Welcome to the June 9, 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."
|
|
|
Pioneer of Multicore Processor Design Receives the ACM-IEEE CS Eckert-Mauchly Award
ACM
June 7, 2023
Stanford University's Kunle Olukotun has been named to receive the ACM-IEEE CS Eckert-Mauchly Award for his pioneering work in developing parallel systems. Olukotun became a preeminent designer of multicore processors in the early 1990s, presenting a pivotal paper on their performance advantages at the 1996 ACM Conference on Architectural Support for Programming Languages and Operating Systems. His multicore design evolved into the industry standard, and he based his exploration of fine-grained multithreading on his understanding of multicore processors and thread-level speculation. Olukotun helped lead the development of the Transactional Coherence and Consistency approach to make parallel programming simpler. His work creating the coarse-grained reconfigurable dataflow has been a core contribution to machine learning and other data-intensive applications.
|
'Incompetent' Driverless Cars Wreak Havoc on San Francisco
SF Gate
Ariana Bindman
June 6, 2023
In San Francisco, autonomous vehicles (AVs) from Cruise and Waymo are undergoing "rigorous" training on local roadways, raising safety concerns among residents. Both companies received approval from the California Public Utilities Commission (CPUC) for their AVs to carry passengers through designated regions of the city. The CPUC appears ready to approve further expansions for the companies, despite complaints from local transportation agencies that the AVs cause traffic congestion and put pedestrians, drivers, and emergency personnel at risk. In a January open letter to the CPUC, the San Francisco Municipal Transportation Agency (SFMTA) reported an increase in 911 calls about AVs driving erratically, blocking lanes, or stopping in the middle of the road. SFMTA's Stephen Chun said, "In the short term, we need AVs to demonstrate that they can operate on our streets without creating new hazards."
|
3D Printing Creates Heart Valves in Minutes
Interesting Engineering
Rupendra Brahambhatt
June 7, 2023
Researchers at Harvard University, Germany's Deutsches Herzzentrum der Charite, and Switzerland's University of Zurich used three-dimensional (3D) printing to fabricate an artificial heart valve in under 10 minutes. A nanofiber meshwork in the synthetic valve mimics the extracellular matrix supporting the growth of natural heart valves. Focused rotary jet spinning enables the rapid additive manufacturing (3D printing) of micro or nanofiber scaffolds with adjustable alignments by fabricating a heart valve-shaped frame, then pushing in a liquid polymer with air jets to form the meshwork. The researchers tested a prototype valve in sheep, which effectively controlled blood flow, with no side effects.
|
Algorithm Uses Phone Camera to Detect Blood Oxygen Levels
Purdue University
June 7, 2023
Purdue University researchers have developed an algorithm to improve the speed and accuracy of medical diagnoses using smartphone sensors. Although smartphone cameras capture only red, green, and blue wavelengths of light in each pixel, the researchers were able to reconstruct the full spectrum of visible light in each pixel of an image taken by a smartphone camera using deep learning, statistical techniques, and an understanding of light-tissue interactions. The researchers found their technology produced information about blood oxygen levels in study participants' eyelids faster than commercially available hyperspectral imaging equipment, while being less expensive and just as accurate.
|
DeepMind AI Creates Algorithms That Sort Data Faster Than Those Built by People
Nature
Matthew Hutson
June 7, 2023
Google DeepMind researchers created an artificial intelligence (AI) system based on the AlphaZero AI to formulate algorithms capable of sorting data up to three times faster than human-produced programs. The researchers initially used the AlphaDev system to sort numbers by size. AlphaDev can choose one of four types of value comparisons, moving values between locations or shifting to a different program segment. It attempts to sort a set of lists after each step, receiving rewards for the number of correctly sorted list items until all lists are sorted perfectly, or it reaches a program length threshold before starting a new program. AlphaDev's optimal algorithms sorted data 4% to 71% faster than human algorithms, depending on the processor used and how many values required sorting.
|
Lingering Effects of Neanderthal DNA Found in Modern Humans
Cornell Chronicle
June 7, 2023
A multi-institutional team of scientists developed a new series of computational genetic tools to determine the lingering effects Neanderthal genes have on modern human traits. The researchers analyzed more than 235,000 likely Neanderthal-originating genetic variants from information on nearly 300,000 non-African-descended British people in the U.K. Biobank dataset. They found 4,303 variants continue to influence 47 distinct genetic traits in modern humans. Said Cornell University’s April (Xinzhu) Wei, ”We found that several of the identified genes involved in modern human immune, metabolic, and developmental systems might have influenced human evolution after the ancestors’ migration out of Africa.”
|
Drone Taxis Take First Test Spin in Israel
The Times of Israel
Sharon Wrobel
June 6, 2023
Israel has begun test flights of autonomous aircraft for heavy cargo, and eventually passenger transport, as part of the Israel National Drone Initiative, a two-year, government-led project to manage traffic congestion. Israel’s Transportation Minister Miri Regev calls it "the first initiative of its kind in the world for an extensive and multidisciplinary examination of new technologies, including the transportation of cargo and later, people." Eleven drone operating and delivery companies took part in test flights of heavy cargo. Israel Innovation Authority's Dror Bin said, "The significance of these demonstrations lies in creating a regulatory sandbox that allows all relevant stakeholders to have practical experience before establishing a regulatory infrastructure that enables a wider economically viable model."
|
Scaling Audio-Visual Learning Without Labels
MIT News
Lauren Hinkel
June 5, 2023
A technique developed by a team led by researchers at the Massachusetts Institute of Technology (MIT) analyzes unlabeled audio and visual data using a combination of contrastive learning and masked data modeling. The goal was to scale machine learning tasks without the need for annotation, the way humans learn. The resulting neural network, the contrastive audio-visual masked autoencoder (CAV-MAE), can perform audio-visual retrieval and audio-visual event classification tasks, learning by prediction and comparison. MIT's Jim Glass said the new model " has the contrastive and the reconstruction loss, and compared to models that have been evaluated with similar data, it clearly does very well across a range of these tasks."
|
Instagram's Algorithms Are Promoting Accounts that Share Child Sex Abuse Content
CNBC
Jonathan Vanian
June 7, 2023
A probe by The Wall Street Journal and researchers at Stanford University and the University of Massachusetts, Amherst, found Instagram's recommendation algorithms connect and advertise accounts that share self-generated child sexual abuse material (SG-CSAM). The researchers explained, "Due to the widespread use of hashtags, relatively long life of seller accounts, and, especially, the effective recommendation algorithm, Instagram serves as the key discovery mechanism for this specific community of buyers and sellers." They also learned the algorithms promoted SG-CSAM "to users viewing an account in the network, allowing for account discovery without keyword searches." A spokesperson for Instagram parent Meta said the company is engaged in correcting these issues and has established "an internal task force" to investigate and resolve them.
|
Researchers Unveil First Chat-GPT-Designed Robot
EPFL (Switzerland)
Celia Luterbacher
June 7, 2023
Researchers from the Swiss Federal Institute of Technology, Lausanne (EPFL) and the Delft University of Technology in the Netherlands used the ChatGPT-3 large language model (LLM) to design a tomato-harvesting robotic gripper. The researchers and LLM held an "ideation" session to formulate the robot's purpose, design parameters, and specifications, then realized it via code refinement, fabrication, and troubleshooting. Said EPFL's Francesco Stella, "While computation has been largely used to assist engineers with technical implementation, for the first time, an AI [artificial intelligence] system can ideate new systems, thus automating high-level cognitive tasks. This could involve a shift of human roles to more technical ones."
|
Nvidia Spools Up Quantum Jet Engine Simulations
IEEE Spectrum
Edd Gent
June 7, 2023
Engineers at aircraft engine manufacturer Rolls-Royce have modeled the largest quantum circuit for computational fluid dynamics (CFD) in partnership with technology company Nvidia and Israeli quantum software firm Classiq. Last month, the partners announced they had tested a 39-quantum-bit (qubit) CFD circuit with a depth of 10 million layers using Nvidia's graphics processing unit (GPU)-based quantum simulator. Rolls-Royce based its quantum CFD simulation on the HHL algorithm, which can solve linear algebraic equations exponentially faster than classical strategies (given enough qubits). Rolls-Royce's Leigh Lapworth said Nvidia's A100 GPUs accelerated the simulation, enabling the circuit simulation to run at a sufficiently large scale to compare it meaningfully against experimental data of jet engines and their components in action.
|
Fitness App Loophole Allows Access to Home Addresses
NC State University News
Matt Shipman
June 7, 2023
North Carolina State University (NC State) researchers found an exploit in the Strava mobile fitness-tracking application that permits anyone to find personal information about certain users, including their home addresses. NC State's Anupam Das said a loophole in the app's heatmap feature aggregates anonymized data so users can see how many others exercise in a given area. The researchers determined anyone can look up all Strava users in a given area, as well as where their fitness routes start and end, by studying the heatmap's aggregate data. Kevin Childs, formerly of NC State, said the researchers contacted Strava about the loophole, and were told the app “Does not share heatmap data unless several users are active in a given area.”
|
3D-Printed Lungs Help UTHealth Houston Study Health Risks of Aerosols
Houston Chronicle
Evan MacDonald
June 8, 2023
Researchers at the UTHealth Houston School of Public Health developed a three-dimensionally printed replica of human lungs that can be transported outside the laboratory to measure real-world aerosol respiratory deposition. The Mobile Aerosol Lung Deposition Apparatus (MALDA)features a replica of the human airway system, including head airways, tracheobronchial airways, and alveoli, along with an aerosol measurement system that calculates within seconds the amount of aerosol deposits reaching each region of the respiratory system. The researchers will use the data to determine the levels of toxic substances in the atmosphere across Houston and compare it to the level deemed safe to breathe by the U.S. Environmental Protection Agency. UTHealth's Wei-Chung Su said, "By comparing the safe dose and the dose I collect from the field study, then I will be able to estimate the health risk."
|
|
