Lifeboat Foundation

Two researchers at Duke University have recently devised a useful approach to examine how essential certain variables are for increasing the reliability/accuracy of predictive models. Their paper, published in Nature Machine Intelligence, could ultimately aid the development of more reliable and better performing machine-learning algorithms for a variety of applications.

“Most people pick a predictive machine-learning technique and examine which variables are important or relevant to its predictions afterwards,” Jiayun Dong, one of the researchers who carried out the study, told TechXplore. “What if there were two models that had similar performance but used wildly different variables? If that was the case, an analyst could make a mistake and think that one variable is important, when in fact, there is a different, equally good model for which a totally different set of variables is important.”

Dong and his colleague Cynthia Rudin introduced a method that researchers can use to examine the importance of variables for a variety of almost-optimal predictive models. This approach, which they refer to as “variable importance clouds,” could be used to gain a better understanding of machine-learning models before selecting the most promising to complete a given task.

How true Eric Klien.

Only 5% of internet users currently connect via satellite in the US. But that number could jump significantly when SpaceX’s service finally gets off the ground, according to a new survey.

Even for the brain of a worm, the best theory on offer would, she says, take several billion years to calculate. That can’t be the right answer.

Stuck at home with time on their hands, millions of amateurs around the world are gathering information on everything from birds to plants to Covid-19 at the request of institutional researchers. And while quarantine is mostly a nightmare for us, it’s been a great accelerant for science.

From backyard astronomy to birding, amateurs have been busy collecting data — and making real discoveries.

Continue reading “Citizen science is booming during the pandemic” »

The process of systems integration (SI) functionally links together infrastructure, computing systems, and applications. SI can allow for economies of scale, streamlined manufacturing, and better efficiency and innovation through combined research and development.

New to the systems integration toolbox are the emergence of transformative technologies and, especially, the growing capability to integrate functions due to exponential advances in computing, data analytics, and material science. These new capabilities are already having a significant impact on creating our future destinies.

The systems integration process has served us well and will continue to do so. But it needs augmenting. We are on the cusp of scientific discovery that often combines the physical with the digital—the Techno-Fusion or merging of technologies. Like Techno-Fusion in music, Techno-Fusion in technologies is really a trend that experiments and transcends traditional ways of integration. Among many, there are five grouping areas that I consider good examples to highlight the changing paradigm. They are: Smart Cities and the Internet of Things (IoT); Artificial Intelligence (AI), Machine Learning (ML), Quantum and Super Computing, and Robotics; Augmented Reality (AR) and Virtual Reality Technologies (VR); Health, Medicine, and Life Sciences Technologies; and Advanced Imaging Science.

The new clinical trial data shows a single shot of the vaccine “gives sustainable antibodies,” Dr. Paul Stoffels, J&J chief scientific officer, told CNBC.

I never thought about it, but these 3D magic eyes also work on computers.

On the path to writing his Ph.D. dissertation, Lucio Milanese made a discovery—one that refocused his research, and will now likely dominate his thesis.

Milanese studies plasma, a gas-like flow of ions and electrons that comprises 99 percent of the visible universe, including the Earth’s ionosphere, interstellar space, the solar wind, and the environment of stars. Plasmas, like other fluids, are often found in a turbulent state characterized by chaotic, unpredictable motion, providing multiple challenges to researchers who seek to understand the cosmic universe or hope to harness burning plasmas for fusion energy.

Milanese is interested in what physicist Richard Feynman called “the most important unsolved problem of classical physics”—turbulence. In this case, the focus is plasma turbulence, its nature and structure.

A team of researchers at Columbia University has developed a way to allow DNA strands to store more data. In their study, published in the journal Science, the group applied a small amount of electricity to DNA strands to allow for encoding more information than was possible with other methods.

For several years, researchers have been looking for ways to increase data storage capacity—storage requirements are expected to exceed capacity in the near future as demand skyrockets. One such approach has involved encoding data into strands of DNA—prior research has shown that it is possible. In the early stages of such research, scientists manually edited strands to add characteristics to represent zeroes or ones. More recently, researchers have used the CRISPR gene editing tool. Most such studies used DNA extracted from the tissue of deceased animals. More recently, researchers have begun efforts to move the research to living animals because it will last longer. And not just in the edited strands—the information they contain could conceivably be passed on to offspring, allowing data to be stored for very long periods of time.

Back in 2017, another team at Columbia University used CRISPR to detect a certain signal—in their case, it was the presence of sugar molecules. Adding such molecules resulted in gene expressions of plasmid DNA. Over time, the editing process was improved as genetic bits were added to represent ones and zeroes. Unfortunately, the system only allowed for storing a few bits of data.