Data Science newsletter – October 28, 2021

Thanks to support from Microsoft, the new IDEaS Cloud Hub will provide faculty and students with new opportunities and resources for both education and research for those wishing to use the cloud for their computational needs.

“Microsoft has been a consistent and generous supporter of IDEaS since its inception, contributing funding for research and events, and providing cloud resources for research and education at Georgia Tech as well as for supporting programs under the South Big Data Hub,” said Srinivas Aluru, executive director of IDEaS. “The launch of the Cloud Hub takes this relationship to the next level and reflects our ongoing commitment to partnership.”

In a project that could unlock the world’s research papers for easier computerized analysis, an American technologist has released online a gigantic index of the words and short phrases contained in more than 100 million journal articles — including many paywalled papers.

The catalogue, which was released on 7 October and is free to use, holds tables of more than 355 billion words and sentence fragments listed next to the articles in which they appear. It is an effort to help scientists use software to glean insights from published work even if they have no legal access to the underlying papers, says its creator, Carl Malamud. He released the files under the auspices of Public Resource, a non-profit corporation in Sebastopol, California that he founded.

A new study led by Kellogg School of Management at Northwestern University found that although researchers’ productivity levels have mostly returned to pre-pandemic highs, scientists who did not pursue COVID-19-related research initiated 36% fewer new projects in 2020 compared to 2019. This dramatic decline in new projects suggests the pandemic’s impact on science may be longer lasting than commonly imagined.

“On the surface, it appears that researchers are as productive as they used to be,” said Dashun Wang, a professor at Kellogg School who led the study. “But, instead of generating new directions, they are busy working on established topics, writing up existing research, reviving legacy projects or revisiting old data. We found this to be true across many scientific disciplines — no fields were immune to the reduced number of projects.”

The pandemic led to a surge in COVID-19 research, but it severely disrupted other fields—shuttering labs, restricting travel, and leaving scientists with young children struggling to work without adequate child care. A flurry of studies indicates the productivity of women scientists slowed during the pandemic to a greater extent than their male colleagues. Now, a survey of a large swath of the scientific community reveals those impacts may be felt for years to come because many investigators—especially women and those with young children—were unable to start new research projects in 2020.

“There may be a cliff looming on the horizon,” says Dashun Wang, a professor at Northwestern University who headed the survey effort, whose results were published today in Nature Communications.

The University of Oregon has launched a new website to provide faculty guidance and resources related to academic freedom. Academic freedom enables faculty members to research and teach without fear of interference within established legal and ethical bounds.

The Academic Freedom site was created by the Office of the Provost as a resource for faculty members who may face questions, criticism or threats related to their research or teaching. With curated materials from across campus and other universities, the site contains easy-to-access links to relevant UO policies and safety resources, communications guidance, and publications addressing academic freedom. The Office of the Provost intends to update the site regularly and welcomes faculty suggestions for its improvement.

As part of Robert and Donna Manning’s recent historic $50 million gift to increase access and opportunity across the five-campus University of Massachusetts system, $18 million will endow the newly named Robert and Donna Manning College of Information and Computer Sciences. Concurrently, Massachusetts Governor Charlie Baker announced a $75 million state capital investment, backed by a $30 million commitment from the Amherst campus, to enhance and expand the college’s facilities. All together, these investments will allow the college to serve many more students while fueling new discoveries in computing research.

CodeCrew, a Memphis-based non-profit that empowers children and adults in underrepresented communities to be tech innovators and leaders, and CSforALL will share in a nearly one million dollar grant from the National Science Foundation (NSF) to increase computer science education (K-12) throughout the state of Tennessee.

The funds from the three-year grant will be used to leverage the Strategic CSforALL Resource and Implementation Planning Tool (SCRIPT) to create opportunities for local facilitator training throughout the state. CodeCrew will execute the training beginning in the Spring of 2022.

For a few weeks in 2017, Wanda Kukulski found herself binge-watching an unusual kind of film: videos of the insides of cells. They were made using a technique called cryo-electron tomography (cryo-ET) that allows researchers to view the proteins in cells at high resolution. In these videos, she could see all kinds of striking things, such as the inner workings of cells and the compartments inside them, in unprecedented detail. “I was so overwhelmed by the beauty and the complexity that in the evenings I would just watch them like I would watch a documentary,” recalls Kukulski, a biochemist at the University of Bern, Switzerland.

In recent years, imaging techniques such as cryo-ET have started to enable scientists to see biological molecules in their native environments. Unlike older methods that take individual proteins out of their niches to study them, these techniques provide a holistic view of proteins and other molecules together with the cellular landscape. Although they still have limitations — some researchers say that the resolution of cryo-ET, for example, is too low for molecules to be identified with certainty — the techniques are increasing in popularity and sophistication. Researchers who turn to them are not only mesmerized by the beautiful images, but also blown away by some of the secrets that are being revealed — such as the tricks bacteria use to infect cells or how mutated proteins drive neurodegenerative diseases such as Parkinson’s.

Why congratulate a college for doing something it should have done a long time ago? Many large public universities, including the University of California, the University of Texas, and Texas A&M, dropped legacy preferences years ago, as did Johns Hopkins University two years ago and MIT and CalTech before it. Pomona College, a highly ranked liberal-arts college like Amherst, dropped legacy preference in 2017. What’s more, even at those places that give the offspring of alumni an edge, or what admissions offices call a “tip,” legacies usually account for a small share of the enrolled class, and many of those admits would likely have gotten in without the advantage, or so the admissions deans defending the practice of birthright advantage like to say.

Wouldn’t it be better to go after all those jocks on campus? Division III sports have been called “affirmative action for rich white students,” and almost a third of Amherst’s roughly 1,800 students play one of the college’s 27 varsity sports.

Fear of never being able to pay off school loans is keeping many young Latinos in the U.S. from going to college or completing a degree, according to a report published in September.

State of play: Latinos tend to have more difficulty repaying school debt than white student borrowers, according to Federal Reserve data, at the same time that they need more loans in order to afford tuition.

Research by the University of North Carolina and the civil rights group UnidosUS also points to the need for reliable transportation between school, home and jobs as a significant factor before 2020 in Latinos successfully completing college.

Enrollment at U.S. colleges and universities is on track to fall by another nearly 500,000 undergraduate students this fall, continuing the historic drops that began with the start of the coronavirus pandemic, according to new data out Tuesday.

The decline of 3.2% in undergraduate enrollment this fall follows a similar drop of 3.4% the previous year, the first fall of the pandemic, according to the research from the National Student Clearinghouse Research Center.

The numbers are from a preliminary data set representing 8.4 million undergrad and graduate students from about 50% of U.S. colleges. The numbers show there are now 240,000 fewer undergraduates enrolled this fall compared with the same time last year, and if that rate of decline holds up for the rest of the colleges, that could translate into almost a half-million fewer undergraduate students.

Miami University held a groundbreaking ceremony on Monday, Oct. 25, for the incoming Richard M. McVey data science building. The ceremony, originally scheduled for Sept. 22 but postponed due to inclement weather, was held to celebrate the ongoing construction of the McVey building.

Miami alumnus Rick McVey (‘81) contributed a $20 million gift to the university – one of the top five largest single gifts in Miami’s history – to fund the project. After graduating with a degree in finance, McVey founded MarketAxess, an electronic trading platform for investors and dealers in 2000. McVey still serves as chairman and CEO.

“It is time for Miami to take a bold step forward in data science,” McVey said at the ceremony. “This is a special day for Miami, and an even more special day for me.”

The College of Engineering is proud to announce a $15 million lead gift from philanthropists and benefactors John and Marcia Price to build a new home for computing. Pending approval of the University of Utah Board of Trustees, the building will be named in their honor.

“We are pleased to lend our support to this effort that is so crucial to Utah’s expanding economy,” said John Price. “The University of Utah has an international reputation for innovation in computer science, and Marcia and I want to help ensure that opportunity for this generation and all future generations of Utah students.”

A data-science training program for equipping leaders to support improvement of health outcomes in Kenya, led by a team from NYU, Brown University, and Moi University in Kenya, was chosen as one of 19 initiatives funded by The National Institutes of Health (NIH) under its new Harnessing Data Science for Health Discovery and Innovation in Africa (DS-I Africa) program.

The $1.7 million award, part of the NIH’s mission to advance data science, catalyze innovation and spur health discoveries across Africa, establishes a consortium consisting of a data science platform and coordinating center, seven research hubs, seven data science research training programs, and four projects focused on studying the ethical, legal and social implications of data science resea

Georgia State University has launched a new graduate certificate program in Trustworthy Artificial Intelligence (AI) Systems, designed to prepare working professionals to advance their careers in this new and fast-growing field.

The fully online certificate program combines AI and machine learning with cybersecurity and privacy, which are among the most in-demand subfields of computer science.

“While there are many certificates in AI/machine learning and in cybersecurity separately, there’s currently no program in trustworthy AI to the best of our knowledge,” said Daniel Takabi, program director, associate professor of computer science and founding director of the Center for Information Security and Privacy: Interdisciplinary Research and Education (INSPIRE).

The University of Illinois Urbana-Champaign (UIUC) has launched a new Master of Science in Biomedical Image Computing (M.S. in BIC) degree program that provides rigorous training focused at the intersection of biomedical imaging science, machine learning and high-performance computing. The program comprises 36 credit hours of coursework that blends biomedical imaging, machine learning and computing concepts, and is spread over three semesters. Applications for the Fall 2022 on-campus cohort are currently being accepted on a rolling basis.

Charlottesville, VA November 19, starting at 11 a.m. Eastern. “Datapalooza is an annual conference hosted by the University of Virginia’s School of Data Science that brings together more than 500 attendees from across higher education, industry, and the greater data science community. This year’s theme “Learning In Action,” is inspired by the interdisciplinary values of data science with the intention to share knowledge and build community.” [registration required]

Hundreds of millions of people use spreadsheets, and formulas in those spreadsheets allow users to perform sophisticated analyses and transformations on their data. Although formula languages are simpler than general-purpose programming languages, writing these formulas can still be tedious and error-prone, especially for end-users. We’ve previously developed tools to understand patterns in spreadsheet data to automatically fill missing values in a column, but they were not built to support the process of writing formulas.

In “SpreadsheetCoder: Formula Prediction from Semi-structured Context”, published at ICML 2021, we describe a new model that learns to automatically generate formulas based on the rich context around a target cell. When a user starts writing a formula with the “=” sign in a target cell, the system generates possible relevant formulas for that cell by learning patterns of formulas in historical spreadsheets. The model uses the data present in neighboring rows and columns of the target cell as well as the header row as context.

Laws of UX is focused on making complex psychology heuristics accessible to more designers through an interactive resource that collects those that are relevant to user experience design. [free]

TL;DR: We often talk about visualizations as tools for amplifying cognition—but what if we took that analogy a bit further, looking to superhero comics and other science fiction as sources of inspiration for visualizations that can enhance human abilities in new and surprising ways? Based on a deep dive into perceptual and cognitive superpowers in fiction, we propose two ways of thinking about the relationship between these powers and visualization, and describe what it means for a visualization to feel empowering. We also illustrate a set of new “visualization superpowers” that highlight opportunities for new empowering data visualizations, as well as the challenges they must confront.