Data Science newsletter – April 4, 2021

If you’ve ever worried a bear might be after your picnic basket, you may want to take the hardest, hilliest trail to your destination. That’s the take-home message of a new study, in which researchers got nine bears to run on treadmills—a first for science—and found that they, like their laziest human counterparts, prefer flat paths to save energy. The study, scientists say, may help explain why bears are often found around popular hiking trails.

Fibr relies on citizen scientists like you (no training needed) to guide the algorithm towards new innovations in #neuroscience and beyond. Watch the demo & get started: http://bit.ly/Fibr-demo

The main objective of Hawkins’s new book is to present his grand theory of intelligence: both how the human brain produces this feature, and what it really means to be intelligent. Happily for engineers, he then goes on to consider how his theories could impact AI research. He spoke with IEEE Spectrum about the path forward for machine learning.

A fast, low-cost technique to see and count viruses or proteins from a sample in real time, without any chemicals or dyes, could underpin a new class of devices for rapid diagnostics and viral load monitoring, including HIV and the virus that causes COVID-19.

Researchers at the University of Illinois Urbana-Champaign described the technique, called Photonic Resonator Interferometric Scattering Microscopy, or PRISM, in the journal Nature Communications.

“We have developed a new form of microscopy that amplifies the interaction between light and biological materials. We can use it for very rapid and sensitive forms of diagnostic testing, and also as a very powerful tool for understanding biological processes at the scale of individual items, like counting individual proteins or recording individual protein interactions,” said study leader Brian Cunningham.

President Joe Biden is set to unveil his infrastructure plan this week. He says he wants to spend 2% of our total economic output — our GDP — on science. That may not sound like much, but it’s more than twice what we’re spending now.

The conflict in Reno underlines the broader tensions between conventional therapy and the cash-flush companies racing to disrupt it. Many therapists are skeptical of the quality of care delivered by Talkspace and startups offering similar services. In Reno, they also say a short-term deal won’t have a lasting impact on mental health care.

But supporters say that apps like Talkspace are effective and that they fill a gap in the nation’s tattered mental health system. In many communities, therapists couldn’t meet demand even before the Covid-19 crisis stretched their services beyond their limit. Conventional therapy is also increasingly expensive, with many providers declining to accept health insurance.

“Talkspace is not going after therapists,” Mark Hirschhorn, Talkspace’s president and chief operating officer, told STAT. “What Talkspace is doing is providing a very, very essential need that has been unmet in the marketplace as a result of the either shortage of available therapists or the inability of individuals to access those therapists.”

There appears to be nowhere left on Earth where astronomers can view the stars without light pollution from space junk and satellites, according to a new analysis. The study considered the tens of thousands of objects in orbit as of 2020—before an onslaught of thousands more satellites that companies plan to launch in the coming years.

Call it a pyrrhic victory.

For years, privacy advocates argued that data collection and sharing among countless hidden ad tech intermediaries, via third-party cookies, was a privacy invasion. Thanks in part to their advocacy, government and consumer pressure for more privacy protections has finally pushed Google and others to disable third-party cookies. Now the digital ad industry is gravitating toward replacements that some in the privacy community consider even more invasive.

“Indeed, that irony is not lost on me,” said technologist Ashkan Soltani, who helped craft the California Consumer Privacy Act and served in the Division of Privacy and Identity Protection at the Federal Trade Commission.

By combining machine learning, molecular dynamics simulations and experiments it has been possible to design antimicrobial peptides from scratch.1 The approach by researchers at IBM is an important advance in a field where data is scarce and trial-and-error design is expensive and slow.

Antimicrobial peptides – small molecules consisting of 12 to 50 amino acids – are promising drug candidates for tackling antibiotic resistance. ‘The co-evolution of antimicrobial peptides and bacterial phyla over millions of years suggests that resistance development against antimicrobial peptides is unlikely, but that should be taken with caution,’ comments Håvard Jenssen at Roskilde University in Denmark, who was not involved in the study.

Academic research published last week looked at the telemetry traffic sent by modern iOS and Android devices back to Apple and Google servers and found that Google collects around 20 times more telemetry data from Android devices than Apple from iOS.

The research, conducted by Professor Douglas J. Leith from Trinity College at the University of Dublin, analyzed traffic originating from iOS and Android devices heading to Apple and Google servers at various stages of a phone’s operation.

With the help of AI software, decentralized energy sources can send any excess electricity they produce to the grid, while utilities direct that power to where it’s needed. Similarly, energy storage in industrial facilities, office buildings, homes, and cars can hold excess energy when demand is low, while AI deploys that power when generation is inadequate or impossible.

That’s a lot of moving parts requiring coordination, forecasting and optimization to keep the grid in balance. If you think of DERs as individual musicians, a utility is a conductor keeping the orchestra in sync as AI composes the symphony in real-time.

This makes an AI-centered system a potential game-changer. Shifting from an infrastructure heavy system to one centered on AI enables forecasting and control in seconds – not days – resulting in a grid that is more resilient and flexible when unforeseen events occur.

Cleveland Clinic and IBM have announced a ten-year partnership to launch the Discovery Accelerator, a center aimed at using artificial intelligence, high performance on the hybrid cloud, and quantum computing technologies to speed discovery in healthcare.

The collaboration is expected to build a comprehensive research and clinical infrastructure to enable big data medical research in ethical, privacy-preserving ways. The partnership will also facilitate discoveries for patient care and innovative approaches to public health threats like COVID-19.

Through the Discovery Accelerator, researchers plan to use advanced technologies to generate and analyze data to help enhance research in the new Global Center for Pathogen Research and Human Health in areas like population health, genomics, clinical applications, and chemical and drug discovery.

Researchers at North Carolina State University have designed and demonstrated a new system that allows them to remotely monitor the behavior of freshwater mussels. The system could be used to alert researchers to the presence of toxic substances in aquatic ecosystems.

“When mussels feed, they open their shells; but if there’s something noxious in the water, they may immediately close their shells, all at once,” says Jay Levine, co-author of a paper on the work and a professor of epidemiology at NC State. “Folks have been trying to find ways to measure how widely mussels or oysters open their shells off and on since the 1950s, but there have been a wide variety of challenges. We needed something that allows the animals to move, can be placed in streams and collects data – and now we have it.”

The public health response to the COVID-19 pandemic introduced the idea of bubbles to our social lives. British Columbia restricted socializing to core bubbles: immediate household members or, for those living alone, a maximum of two people who could be seen regularly.

Last summer, when “good times” were in the cards, public health guidelines provided for a “safe six” bubble of friends. The March update allows for outdoor bubbles of ten, with an emphasis on the fact that these should be a consistent set of ten people.

These strict directives to stick to our bubbles are essential to prevent the transmission of COVID-19. At the same time, this way of thinking disrupts our understanding of the kinds and quantities of social interactions needed to make healthy cities possible.

Online April 12, starting at 3 p.m. Eastern. “For the next seminar in our Bold Predictions seminar series, Drs. Tom Gingeras and Tuuli Lappalainen will describe their research and how it is helping to understand general features about the epigenetic landscape and transcriptional output!”

These two technologies are notoriously tricky to combine. The crux of the issue is that they both want to handle the DOM.

New developments and features from our sponsored projects, announcements about our programs including PyData, job opportunities, and more!

— a public good software signing service under the @LinuxFoundation
. Courtesy of @BUCompSci
@BU_Spark
@BU_Tweets
alumn @MBestavros
@RedHat
!

This talk provides a foundational intuition coming from the core machine learning principles, and introduces the challenges around deployed models such as re-training, bias/variance, and general MLE challenges https://youtube.com/watch?v=06-AZX