Welcome to the March 17, 2023, 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."

An illustration of Codon—the Python-based compiler. Python-Based Compiler Achieves Orders-of-Magnitude Speedups
MIT News
Rachel Gordon
March 14, 2023


The Codon compiler, available on GitHub, allows developers to create new domain-specific languages within Python while achieving the performance speeds of programming languages such as C and C++. Led by researchers at the Massachusetts Institute of Technology (MIT), the research team compiled about 10 commonly used genomics applications written in Python and observed speedups of 5 to 10 times over the original hand-optimized implementations. "Instead of needing to rewrite the program using a C-implemented library like NumPy or totally rewrite in a language like C, Codon can use the same Python implementation and give the same performance you'll get by rewriting in C,” said MIT's Saman Amarasinghe.

Full Article

Libyan youths take part in annual First Tech robotics competition in Benghazi. Can Young Robotics Fans Unite a Troubled Libya?
Al Jazeera
March 15, 2023


Young robotics enthusiasts such as 18-year-old Youssef Jira hope unify a divided Libya through high-tech modernization. Jira and others participated in the Libya Regional Championship for robotics, where about 20 teams of 12- to 18-year-olds competed with wheeled robots. Such programs help "open new horizons" for young Libyans, said event coordinator Mohammed Zayed. Participants "had to manage their relationships and work towards inclusion, unity, and peace." The competition’s intent is to also "prepare the workers of the future and make the country aware of the importance of technology and innovation," Zayed added.

Full Article

An infographic showing chair sensors and projector actuation. Researchers Create Office Chair with Spatial Augmented Reality
University of Waterloo Cheriton School of Computer Science (Canada)
March 13, 2023


An office chair that supports personal spatial augmented reality (AR) has been designed by researchers from Canada’s University of Waterloo Cheriton School of Computer Science. Working out of the university's Human-Computer Interaction (HCI) Laboratory, researchers incorporated capacitive sensors into the chair's armrests for touch input, while the back can detect tilt and rotation via an accelerometer; the seat and backrest can read seated postures and leg position with force-sensitive resistors. A depth camera tracks chair position and office surfaces while battery-powered computers process data and determine projector output, and a right-side servo-actuated pan/tilt head directs the projection of digital content.

Full Article
Teen Uses AI to Study Media Coverage of Homicides, Wins $175,000
The Washington Post
Theresa Vargas
March 15, 2023


The Regeneron Science Talent Search named Northern Virginia high school student Emily Ocasio as the recipient of its $175,000 second-place prize for using artificial intelligence to study the media's humanization of homicide victims. Ocasio used the GPT-3 language model, data from the U.S. Federal Bureau of Investigation, and 5,000 articles from the Boston Globe to review coverage of homicides in Massachusetts between 1976 and 1984. She evaluated each article for humanizing or impersonal language, then analyzed coverage across race, gender, age, and the junction of those identities to produce a "composite humanizing score." Ocasio said Black males under 18 years old were 30% less likely to be humanized by media coverage than white males, while Black women aged 18 to 29 were 23% less likely to be humanized as white women.

Full Article
*May Require Paid Registration

A playful illustration of an entity overseeing a self-driving lab. Self-Driven Laboratory, AlphaFlow, Speeds Chemical Discovery
NC State University News
Matt Shipman
March 15, 2023


Chemical engineering researchers at North Carolina State University (NC State) and the University at Buffalo developed the self-driven AlphaFlow laboratory to more efficiently synthesize semiconductor nanocrystals used in optical and photonic devices. AlphaFlow applies reinforcement learning to accelerate materials discovery in combination with automated microfluidic devices, said NC State's Milad Abolhasani, explaining that "AlphaFlow can conduct more experiments than 100 human chemists in the same period of time, while using less than 0.01% of the relevant chemicals.” Abolhasani added that AlphaFlow “effectively miniaturizes the experiments and performs the same laboratory operations that would require an entire wet chemistry lab in a suitcase-sized end-to-end experimental platform."

Full Article

A photo of the neuron-based living computer. 80,000 Mouse Brain Cells Used to Build a Living Computer
New Scientist
Karmela Padavic-Callaghan
March 16, 2023


An organic computer composed of about 80,000 neurons cultivated from repurposed mouse stem cells has been built by researchers at the University of Illinois at Urbana-Champaign. The researchers arrayed the neurons two-dimensionally, positioning them under an optical fiber onto a grid of electrodes so the cells could be activated with electricity and light. The electrodes also could detect when the neurons responded with their own electrical signals. The researchers trained the neural network on 10 different electricity/light sequences over 60 minutes, recording and processing neuron-produced electrical signals with a conventional computer chip. When re-exposed to the 10 sequences, the computer achieved a best F1 performance score of 0.98 on a scale of 0 to 1.

Full Article
Zipline Wants to Green Last-Mile Delivery with Drones
Bloomberg
Ira Boudway
March 15, 2023


Drone delivery startup Zipline announced its intended 2024 launch of an environmentally friendly last-mile service that will allow various retailers to deliver goods to customers' yards. The drone, or "zip," hovers roughly 300 feet overhead and lowers a tether with a container or "droid" at the end. The droid keeps steady with fans while avoiding obstacles during descent using cameras and other sensors. A bottom hatch opens to deposit the cargo upon touchdown, then the drone flies back to the retailer to recharge. Zipline calculates that its latest drones cut the carbon output of deliveries by nearly 98% versus gasoline-powered cars while emitting up to 30 times less carbon per mile than the average electric vehicle.

Full Article
*May Require Paid Registration

Illustration of the MIT system that enables makers to incorporate sensors directly into rotational mechanisms. 3D-Printed Revolving Devices Can Sense how They Are Moving
MIT News
Adam Zewe
March 16, 2023


MechSense, a system that allows manufacturers to 3D-print rotational mechanisms integrating motion sensors in a single pass has been designed by engineers at the Massachusetts Institute of Technology, Denmark's Aarhus University, and technology developer Accenture Labs. The system uses a conductive 3D printing filament so the devices can detect their angular position, rotation speed, and spin direction. The sensor incorporates three conductive patches partitioned by nonconductive material printed into the stationary plate, while a fourth patch is printed into the rotating plate. The researchers made a plug-in for the computer-aided design software SolidWorks that automatically incorporates sensors into a mechanism model, which can be fed to the multi-material 3D printer for production.

Full Article
Meta AI Unlocks Hundreds of Millions of Proteins to Aid Drug Discovery
The Wall Street Journal
Eric Niiler
March 16, 2023


The ESMFold program developed by Meta Platforms’ Meta AI unit uses artificial intelligence (AI) to predict the structure of hundreds of millions of proteins, which researchers believe could possibly accelerate drug discovery. Meta researchers used a large language model (LLM) to make predictions, providing ESMFold with a sequence of letters representing the amino acids comprising a protein's genetic code. The LLM learned to fill in blank or hidden areas, then determined how known protein sequences relate to structures that are already well-understood to anticipate new sequence structures. Meta AI compiled a public database of 617 million predicted proteins with ESMFold. Meta said the tool is 60 times faster but less accurate than Alphabet subsidiary DeepMind Technologies' AlphaFold protein-prediction computer model.

Full Article
*May Require Paid Registration

A worker holding the SEEQC's DQM system-on-a-chip with a pair of tweezers. Digital Chip Operates in Super-Cold Temps
Reuters
Jane Lee
March 15, 2023


A digital chip developed by quantum computer startup SEEQC can be operated at temperatures near zero Kelvin (-273.15 degrees Celsius) for use with quantum processors, which generally are stored in cryogenic chambers. The chip is located directly under the quantum processor and controls the qubits. Other chips under development will be used in slightly warmer parts of the cryogenic chamber. The chips use silicon wafers but no transistors.

Full Article
Teaching Robots to Perform Tasks like Human
USC Viterbi School of Engineering
Greg Hardesty
March 14, 2023


Researchers at the University of Southern California (USC), China's Fudan University, and Singapore's Sea AI Lab investigated the challenge of training robots to conduct tasks more independently, like humans. The analysis is a test for linking language models (LMs) with realistic environments for grounded planning of embodied tasks, said USC's "Bill" Yuchen Lin. Among the challenges to be addressed is providing robots with physical hardware to perform tasks, and a high-precision tracking system, according to Lin. The first step involves researchers closing the gap between natural language processing and robotics by designing and testing agents that can render language instructions as sequences of actions. The researchers developed an LM to transfer a teapot from a stove to a shelf, training it to generate subgoals step by step based on previous steps, rather than completing all steps simultaneously.

Full Article
'Spell-Checker for Statistics' Reduces Errors in Psychology Literature
Nature
Dalmeet Singh Chawla
March 15, 2023


The statcheck tool developed by Michèle Nuijten at Tilburg University in the Netherlands has reduced statistical errors in the peer review of psychology studies. Nuijten and Tilburg's Jelte Wicherts compared the P value error rate in two psychology journals before and after they implemented the tool, and in two publications that have never used it. The researchers reviewed more than 7,000 papers published since 2003; papers published in Psychological Science and Journal of Experimental and Social Psychology had 4.5% fewer statistical errors after deploying statcheck, while articles in the non-statcheck-using Journal of Experimental Psychology: General and Journal of Personality and Social Psychology cut errors by only about 1%. Nuijten credited the lower error rate among non-statcheck-using journals to greater awareness of issues concerning statistical significance during the review period, and to researchers using statcheck themselves prior to manuscript submission.

Full Article

Children wearing surgical masks awaiting treatment for TB in a clinic in South Africa. Algorithms Could Improve Pediatric Tuberculosis Diagnosis
Yale School of Public Health
Matt Kristoffersen
March 13, 2023


A team led by researchers at the Yale School of Public Health developed two evidence-based algorithms that aim to help health care providers diagnose tuberculosis (TB) in children. The researchers developed the algorithm-based scoring systems using data from 4,718 children with TB from 13 studies and 12 countries. Both algorithms generate a score based on the patient's symptoms and other details. Whether chest X-rays are available for the patient will determine which algorithm is used. An average of 85% of children with TB with scores above 10 will receive accurate diagnoses. Although the researchers have yet to formally test the algorithms, the World Health Organization's (WHO) latest consolidated guidelines for treating pediatric TB recommend the use of algorithms, and WHO specifically encourages the use of these new algorithms. Yale's Kenneth Gunasekera said the tool "is evidence-based and pragmatic in that it's easy for a health care worker to implement this score."

Full Article
Master's Degree in Microelectronics & Semiconductors
 
Ambassadors for ACM Program
 

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]