Welcome to the March 10, 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."
|
|
Hackers Breach Thousands of Security Cameras, Exposing Tesla, Jails, Hospitals
Bloomberg William Turton March 9, 2021
Hackers say they have compromised data from as many as 150,000 surveillance cameras, including footage from electric vehicle company Tesla. An international hacking collective executed the breach to demonstrate the ease of exposing video surveillance by targeting camera data provided by enterprise security startup Verkada. In addition to footage from Tesla factories and warehouses, the hackers exposed footage from the offices of software provider Cloudflare, and from hospitals, schools, jails, and police stations. Tillie Kottmann, one of the hackers claiming credit for the breach, said the collective obtained root access to cameras, enabling them to execute their own code; they exploited a Super Admin account to access the cameras, and found a username and password for an administrator account online. A Verkada spokesperson said the company has disabled all internal administrator accounts to block unauthorized access.
|
Smart Speakers Can Detect Abnormal Heart Rhythms, Researchers Find
The Washington Post Dalvin Brown March 9, 2021
University of Washington (UW) researchers have developed a contactless method of screening for irregular heartbeats using smart speakers and an artificial intelligence-powered system that employs sonar. UW's Arun Sridhar said the goal was to use existing appliances to advance edge cardiology and health monitoring. The system emits audio signals into a room at a volume undetectable to humans, and an algorithm identifies heartbeat vibrations from a person's chest wall as the pulses bounce back to the speaker; a second algorithm measures inter-beat intervals. The UW researchers trained the speakers to detect regular and irregular heart rhythms, and their readings were relatively accurate in comparison to those of medical-grade electrocardiogram monitors.
*May Require Paid Registration
|
Magnetic Boost Helps Squeeze More Data Onto Computer Hard Disks
New Scientist Matthew Sparkes March 9, 2021
A short-term technology developed by researchers at Japan’s Toshiba may help clear a path toward next-generation computer hard disks by utilizing microwaves with existing platter material. Next-generation disks are expected to have higher data-storage capacity as a result of microwave-assisted switching, and Toshiba's Hirofumi Suto and colleagues' approach works by amplifying the magnetic field from the read/write head. The team used this method to fabricate a commercial hard disk in a helium-filled container that is being sold in capacities of up to 18 terabytes. Siva Sivaram at hard-disc manufacturer Western Digital said future disks will utilize heat as a long-term measure for boosting data storage, "[b]ut it adds a lot of cost and complexity and reliability issues."
|
Drones vs. Hungry Moths: Dutch Use Tech to Protect Crops
Associated Press Mike Corder March 7, 2021
Dutch farmers are adopting palm-sized drones from autonomous systems developer PATS Indoor Drone Solutions to protect their crops against moths. The drones patrol greenhouses and kill moths by flying into them, directed by smart technology that uses special cameras to scan the airspace. PATS' Kevin van Hecke said the drone "sees the moth flying by, it knows where the drone is ... and then it just directs the drone towards the moth." Dutch cress grower Rob Baan said the system can distinguish between helpful and destructive insects. PATS' Bram Tijmons said, "We are targeting moths and we are taking out moths every night in an autonomous way without human intervention."
|
Facebook Researchers Report Advance in Computer Vision
The Wall Street Journal John McCormick March 4, 2021
A software toolkit developed by Facebook's artificial intelligence (AI) research arm could enable companies to create highly accurate computer vision software more quickly. Facebook AI's Vissl toolkit leverages self-supervised learning, in which AI models train themselves on large datasets without external labels. Facebook's Yann LeCun said the techniques "allow you to basically reduce the amount of labeled data that is required to reach reasonable performance." Gartner's Carlton Sapp said the time required to build computer vision systems potentially could be halved using such self-supervised learning methods. LeCun, named 2018 ACM A.M. Turing Award laureate for his work on deep neural networks, said the technique also will boost the accuracy of computer vision systems by allowing analysis of more items in an image. In tests on the ImageNet database, Facebook's techniques achieved 85% accuracy, compared to 80% for computer vision systems trained with supervised learning.
*May Require Paid Registration
|
Researchers Are Peering Inside Computer Brains. What They've Found Will Surprise You
Fortune Jeremy Kahn March 4, 2021
Researchers at artificial intelligence (AI) research company OpenAI developed new techniques to examine the inner workings of neural networks to help interpret their decision-making. As neuroscientists have found in studies of the human brain, the researchers found individual neurons in a large neural network used to identify and categorize images can encode a particular concept. This finding is important given the challenges of understanding the rationale behind decisions made by neural networks. The researchers used reverse-engineering techniques to determine what most activated a particular artificial neuron. Among other things, the researchers identified a bias that could enable someone to trick the AI into making incorrect identifications. Said OpenAI's Gabriel Goh, "I think you definitely see a lot of stereotyping in the model."
|
Study Reveals Extent of Privacy Vulnerabilities With Amazon's Alexa
North Carolina State University March 4, 2021
Researchers at North Carolina State University identified a number of privacy concerns related to programs, or skills, run on Amazon's voice-activated assistant Alexa. The researchers used an automated program to collect 90,194 unique skills in seven different skill stores, and an automated review process to analyze each skill. They found Amazon does not verify the name of the developer responsible for publishing the skill, meaning an attacker could register under the name of a trustworthy organization. They also found multiple skills can use the same invocation phrase, so consumers might think they are activating one skill but are inadvertently activating and sharing information with another. In addition, the researchers found that developers can modify the code of their programs to request additional information after receiving Amazon approval. Moreover, nearly a quarter (23.3%) of 1,146 skills requesting access to sensitive data had misleading or incomplete privacy policies, or lacked them altogether.
|
Tiny Computers Reveal How Wild Bats Hunt So Efficiently
Aarhus University (Denmark) Peter F. Gammelby; Laura Stidsholt March 3, 2021
Researchers at Denmark's Aarhus University and Germany's Max Planck Institute of Ornithology used 3-gram computers attached to wild greater mouse-eared bats in Bulgaria to study how they hunt. The miniature tags record each bat's echolocation calls and movement in three dimensions. The institute's Holger Goerlitz said, "We found that hunting bats narrow their sensory volumes by more than a thousand times to only focus on the prey, and thereby reduce the clutter from other echoes. It's like an acoustic version of a tunnel vision that briefly makes their world much simpler." Said Aarhus University's Mark Johnson, who developed the tags, "It was a real challenge to make a computer so small that it could work on a flying bat and still be sensitive enough to pick up these weak sounds."
|
Algorithm Could Reduce Complexity of Big Data
Texas A&M Engineering News Stephanie Jones March 8, 2021
Researchers at Texas A&M University, the University of Texas at Austin, and Princeton University have developed an algorithm that can be applied to large datasets, with the ability to extract and directly order features from most to least salient. Texas A&M's Reza Oftadeh said, "There are many ad hoc ways to extract these features using machine learning algorithms, but we now have a fully rigorous theoretical proof that our model can find and extract these prominent features from the data simultaneously, doing so in one pass of the algorithm." The algorithm adds a new cost function to an artificial neural network to provide the exact location of features directly ordered by their relative performance, allowing it to perform classic data analysis on larger datasets more efficiently.
|
Storing the Declaration of Independence in a Single Molecule
UCI News Ian Anzlowar March 1, 2021
Researchers at the University of California, Irvine (UCI) are using an artificial variation of DNA for data storage. The researchers are using the four-letter nucleotide code in DNA instead of the binary system to transcribe data to a DNA strand. They sequentially assign each nucleotide a specific binary number, which allows them to write a binary sequence using the nucleotides. A special enzyme that connects the two sequences is added when the genetic code must be retrieved. For their experiment, the researchers chose threose nucleic acid (TNA), a synthetic genetic polymer less prone to degradation from physical factors. The researchers were able to transcribe the Declaration of Independence and the UCI seal to a solution of TNA, and recover them. UCI's John Chaput suggested all data generated during all of human history could be stored in just a half-cup of liquid TNA.
|
Engineering Platform Offers Collaborative Cloud Options for Sustainable Manufacturing
Purdue University News Chris Adam March 3, 2021
A new cloud-based platform designed by Purdue University researchers aims to map inter-industry dependencies for materials and waste generation among manufacturers in sectors linked to bio-based economies. Purdue's Shweta Singh and colleagues invented a method for automatically generating physical input-output tables to track flows in these networks. Said Singh, "Our new platform allows for dynamic changes in manufacturing network via mechanistic models developed as computer codes or simulation systems to update network structure for industrial interactions. The goal of this technology is to assist manufacturers to track the materials flow and supply network demand to optimize the process and reduce overall waste, as well as assist in the decision-making process to pick the most sustainable and resilient technology in any supplier network."
|
Algorithmic Approaches for Assessing Pollution Reduction Policies Can Reveal Shifts in Environmental Protection of Minority Communities
Stanford News Rob Jordan March 8, 2021
Analysis of a U.S. Environmental Protection Agency (EPA) initiative by researchers at Stanford University's Regulation, Evaluation, and Governance Laboratory (RegLab) and Stanford Law School showed how algorithmic design determines which communities are targeted for pollution compliance and penalized for violations. RegLab investigators analyzed how each state classified permits with similar functions to influence their inclusion in EPA's initiative to reduce Clean Water Act violations. Machine learning models screened hundreds of millions of observations from EPA databases to predict severe violations, and how much pollution each facility would generate. The researchers evaluated demographic data, such as household income and minority population, for the areas where each model indicated the riskiest facilities were located. RegLab's Elinor Benami said, "Careful algorithmic design can help regulators transparently identify how objectives translate to implementation while using these techniques to address persistent capacity constraints."
|
Helping Soft Robots Turn Rigid on Demand
MIT News Daniel Ackerman March 3, 2021
Massachusetts Institute of Technology (MIT) researchers used computer models to design a soft-bodied, cable-driven robot that turns rigid on demand by simultaneously controlling its position and stiffness. MIT's James Bern said, "It's just encoding that idea [of on-demand rigidity] into something a computer can work with." Bern used this roadmap to model the tuning of movement and rigidity in robots of various shapes, and tested their ability, when stiffened, to resist displacement when pushed. In simulation, the robots generally remained rigid as intended, but were not equally resistant from all angles. Bern is constructing a prototype robot to test the control system, and said his hope is to "start making soft robots that are safe but can also act rigid on demand, and expand the spectrum of tasks robots can perform."
|
|