of other tweets: 76.24% were negative, 12.33% were positive, and 11.43% were neutral. Said Kolbinger, "The football associations should attempt to communicate all VAR decisions with greater transparency."
|
Scientists Digitally Rebuild Dinosaur's Brain, Make Some Surprising Discoveries
CNN Amy Woodyatt December 14, 2020
The reconstruction of a dinosaur's brain through advanced imaging and three-dimensional modeling by researchers at the U.K.'s University of Bristol suggests the sauropod may have been bipedal and occasionally carnivorous, unlike its later relatives. The software used to three-dimensionally model the brain and inner ear configuration of Thecodontosaurus from computer tomography scans of its braincase fossil revealed large floccular lobes, indicating bipedalism and agility. Bristol's Antonio Ballell said, "Our analysis showed parts of the brain associated with keeping the head stable and eyes and gaze steady during movement were well-developed. This could also mean Thecodontosaurus could occasionally catch prey."
|
An AI Used Facebook Data to Predict Mental Illness
Wired Grace Huckins December 14,
2020
Researchers at New York's Feinstein Institutes for Medical Research used an artificial intelligence (AI) algorithm to scan 223 volunteers' Facebook messages in order to predict psychiatric diagnoses, including messages sent up to 18 months before a user received an official diagnosis. The algorithm mined the messages and posted photos to predict whether each participant had a mood disorder, a schizophrenia spectrum disorder, or no mental health issues. The results indicated the use of swear words signaled mental illness overall, while using perception words (see, feel, hear) and words related to negative emotions suggested schizophrenia, and more bluish colors in photos were associated with mood disorders. Michael Birnbaum of the Feinstein Institutes for Medical Research said this type of AI tool could make an enormous difference in the treatment of psychiatric illnesses.
|
Cryptographers Unveil Breakthrough in Achieving Indistinguishability Obfuscation
Forbes Tony Bradly December 15, 2020
Nearly 20 years after indistinguishable obfuscation (iO) was theorized, researchers at the University of California, Los Angeles and the University of Washington have shown that iO???encrypting a computer program so the code is unintelligible but all the functionality is retained???can be achieved. The solution combines Symmetric external Diffie-Hellman on pairing groups, Learning with Errors, Learning Parity with Noise (LPN) over large fields, and a simple Boolean Pseudo-Random Generator (PRG), each of which is computationally expensive to solve. The researchers leveraged LPN over fields and simple Boolean PRG to develop a structured-seed PRG, which then allowed them to draw on previous research to achieve iO. NTT Research's Tatsuaki Okamoto said, "Because iO implies many strong cryptographic functionalities that are considered hard to realize without iO, what this means is that our cryptographic world is now richer and more powerful."
|
Wearable Device May Help Shine Light on Health Benefits of Outdoor Lighting
Penn State News Matt Swayne December 14, 2020
A wearable device developed by Pennsylvania State University Berks (Penn State Berks) researchers can distinguish between indoor and outdoor lighting. The device features sensors that collect data on light sources like wavelength and frequency data. An artificial neural network was used to analyze 3,640 indoor and 1,368 outdoor data samples to determine whether the device was placed indoors or outdoors. The device, created using off-the-shelf components for about $70, could help researchers better understand the benefits of outdoor light and lead to wearables that could encourage users to spend more time outdoors. Penn State Berks' Matthew Rhudy said the device could be made less expensive "if we could maybe work with an electronics company and mass-produce it."
|
Free Tool Adds Layer of Security for Software Supply Chain
New York University Tandon School of Engineering December
14, 2020
An open source tool developed by New York University (NYU) Tandon School of Engineering researchers could safeguard the software supply chain against cyberattacks. The tool, called in-toto, works like blockchain for the software development process to ensure all steps performed on a piece of software throughout its design and development lifecycle can be trusted, by providing information crucial to security. The researchers said their experiments show in-toto could have prevented at least 83% of real-life software supply chain compromises affecting millions of users last year. NYU Tandon's Santiago Torres-Arias said, "As it moves from development to testing to packaging, and finally to distribution, a piece of software passes through a number of hands. By requiring that each step in this chain conforms to the layout specified by the developer, it confirms to the end-user that the product has not been altered for malicious purposes, such as by adding backdoors in the source code."
|
App Tracks Black Rhinos Through Their Footprints
Scientific American Helen Santoro
The footprint-identification technique (FIT) developed by the WildTrack conservation organization uses software to track black rhinoceros' movements via smartphone-recorded footprints, in order to protect the animals from poachers. Scientists first collect footprint images with a smartphone application and upload them to a global database; FIT software analysis then can identify the specific rhino and tell its age and gender with up to 99% accuracy. Researchers also can use the technique to calculate the number of black rhinos in an area and monitor their movements. WildTrack's Sky Alibhai and Zoe Jewell are training wildlife conservationists, land managers, local guides, and antipoaching agents to utilize FIT in Namibia. Jewell said, "FIT is scalable, and we would always welcome the opportunity to use it with endangered species that are in larger populations."
|
Academics Turn RAM Into Wi-Fi Cards to Steal Data From Air-Gapped Systems
ZDNET Catalin Cimpanu December 11, 2020
The AIR-FI technique developed by academics at Israel's Ben-Gurion University of the Negev (BGU) transforms a random-access memory (RAM) card into an impromptu Wi-Fi card to transmit data from within a non-networked air-gapped computer. AIR-FI's effectiveness hinges on the fact that electronic components emit electromagnetic waves as electric current passes through them. BGU's Mordechai Guri said malware planted on an air-gapped system could rig the current inside the RAM card to produce electromagnetic waves in the 2,400-gigahertz frequency range of Wi-Fi. Guri demonstrated how perfectly timed read-write operations to a computer's RAM card can induce its memory bus to emit waves consistent with a weak Wi-Fi signal, which can be received by any nearby equipment with a Wi-Fi antenna. He said AIR-FI can be launched from an ordinary user-space process and operate across on any operating system.
|
UGA Engineers Develop Soft Robotic Gripper
UGA Today Mike Wooten December 16, 2020
Researchers at the University of Georgia (UGA) College of Engineering have developed a soft robotic gripper inspired by pole beans, which can grasp objects as small as 1 millimeter in diameter. The pole bean uses touch-sensitive shoots to wrap around supports and grow upwards. UGA's Mable Fok said, "Our robot's twining action only requires a single pneumatic control ??? eliminating the need for complex coordination between multiple pneumatic controls.??? Fok added that the new gripper ???needs only a small operational space," so it works well in confined spaces. The device has an embedded fiber-optic sensor that supplies real-time feedback of the twining angle, the target's physical parameters, and any outside disruptions that could loosen the target. In sum, Fok said, ???Having a simpler design and control is definitely an advantage."
|
Predicting British Railway Delays Using Artificial Intelligence
University of Illinois at Urbana-Champaign December 10, 2020
Researchers at the University of Illinois at Urbana-Champaign (UIUC) and the Zhejiang University-University of Illinois at Urbana-Champaign Institute used an artificial intelligence (AI) model and data from British Railway, which serves 1.7 billion passengers annually, to predict railway network delays. They predicted delays in a portion of the British rail network by applying the Spatial-Temporal Graph Convolutional Network model, which UIUC's Huy Tran said outperforms other statistical models "for forecasting delays up to 60 minutes in the future." The researchers developed a new formulation that can approximate the leg of the trip where the delay occurred. Said Tran, "We learned that using our formulation with this class of AI models can work well with real-world networks to predict behaviors." However, Tran said, while the model aims to predict a delay, it does not provide any insight into the reason for the delay.
|
Accurate Neural Network Computer Vision Without the 'Black Box'
Duke Today December 15, 2020
Duke University researchers have formulated a method for addressing the "black box" problem, the challenge of trusting computer-vision predictions when their operational mechanisms are unknown. Duke's Cynthia Rudin, Zhi Chen, and Yijie Bei modified the prediction process' underlying reasoning in order to better resolve deep neural network problems or determine networks' trustworthiness. Rather than trying to account for a network's decision-making on a post hoc basis, the technique trains the network to reveal its work by expressing conceptual comprehensions along the way. This method exposes how much the network considers different concepts to help understand what it visualizes. The Duke team learned a slight network adjustment enables accurate identification of object and scenes and images comparative to the original network, with significant gains in interpretability.
|
|