Data Science newsletter – March 3, 2021

Despite strides in family-leave offerings, and men taking a greater role in parenting, women in academia still experience about a 20% drop in productivity after having a child, while their male counterparts generally do not, according to new CU Boulder research.

The study, publlished this week in the journal Science Advances, suggests that persistent differences in parenting roles are the key reason that men tend to publish more research papers than women. Because publishing is closely linked to promotion, this gender gap could have long-term impacts on what academia looks like in the future.

The researchers also found that while parental leave is critically important for women seeking faculty positions, 43% of institutions have no such policy.

Using machine learning, researchers found that people’s biases about COVID-19 and its treatments are exacerbated when they read tweets from other users, a study published in JMIR showed.

The analysis also revealed that scientific events, like scientific publications, and non-scientific events, like speeches from politicians, equally influence health belief trends on social media.

The rapid spread of COVID-19 has resulted in an explosion of accurate and inaccurate information related to the pandemic – mainly across social media platforms, researchers noted.

“In the pandemic, social media has contributed to much of the information and misinformation and bias of the public’s attitude toward the disease, treatment and policy,” said corresponding study author Yuan Luo, chief Artificial Intelligence officer at the Institute for Augmented Intelligence in Medicine at Northwestern University Feinberg School of Medicine.

Traditional computer scientists and engineers are trained to develop solutions for specific needs, but aren’t always trained to consider their broader implications. Each new technology generation, and particularly the rise of artificial intelligence, leads to new kinds of systems, new ways of creating tools, and new forms of data, for which norms, rules, and laws frequently have yet to catch up. The kinds of impact that such innovations have in the world has often not been apparent until many years later.

As part of the efforts in Social and Ethical Responsibilities of Computing (SERC) within the MIT Stephen A. Schwarzman College of Computing, a new case studies series examines social, ethical, and policy challenges of present-day efforts in computing with the aim of facilitating the development of responsible “habits of mind and action” for those who create and deploy computing technologies.

Major universities are using their students’ race, among other variables, to predict how likely they are to drop out of school. Documents obtained by The Markup through public records requests show that some schools are using education research company EAB’s Navigate advising software to incorporate students’ race as what the company calls a “high-impact predictor” of student success—a practice experts worry could be pushing Black and other minority students into “easier” classes and majors.

The documents, called “predictive model reports,” describe how each university’s risk algorithm is tailored to fit the needs of its population. At least four out of seven schools from which The Markup obtained such documents incorporate race as a predictor, and two of those describe race as a “high impact predictor.” Two schools did not disclose the variables fed into their models.

More than 500 universities across the country use Navigate’s “risk” algorithms to evaluate their students. In addition to documents on how the models work, the Markup obtained aggregate student risk data from four large public universities—the University of Massachusetts Amherst, the University of Wisconsin–Milwaukee, the University of Houston, and Texas A&M University—for the fall 2020 semester. We found large disparities in how the software treats students of different races, and the disparity is particularly stark for Black students, who were deemed high risk at as much as quadruple the rate of their White peers.

The National Security Commission on AI believes that China could soon replace the U.S. as the world’s “AI superpower” and said there are serious military implications to consider.

In a report, it warned that AI systems will be used in the “pursuit of power” and that “AI will not stay in the domain of superpowers or the realm of science fiction.”

It also urges President Biden to reject calls for a global ban on highly controversial AI-powered autonomous weapons, saying that China and Russia are unlikely to keep to any treaty they sign.

Most of the biggest tech companies, which have been at the forefront of the AI revolution, are well aware of the risks of deploying flawed systems at scale. Tech companies publicly acknowledge the need for societal acceptance if their systems are to be trusted. Although historically allergic to government intervention, some industry bosses are even calling for stricter regulation in areas such as privacy and facial recognition technology.

A parallel is often drawn between two conferences held in Asilomar, California, in 1975 and 2017. At the first, a group of biologists, lawyers and doctors created a set of ethical guidelines around research into recombinant DNA. This opened an era of responsible and fruitful biomedical research that has helped us deal with the Covid-19 pandemic today. Inspired by the example, a group of AI experts repeated the exercise 42 years later and came up with an impressive set of guidelines for the beneficial use of the technology.

Translating such high principles into everyday practice is hard, especially when so much money is at stake.

The mandate of the National Security Commission on Artificial Intelligence’s (NSCAI) is to make recommendations to the President and Congress to “advance the development of artificial intelligence, machine learning, and associated technologies to comprehensively address the national security and defense needs of the United States.”

This Final Report presents the NSCAI’s strategy for winning the artificial intelligence era. The 16 chapters in the Main Report provide topline conclusions and recommendations. The accompanying Blueprints for Action outline more detailed steps that the U.S. Government should take to implement the recommendations.

As her chemistry laboratory got rolling here one winter morning, with masked students taking their stations at a pandemic-safe distance, community college professor Sherry Heidary aimed to teach more than how to separate mixtures through chromatography.

She exhorted one student to believe in herself. “You’re getting an A, right?” she said, more of a demand than a question. Then the professor gently chided a latecomer who didn’t turn in an assignment. “Not going to happen again, right? Good.”

The lesson behind the lesson: They all belong. There in the lab, there at Union County College, there on the path to a degree.

It’s an urgent message for community colleges everywhere as they struggle to fill classes after a shocking nationwide enrollment plunge in the fall that educators blame on the coronavirus crisis and economic and social upheaval.

For years, tech companies have touted blockchain technology as a means to develop identity systems that are secure and decentralized. The goal is to build a platform that could store information about official data without holding the actual documents or details themselves. Instead of just storing a scan of your birth certificate, for example, a decentralized ID platform might store a validated token that confirms the information in it. Then when you get carded at a bar or need proof of citizenship, you could share those pre-verified credentials instead of the actual document or data. Microsoft has been one of the leaders of this pack—and is now detailing tangible progress toward its vision of a decentralized digital ID.

At its Ignite conference today, Microsoft announced that it will launch a public preview of its “Azure Active Directory verifiable credentials” this spring. Think of the platform as a digital wallet like Apple Pay or Google Pay, but for identifiers rather than credit cards. Microsoft is starting with things like university transcripts, diplomas, and professional credentials, letting you add them to its Microsoft Authenticator app along with two-factor codes. It’s already testing the platform at Keio University in Tokyo, with the government of Flanders in Belgium, and with the United Kingdom’s National Health Service.

Increasing mortality rates among working-age Americans since 2010 have been mainly driven by drug- and alcohol-related deaths, suicide, and cardiometabolic conditions like diabetes and heart disease, according to a sweeping new report.

The report, released Tuesday by the National Academies of Sciences, Engineering, and Medicine, examined data from 1990 to 2017, and revealed that the rise in deaths among working-age adults (those between 25 and 64) was seen across rural and metropolitan areas and racial and ethnic groups. The increase also appears to be behind the recent fall in U.S. life expectancy, which is significantly lower than in other high-income countries.

“Things have really gotten worse since 2010,” said Kathleen Mullan Harris, the James Haar Distinguished Professor of Sociology at the University of North Carolina at Chapel Hill and the chair of the Committee on Rising Midlife Mortality Rates and Socioeconomic Disparities, which completed the report.

A research team funded by the National Institutes of Health has launched a study to assess performance and usability of a smartphone app paired with the Quidel QuickVue At-Home COVID-19 Test, which just received an Emergency Use Authorization(link is external) (EUA) from the U.S. Food and Drug Administration for use with a prescription. The home test was supported by NIH through the Rapid Acceleration of Diagnostics (RADx) initiative, which has spurred the development and commercial availability of millions of COVID-19 tests over the past year.

In the spring of 2019, the sky in parts of West Texas opened up, in some areas dropping hailstones as big as baseballs, according to the National Weather Service. Beyond cracking car windows and damaging rooftops, the hailstorm struck a 180-megawatt solar project developed by 174 Power Global, causing an estimated $70 million to $80 million in damages as ice smashed the project’s panels, made by Hanwha Q Cells.

The event got the insurance market’s attention.

“That’s really when the market changed overnight,” said Sara Kane, a senior vice president overseeing energy risk management at insurance broker Beecher Carlson.

Solar came up at a time when insurance was relatively affordable and easy to procure. Insurance was never an insignificant cost for developers, according to a 2010 report from the National Renewable Energy Laboratory (NREL). But it’s gotten significantly more expensive in recent years as natural disasters exacerbated by climate change have proliferated. Hurricanes have soaked the South and wildfires have destroyed property in the West, compelling insurers to reckon with what experts say are years’ worth of underpricing the risk of damages.

When it comes to improving nutrition in marginalized communities, education and infrastructure are some of the first investment targets that come to mind. But these kinds of efforts are not sustainable unless they take into account the differing impacts of climate change on local communities. In a new study published in Environmental Research Letters, researchers found that climate variables influence the diversity of children’s diets at levels that are statistically comparable to traditional influences like access to clean water, education quality, and distance to urban centers.

The unprecedentedly large scale research included analyzing health and climate data for 107,000 children under 5 from 19 different countries across six regions in Africa, Asia, and Latin America. Most previous work linking climate to diet diversity was focused on individual countries or regions, according to Meredith Niles, an assistant professor at the University of Vermont and lead author of the study. The new study takes a big data approach to encompass a much broader geographic and temporal scope.

Deciphering what flavor notes are present in a given coffee is more art than science. Tasters spend hours upon hours training their palates, translating the sensory information through the lens of past (highly subjective) flavor experiences to arrive at descriptors that best align with the original sensory input. But one startup is looking to take a more scientific approach to identifying flavors in coffee. The company is called Demetria and they have created an app to “detect a specific and high value sensory (“taste”) profile of green coffee.”

In a press release announcing the company, Demetria touts itself as “the first AI-powered taste and quality intelligence [software as a service] startup for the coffee supply chain.” The startup has already secured $3 million in seed funding led by Colombian-Israeli investment firm Celeritas as well as private investors Mercantil Colpatria.

With their app, Demetria is attempting to digitize the entire coffee assessment process. Using a portable near-infrared sensor, a coffee is analyzed and fingerprinted for biochemical markers, which then gets run through the company’s AI-based platform, the “e-Palate,” that then matches the biomarkers to their associated flavors as defined by the SCA flavor wheel. Per the press release, the analysis can be done at any stage of production line, from green to roasted coffee.

Virginia Gov. Ralph Northam (D) signed the Consumer Data Protection Act on Tuesday, making Virginia the second state in the U.S. to pass a comprehensive data privacy law.

The bill will give consumers the right to opt out of having their personal data processed for targeted advertising and the right to confirm if their data is being processed.

Online May 11-12. “You’ll hear from experts working at the intersection of tech and health care — at both Silicon Valley giants and upstarts just emerging from stealth — as they design the future of health.” [$$$]

For years Python and R have been pitted as mortal enemies in the world of data science, enticing its practitioners to choose a side and never look back – not anymore. It’s 2021 and it’s time for these two titans to join forces through reticulate which allows us to use Python and R together!

Conceptual abstraction and analogy-making are key abilities underlying humans’ abilities to learn, reason, and robustly adapt their knowledge to new domains. Despite of a long history of research on constructing AI systems with these abilities, no current AI system is anywhere close to a capability of forming humanlike abstractions or analogies. This paper reviews the advantages and limitations of several approaches toward this goal, including symbolic methods, deep learning, and probabilistic program induction. The paper concludes with several proposals for designing challenge tasks and evaluation measures in order to make quantifiable and generalizable progress in this area.