MS Programs
 
Welcome to the April 9, 2021 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."

The critical care unit in Israel’s Sheba Medical Center The People in This Medical Research Are Fake. The Innovations Are Real.
The Wall Street Journal
Dov Lieber
April 6, 2021


Medical researchers and data scientists are generating artificial patients algorithmically from real-life datasets to accelerate the development of innovations with real-world applications. Allan Tucker at the U.K.'s Brunel University London said, "The key advantage that synthetic data offers for healthcare is a large reduction in privacy risks that have bugged numerous projects [and] to open up healthcare data for the research and development of new technologies." The Covid-19 pandemic fueled demand for synthetic-data solutions as medical providers and researchers raced to understand the pathogen and develop treatments. Israel is a major testbed, using the MDClone startup's platform for creating synthetic data from medical records, for example. Not all synthetic-data research relies on real-life medical records: U.S. nonprofit Mitre's open source Synthea tool can generate populations of artificial patients from scratch, using publicly available data sources.

Full Article
*May Require Paid Registration

A swarm of desert locusts. As Locusts Swarmed East Africa, This Tech Helped Squash Them
The New York Times
Rachel Nuwer
April 8, 2021


A 2020 locust plague in East Africa was mitigated by technology-driven countermeasures, spearheaded by Keith Cressman at the United Nations Food and Agriculture Organization (FAO) and Pennsylvania State University's David Hughes. The partners created the eLocust3m application for collecting dependable and detailed locust data, using a mobile tracking tool Hughes previously created with the FAO as a template. The smartphone-enabled app presents photos of locusts at different developmental stages, so users can diagnose what they observe in the field; global positioning system coordinates are automatically recorded, and algorithms double-check photos submitted with each entry. Wildlife-focused security and logistics company 51 Degrees repurposed anti-poaching aerial surveys to find and kill locust swarms, using a customized version of the EarthRanger program from philanthropic firm Vulcan. The program integrated data from the eLocust programs and the computer loggers on pesticide sprayers.

Full Article
*May Require Paid Registration

A macaque with a brain implant playing Pong with its neural activity. Monkey Equipped with Elon Musk's Neuralink Device Plays Pong with Its Brain
TechCrunch
Darrell Etherington
April 8, 2021


Elon Musk's Neuralink company released a blog post and video displaying a monkey playing the game of Pong by thought, via the firm's sensor hardware and brain implant. The technology recorded a baseline of activity from a macaque playing a game onscreen where it had to move a token to different squares using a joystick. Neuralink then employed machine learning to predict where the monkey would be moving the controller, and was eventually able to anticipate this accurately before the move was made. Researchers eliminated the paddle entirely and repurposed Pong to enable the animal to control in-game action entirely with its mind. Neuralink envisions the technology helping paralyzed patients, and Musk suggested future versions would facilitate communication between Neuralinks in different parts of a patient's body.

Full Article
Computational Language Models Can Further Environmental Degradation, Language Bias
The Daily of the University of Washington
Kate Companion
April 8, 2021


Natural language processing (NLP) technology used for modeling and predicting language patterns can promote linguistic bias and damage the environment, according to University of Washington (UW) researchers. NLP utilizes large-scale pattern recognition to generate predictive language models, and UW's Emily M. Bender said such models manifest in predictive text and autocorrect features. Although the algorithms are trained on vast datasets from the Internet to recognize patterns, the Internet's scale does not ensure diversity; when considering people who lack or shun Internet access and the weeding out of certain words, Bender said the datasets can exclude underrepresented voices. The UW study also determined that biases and abusive language patterns that perpetuate racism, sexism, or other harmful perspectives can be picked up in the training data.

Full Article

Illustration of amplifiers with female and male symbols respectively. Not OK, Computer: Music Streaming's Diversity Problem
Financial Times
Gillian Tett
April 7, 2021


Female artists represented just 25% of the music listened to by users of a streaming service, according to researchers at the Netherlands' Utrecht University and Spain's Universitat Pompeu Fabra. The researchers said their analysis of publicly available listening records of 330,000 users of a single streaming service revealed that "on average, the first recommended track was by a man, along with the next six. Users had to wait until song seven or eight to hear one by a woman." Streaming service algorithms recommend music based on what has been listened to before, which creates a vicious feedback cycle if it already offers more music by men. The researchers simulated and modified the algorithm to elevate the rankings of women while lowering those of men, which created a new feedback loop. The algorithm recommended female artists earlier, increasing user awareness so the program would recommend female artists more often when that content was selected.

Full Article
*May Require Paid Registration

Illustration of a tech-based digital fingerprint. Scientists Harness Chaos to Protect Devices From Hackers
Ohio State News
Jeff Grabmeier
April 7, 2021


A new version of physically unclonable functions (PUFs) developed by a research team led by The Ohio State University (OSU) could prevent even the most sophisticated hackers from accessing electronic devices. The team's method leverages small manufacturing variations in computer chips—sometimes visible only at the atomic level—to create PUFs that could be used in secure ID cards, supply chain tracking, and authentication applications. Using these tiny variations, researchers can create unique sequences of 0s and 1s, dubbed "secrets." The researchers calculated one of their PUFs could create 1077 secrets. OSU's Daniel Gauthier said that if a hacker could guess one secret every microsecond, or 1 million secrets per second, it would take about 20 billion years to guess all the secrets in one microchip. Verilock's Jim Northup said, "This novel approach to a strong PUF could prove to be virtually un-hackable."

Full Article

A robot disinfects an airplane interior. Swiss Robots Use UV Light to Zap Viruses Aboard Passenger Planes
Reuters
Arnd Wiegmann; John Miller
April 1, 2021


A robot developed by Swiss startup UVeya is using ultraviolet (UV) light to kill viruses aboard Swiss passenger planes in a test being conducted with airport services company Dnata in the United Arab Emirates. The UVeya team has built three prototypes of the robot, one of which co-founder Jodoc Elmiger demonstrated inside a Helvetic Airways jet at Switzerland's Zurich Airport. One robot can disinfect a single-aisle plane in 13 minutes. Dnata's Lukas Gyger said, "We were looking for a sustainable, and also environmentally friendly, solution to cope with [requests to ensure air travelers do not get sick]."

Full Article

Rice University’s Anshumali Shrivasta, who led a team to explore how to implement deep learning technology without specialized hardware. Rice, Intel Optimize AI Training for Commodity Hardware
Rice University News
Jade Boyd
April 7, 2021


Rice University computer scientists and collaborators at Intel have demonstrated artificial intelligence (AI) software that operates on commodity processors and trains deep neural networks (DNNs) significantly faster than graphics processing unit (GPU)-based platforms. Rice's Anshumali Shrivastava said the cost of DNN training is the biggest bottleneck in AI, and the team's sub-linear deep learning engine (SLIDE) overcomes it by running on commodity central processing units (CPUs), and by approaching DNN training as a search problem to be addressed with hash tables. The latest research considered the impact of vectorization and memory optimization accelerators on CPUs. Shrivastava said, "We leveraged those innovations to take SLIDE even further, showing that if you aren't fixated on matrix multiplications, you can leverage the power in modern CPUs and train AI models four to 15 times faster than the best specialized hardware alternative."

Full Article
Separating Fact From Fiction: UA Professors Study Online 'Pseudo-Reviews'
University of Akron
April 7, 2021


A study by researchers at the University of Akron (UA) and the University of Georgia looks at the impact of pseudo-reviews on online platforms found that pseudo-reviews appear like authentic reviews in telling a story about product use, but often use humor to mock aspects of the product. They found that pseudo-reviews on their own have little impact on consumers' attitudes about a product, but when the number of pseudo-reviews and authentic reviews is the same, consumers' attitudes and purchase intentions are negatively affected. The researchers also noted that consumers could abandon platforms that feature too many pseudo-reviews. UA's Alexa K. Fox said, "While pseudo-reviews may not appear problematic on the surface due to their humorous nature, indeed, they have the potential to be damaging to consumers’ decision-making processes."

Full Article
Quantum Technology Emerges From the Lab to Spark a Mini Start-Up Boom
The Washington Post
Jeanne Whalen
April 7, 2021


The University of Chicago, the University of Illinois, and Argonne National Laboratory have rolled out the first program in the U.S. to support quantum-tech start-ups. The University of Chicago's David Awschalom said, "We are at the birth of a new field of technology. It's like we're at the point where the transistor is being invented. People are beginning to think about systems, software, applications." The Duality accelerator program, based at the University of Chicago's Booth School of Business, will invest $20 million over the next 10 years to assist as many as 10 quantum start-ups annually. The start-ups will benefit from $50,000 grants, access to lab and office space, and faculty mentoring. The University of Chicago's Fred Chong noted that "there is very little on the software side" of quantum computing, and the challenge for developers is to develop programs that work with today’s "imperfect quantum machines."

Full Article

A series of slides showing an organoid grown in a microfluidic bioreactor. Human Brain Organoids Grown in Cheap 3D-Printed Bioreactor
New Scientist
Christa Lesté-Lasserre
April 6, 2021


A human brain organoid was cultured in a week in a three-dimensionally (3D)-printed microfluidic bioreactor developed by researchers at the Indian Institute of Technology Madras (IIT Madras) and the Massachusetts Institute of Technology. The ultra-cheap bioreactor consists of a $5 washable, reusable microchip containing wells where brain tissue grows. Nutrient fluids are automatically pumped through these channels, feeding the tissue. The chip can be 3D-printed using the same kind of biocompatible resin used in dental surgery, while the bioreactors control the flow of nutrient fluid and purge waste through tubes in an enclosed incubator. IIT Madras' Ikram Khan said, "My goal is to see this technology reach people throughout the world who need access to it for their healthcare needs."

Full Article

Illustrations showing how to study, design, and manufacture composite materials in three dimensions. ML Tool Converts 2D Material Images Into 3D Structures
Imperial College London (U.K.)
Caroline Brogan
April 5, 2021


A new machine learning algorithm developed by researchers at the U.K.'s Imperial College London (ICL) can render two-dimensional (2D) images of composite materials into three-dimensional (3D) structures. ICL's Steve Kench said, "Our algorithm allows researchers to take their 2D image data and generate 3D structures with all the same properties, which allows them to perform more realistic simulations." The tool uses deep convolutional generative adversarial networks to learn the appearance of 2D composite cross-sections, and expands them so their “phases” (the different components of the composite material) can be studied in 3D space. The researchers found this method to be less expensive and faster than generating 3D computer representations from physical 3D objects, and able to identify different phases more clearly. ICL's Sam Cooper said, "We hope that our new machine learning tool will empower the materials design community by getting rid of the dependence on expensive 3D imaging machines in many scenarios."

Full Article

A game on the Emojify website demonstrating AI emotion recognition technology. Scientists Create Online Games to Show Risks of AI Emotion Recognition
The Guardian (U.K.)
Nicola Davis
April 4, 2021


Scientists at the U.K.'s University of Cambridge have created emojify.info, a website where the public can test emotion recognition systems via online games, using their own computer cameras. One game has players make faces to fake emotions in an attempt to fool the systems; another challenges the technology to interpret facial expressions contextually. Cambridge's Alexa Hagerty cited a lack of public awareness of how widespread the technology is, adding that its potential benefits should be weighed against concerns about accuracy, racial bias, and suitability. Hagerty said although the technology's developers claim these systems can read emotions, in reality they read facial movements and combine them with existing assumptions that these movements embody emotions (as in, a smile means one is happy). The researchers said their goal is to raise awareness of the technology and to encourage dialogue about its use.

Full Article
Master of Science in Biomedical Informatics
 
ACM Career and Job Center
 

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]