Master's Degree in Electrical and Computer Engineering
Welcome to the April 25, 2022, 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."

Robots are socialize with nursing home residents, telling them jokes and leading their yoga classes. Can Robots Save Nursing Homes?
The New York Times
John Leland
April 24, 2022

As more nursing home and assisted living facility staff leave the profession, some facilities are turning to robots to fill in the gaps. A $2-million contribution from the Minnesota Department of Human Services could enable 16 robots programmed by University of Minnesota Duluth’s Arshia Khan to be deployed to eight nursing homes later this year, if approved by the university's institutional review board. NAO robots will lead residents in yoga, tai chi, and strength-training classes, while Pepper robots will socialize with and entertain residents. Concerns about nursing homes' use of robots include a decline in human-to-human contact, whether families should have access to monitoring data, and how robots are presented in terms of gender and ethnicity.

Full Article
*May Require Paid Registration
Time to Get Social: Tracking Animals with Deep Learning
EPFL (Switzerland)
April 22, 2022

Researchers at Switzerland's École polytechnique fédérale de Lausanne (EPFL) enhanced their DeepLabCut software to enable high-performance tracking of animals in videos. The open source software uses deep learning to teach computers "pose estimation" without requiring physical or virtual markers on the animals. The update allows DeepLabCut to track the movements of social animals, like mice or fish, which can confuse the computer because they look too similar or obscure each other. With the help of researchers at the Massachusetts Institute of Technology and Harvard University, the EPFL team developed four datasets for benchmarking multi-animal pose estimation networks. The researchers also developed a multi-task neural network that can predict keypoints, limbs, and animal identity from single frames, and an assembly algorithm to help identify animals with varied body shapes.

Full Article
Cleaning Up Online Bots' Act—and Speech
UC San Diego News Center
Ioana Patringenaru
April 21, 2022

University of California, San Diego (UCSD) computer scientists have developed algorithms to keep offensive speech from being generated by online chatbots. The researchers inputted toxic prompts to a pre-trained "evil" language model to induce production of toxic content, then trained it to predict the probability of the resulting content's toxicity. They taught a "good model" to avoid all high-ranking content from the evil model, and found it performed as well as state-of-the-art methods, cleaning up speech by as much as 23%. "Our lab has expertise in algorithmic language, in natural language processing, and in algorithmic de-biasing," explained UCSD's Julian McAuley. "This problem and our solution lie at the intersection of all these topics."

Full Article

The compound eye of an insect, such as this hornet mimic hoverfly, excels at distinguishing signal from noise. AIs Spot Drones with Help from Fly Eye
Scientific American
Monique Brouillette
April 20, 2022

Researchers at the University of South Australia, defense firm Midspar Systems, and Australia's Flinders University have developed an artificial intelligence (AI) algorithm for visual drone detection. The researchers reverse-engineered the visual system of the hoverfly—whose compound eyes can separate relevant information from noise—to develop a tool that filters out noisy data. They fed the algorithm spectrograms based on acoustic data from outdoors as drones flew by. The algorithm was able to amplify data related to the frequencies emitted by drones, while reducing background noise from other sources. The researchers found that it could identify drones up to 50% farther away than conventional AI systems.

Full Article
Your iOS App May Still Be Covertly Tracking You, Despite What Apple Says
Ars Technica
Dan Goodin
April 18, 2022

Researchers at the U.K.’s University of Oxford found that iOS apps can still track users despite Apple's App Tracking Transparency (ATT) policy prohibiting app developers from tracking user activity across apps without explicit permission. The researchers found nine iOS apps using server-side code to generate mutual user identifiers that can be used for cross-app tracking by a subsidiary of China's Alibaba. They also compared 1,685 apps before and after the implementation of ATT and found that the number of tracking libraries used by the apps held steady. Although 25% of the apps said they collected no user data, 80% had at least one tracker library. The researchers also found that over half of the apps using SKAdNetwork, Google Firebase Analytics, and Google Crashlytics, and 47% of those using Facebook SDK, did not disclose having access to user data.

Full Article

Stay-at-home dad Marc Perry says the strategy employed by bots on ‘Words With Friends’ is terrible. Friend or Bot? Phony Gamers Leave Players Feeling Betrayed
The Wall Street Journal
Sarah E. Needleman
April 19, 2022

Zynga Inc.'s word game "Words with Friends" became more popular as people sought connections during the pandemic, but many players were upset to learn they have been playing against bots the company introduced in 2019 to ensure players find suitable opponents. The app-analytics firm Sensor Tower Inc. found that 1 out of 10 mobile games used bots at the beginning of 2016, a figure that has risen to 7 of the 10 most-played mobile games. "Words with Friends" co-creator Paul Bettner said players should get used to playing against bots because "We're heading into a world where the definition of a friend is going to include artificial intelligence."

Full Article
*May Require Paid Registration
3D-Printing Approach Melds Solids, Liquids
CU Boulder Today
Daniel Strain
April 18, 2022

University of Colorado Boulder (CU Boulder) engineers have created three-dimensional (3D) printing technology that combines solid and liquid components. The researchers designed computer models to explore the physics of printing different materials contiguously, and CU Boulder's Robert MacCurdy said choosing a liquid that is denser than a solid can prevent mixing. In doing so, the team established a set of rules to help keep the droplets of solid materials from mixing into the liquid materials. Said MacCurdy, “If you have a printer that can use multiple kinds of materials, you can combine them in new ways and create a much broader range of mechanical properties.”

Full Article
Reducing COVID-19 Patients' Breathing Efforts Could Be Key to Success of Non-Invasive Respiratory Support
University of Warwick (U.K.)
April 21, 2022

A team of U.S., U.K., and Irish researchers has used computational modeling to demonstrate that non-invasive respiratory support is more likely to be successful if it relies on significantly reducing patients' efforts to breath. Researchers at the U.K.'s University of Warwick created computer simulations of 120 COVID-19 patients to measure the internal mechanics generated by different types of non-invasive support at different levels of breathing intensity. They found that while non-invasive measures improved oxygenation, stresses and strains within the lung could be elevated to potentially dangerous levels without any reduction in breathing effort. Said Dr. Luigi Camporata, “These results provide urgently needed evidence to help clinicians manage and optimize the treatment of COVID-19 patients in a way that averts additional and preventable lung injury.”

Full Article

The Stanford research team built this prototype lidar system that captures megapixel-resolution depth maps using a commercially available digital camera. LiDAR System Promises 3D Vision for Cameras, Cars, Bots
IEEE Spectrum
Mark Harris
April 21, 2022

Stanford University researchers, working with a colleague at Sweden’s Chalmers University of Technology, have developed a novel light detection and ranging (LiDAR) system that taps the piezoelectric effect to capture three-dimensional data at potentially vastly less cost than current systems. The researchers coated thin-film lithium niobate with transparent electrodes in order to excite the crystal, generating an acoustic standing wave that modulates the intensity of laser light passing through it. They were able to execute a modulated time-of-flight calculation that captured distance data to objects in the scene, using less than a watt of power. Coupling the opto-acoustic modulator with a standard four-megapixel digital camera yielded a relatively high-resolution depth map of several metal targets, locating them to within a few centimeters.

Full Article

The Susquehanna River. University of Georgia researchers created a model to help identify locations best suited for conservation. Improving Georgia Land Conservation Through Algorithms
UGA Today
Ian Bennett
April 20, 2022

Researchers at the University of Georgia (UGA) have developed an algorithm for assessing a tract of land's conservation value by factoring in variables excluded from other models. "Conservation efforts almost always have a fixed budget," said UGA's Puneet Dwivedi. "Our model allows users to find the best parcels of land for conservation and maximize impact over time." The model favors higher connectivity among parcels of land by imposing penalties on adjacent developments. "As we added boundary penalties, larger parcels were selected with higher connectivity," explained UGA's Fabio Jose Benez-Secanho. "Larger protected areas benefit wildlife habitat, biodiversity, and other ecological functions. The financial trade-offs of selecting larger parcels is relatively low when compared with the extra benefits they provide."

Full Article
Scientists Use ML to Identify Antibiotic-Resistant Bacteria
University of Nottingham (U.K.)
April 20, 2022

Experts at the U.K.'s University of Nottingham integrated DNA sequencing and machine learning (ML) to determine the site and extent of antibiotic-resistant bacteria's transmission between humans, animals, and the environment. In examining a commercial poultry farm in China, the researchers compiled 154 samples from animals, carcasses, laborers, and their households and environments. They isolated E. coli bacteria from the samples, and classified distinct pathogens found at the farm through ML, whole genome sequencing, gene sharing networks, and mobile genetic components. The team uncovered a complete network of genes that correlated with antimicrobial resistance, shared between animals, farm workers, and the surrounding environment. "We cannot say at this stage where the bacteria originated from, we can only say we found it and it has been shared between animals and humans,” said Nottingham's Tania Dottorini.

Full Article
2022 CACM Website Survey Notice
ACM Survey on Research Publishing in AI and Machine Learning

Association for Computing Machinery

1601 Broadway, 10th Floor
New York, NY 10019-7434

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]