Five independent groups got different results in a drug-response experiment, despite sharing protocols, reagents, and cell lines. The researchers identify technical variables that could be to blame.
ShelfVision is a virtual-reality tool that uses artificial intelligence to monitor store shelves for product shortages, potentially solving a $1 trillion problem for retailers globally. Evendale-based Kinetic Vision Inc. is testing the product for retail clients it is not allowed to name. It’s the latest in a growing list of AI innovations targeting the retail industry.
“They perform these repetitive manual labor tasks so much faster,” said Greg Sweeney, an AI software engineer at Kinetic Vision. “We can deploy these models, have them assist humans in performing what’s repetitive: Checking for stock outs on a shelf, knowing which shelves to restock first, knowing when and how often products are displaced. Which products are most popular? Which locations are most popular? All of that can be determined automatically with an algorithm like this.”
The Brookings Institution, Heather Hurlburt and Tamara Cofman Wittes
from
Though the U.S. leads the world in national security in many respects, the gender diversity of the field has changed little from the mid-20th century. The ban on married women in the U.S. Foreign Service was lifted in 1971, and women first entered the service academies in 1975. Since then, women have broken glass ceilings as Cabinet members, flag officers, leading negotiators and combat warriors. Today, more than half of graduate students of international affairs are female, and entry-level ranks in the field include plenty of women as well as people outside the gender binary. But this surge of talent is not reflected in the upper ranks.
There are multiple reasons for this. For one thing, women aren’t getting promoted equitably. For another, more women than men leave the field before they move up to higher-ranking positions—some because of the “gender tax” of workplace sexual harassment outlined by our colleague Daniel Drezner. As a result, women have never exceeded 40 percent of senior positions (that is, assistant secretary level and above) at the State Department; at the Department of Defense, it is closer to 20 percent.
Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), Dartmouth, and the University of Washington are investigating a new drone design approach that combines the best of quadcopters and fixed-wing airplanes. Their work, which they detail in a newly published paper that’ll be presented later this month at the Siggraph conference in Los Angeles, resulted in a novel AI system that allows users to dream up drones of different sizes and shapes that can switch between hovering and gliding with a single flight controller.
The 10-year initiative will focus on digital health solutions, precision care, and workforce development through access to diagnostic and therapeutic equipment and educational and training resources.
Valued at $133 million, the partnership agreement will also support the research, educational and clinical care goals of the state’s NextGen Precision Health Initiative.
As part of the collaboration, Siemens Healthineers will supply medical imaging equipment and supporting infrastructure in the NextGen Precision Health Institute, and the parties will jointly contribute to the establishment of the Alliance for Precision Health that could reach north of $40 million.
A less-noticed win for DeepMind, the artificial-intelligence arm of Google’s parent Alphabet Inc., at a biennial biology conference could upend how drugmakers find and develop new medicines. It could also dial up pressure on the world’s largest pharmaceutical companies to prepare for a technological arms race. Already, a new breed of upstarts are jumping into the fray.
In December, at the CASP13 meeting in Riviera Maya, Mexico, DeepMind beat seasoned biologists at predicting the shapes of proteins, the basic building blocks of disease. The seemingly esoteric pursuit has serious implications: A tool that can accurately model protein structures could speed up the development of new drugs.
More than 300 people from multiple national laboratories, academia, and industry attended the inaugural “GPUs for Science” workshop at Berkeley Lab, held July 2-3, 2019.
When Perlmutter, the National Energy Research Scientific Computing Center’s (NERSC) next supercomputer, arrives in 2020, researchers computing at NERSC will need to be prepared to use GPUs for their simulation, data processing, and machine learning workloads. Toward this end, this event – which was the idea of, and conducted by, a group of NERSC postdoctoral fellows – was designed to facilitate the transition to GPU systems by giving attendees the motivation, tools, and expertise needed to make this change.
“This event was really the brainchild of the NERSC postdocs,” said Jack Deslippe, acting lead for NERSC’s Application Performance Group. “A lot of the excitement around GPUs for science is coming from some of the staff who are newest at NERSC. They are really driving this.”
EurekAlert! Science News, Cambridge University Press
from
A new open access journal from Cambridge University Press will explore the transformative effect of data science on engineering.
Launched this week with support from global charity Lloyd’s Register Foundation, Data-Centric Engineering will look at how the revolution in sensing, measurement, and data capture is driving the discovery of new materials, and new methods of manufacture, operation, control and construction.
A team of researchers at Caltech has developed a set of guidelines for designing biological circuits using tools from mechanical and electrical engineering. Like electric circuits—but made out of cells and living matter—biological circuits show promise in producing pharmaceuticals and biofuels.
For example, the antimalarial compound artemisinin is produced by an expensive tropical plant, so scientists at UC Berkeley have engineered the plant’s metabolism into yeast cells in order to synthesize artemisinin without using plants or soil. However, the ability to predict the behavior of these circuits—to design them on paper and then successfully implement the design—is still rudimentary.
The set of bioengineering design principles described in this new work could make building cellular systems more efficient and predictable.
In the beginning, God created the heavens and the Earth, and the Earth was shapeless and barren, so God added light and water and land and sky and plants and animals and humans. If you extend this belief forward, then God also created coal and oil and gas, which we began burning to do our own creating, on a massive scale. Health and wealth flowered across the planet, but there were consequences: first for the poor and marginalized, who were more exposed to the pollution, and then for everyone, in the form of a changing climate that is endangering creation. Stretch the belief a bit further, and in 1972, God created Katharine Hayhoe, who would grow up to be both an evangelical Christian and a climate scientist. Join these identities together, and you get another of God’s creations: a prophet.
The BU School of Public Health has signed a five-year partnership with digital health company Sharecare to mine health and community data to improve well-being nationwide.
SPH and Sharecare will build a Community Well-Being (CWI) Index using a vast trove of health data collected by Sharecare combined with SPH data on social determinants of health, all mined by the SPH Biostatistics and Epidemiology Data Analytics Center (BEDAC).
“There’s going to be new apps coming online every single day,” Steven Lane, clinical informatics director of Sutter Health and a member of ONC’s HIT Advisory Committee, told POLITICO. “[M]ost patients who are using these tools don’t fully understand the privacy implications.”
HHS has been clear that providers aren’t responsible for the health data once a patient decides to send it to an app outside the health system. But some of them feel an ethical responsibility to ensure patients don’t share their data irresponsibly, and are calling for federal guidelines outlining what health apps can do with patient data and how to get consent.
Others say that perspective is paternalistic, pointing out that consumers regularly download apps outside of health care and can be trusted to protect their own privacy.
To respond to the explosion of smart machines in the workplace, higher education should develop a new curriculum tailored to the age of artificial intelligence, promote learning through experience, and expand lifelong learning, experienced hands in the business say here (caveat auditor: start of audio is a bit garbled) and here. Issues extensively surveyed this landscape recently in a series of articles that notably are oriented toward solutions rather than diagnosis of problems, and that center on the particular needs of students themselves as they prepare for a very different world.
Interpretability remains one of the biggest challenges of modern deep learning applications. The recent advancements in computation models and deep learning research have enabled the creation of highly sophisticated models that can include thousands of hidden layers and tens of millions of neurons. While its relatively simple to create incredibly advanced deep neural network models, its understanding how those models create and use knowledge remains a challenge. Recently, researchers from the Google Brain team published a paper proposing a new method called Concept Activation Vectors(CAVs) that takes a new angle to the interpretability of deep learning models.
To increase the performance of Facebook’s apps, we have teams that continuously improve our JavaScript code and platforms. As we analyzed performance data, we noticed that the JavaScript engine itself was a significant factor in startup performance and download size. With this data in hand, we knew we had to optimize JavaScript performance in the more constrained environments of a mobile phone compared with a desktop or laptop. After exploring other options, we built a new JavaScript engine we call Hermes. It is designed to improve app performance, focusing on our React Native apps, even on mass-market devices with limited memory, slow storage, and reduced computing power.
The ML Enabler, as the name suggests, enables applications to take advantage of Machine Learning. It organizes and efficiently stores ML-derived map data so Tasking Manager and other tools can draw on this information via a standard API and enhance their own functionality. For a further explanation of how Tasking Manager and iD would use ML-derived data, check out our previous post and the recent announcement from HOT. With ML Enabler, we hope to bring more models to better support project manager’s plan task areas and help volunteers add new map features faster.
“Today’s paper choice builds on that by distilling wisdom collected from Microsoft, Airbnb, Snap, Skyscanner, Outreach.io, Intuit, Netflix, and Booking.com into a series of checklists that you can use as a basis for your own processes.”