Data Science newsletter – April 6, 2021

Southwest Research Institute, in collaboration with Vanderbilt University, is developing machine learning algorithms to help the Tennessee Department of Transportation (TDOT) coordinate traffic management and incident response along portions of Interstate 24 in the rapidly growing Nashville region.

The project will use artificial intelligence to enhance an integrated corridor management (ICM) system, using software and systems to promote smart mobility and improve collaboration among various transportation agencies.

“SwRI’s ICM solutions fuse data across freeways, surface streets and transit systems to help balance traffic flow and improve performance of the entire corridor,” said Samantha Blaisdell, a program manager at SwRI.

The science committee in the U.S. House of Representatives wants to more than double the budget of the National Science Foundation (NSF) in the next 5 years, devoting a sizeable chunk of the extra money to a new directorate that would accelerate the process of turning basic research into new technologies and products. But its version of a technology directorate would be much smaller and more in line with the way NSF traditionally funds research than the one already proposed by Senate Majority Leader Chuck Schumer (D–NY), which emphasizes the economic and security threats posed by China.

The House bill, introduced today with bipartisan support, would lift NSF’s overall budget from the current $8.5 billion to $18.3 billion in 2026. In addition to growing the agency’s existing seven research and education directorates, the bill would create an eighth, called Science and Engineering Solutions (SES). Its budget would start at $1 billion in 2022 and grow to $5 billion by 2026.

Both the House legislation and Schumer’s Endless Frontier Act (EFA) see the new directorate as a way for NSF to do better in applying basic research findings to major societal challenges, from combating climate change and health inequities to strengthening economic and national security.

The country eliminated Covid 19 in the community by shutting its border in mid-March 2020, introducing compulsory quarantine for all returnees, and instituting a series of lockdowns to stamp out existing clusters. Every so often, a case slips through the border, causing a small outbreak. But without community transmission in the background, New Zealand can drill down into individual cases with forensic detail.

Using a mixture of genomic sequencing and epidemiology, the country works to identify precisely who gave the virus to whom and – often enough – the environment in which it happened.

That knowledge has proven critical for New Zealand’s pandemic response, allowing it to avoid longer lockdowns by more precisely mapping the spread. But it’s also yielded scientific insights to the rest of the world into exactly how and where Covid spreads. With conditions controlled to a degree that’s simply not possible elsewhere, the country behaves like a laboratory for the world.

Using a large, nationally representative sample of first year undergraduate students we tested whether instructors’ use of diversity philosophies could impact the learning of new math and science content among Students of Color and White students. Participants (N = 688) were randomly assigned to one of nine simulated online course environments using a 3 (diversity philosophy: Multicultural, Colorblind, Control) × 3 (lesson: Chemistry, Physics, Math) × 2 (participant race: Students of Color, White students) between-participants experimental design. After listening to an audio welcome message from the instructor and reading the course syllabus, both of which contained the embedded diversity philosophy manipulation, participants watched a novel 10-minute lesson, completed a comprehension quiz, as well as measures of belonging and perceived instructor bias. Students of Color showed greater comprehension of the math/science lesson in the multicultural condition compared to the colorblind condition. Students of Color also perceived the instructor to be less biased in the multicultural condition compared to the colorblind condition. White students tended to either be unaffected or oppositely affected by the diversity philosophy manipulation. Overall, results suggest that college instructors’ use of multicultural (or colorblind) language sends a signal of inclusion (or exclusion) to Students of Color, affecting not only their social experience in the class but also their learning potential.

How the pandemic has thrown college admissions process into a kind of slow-motion chaos. One of the biggest changes: most colleges have stopped requiring the SAT. For decades, there’s been a debate over whether schools should drop the test. What’s it mean that it finally happened? [audio, 1:00:56]

University of Iowa faculty, staff, and students are working to increase the accessibility and usability of nearly a half-million pages of emails related to the Flint water crisis.

Telehealth has become a critical way for doctors to still provide health care while minimizing in-person contact during COVID-19. But with phone or Zoom appointments, it’s harder for doctors to get important vital signs from a patient, such as their pulse or respiration rate, in real time.

A University of Washington-led team has developed a method that uses the camera on a person’s smartphone or computer to take their pulse and respiration signal from a real-time video of their face. The researchers presented this state-of-the-art system in December at the Neural Information Processing Systems conference.

Now the team is proposing a better system to measure these physiological signals. This system is less likely to be tripped up by different cameras, lighting conditions or facial features, such as skin color. The researchers will present these findings April 8 at the ACM Conference on Health, Interference, and Learning.

The Daily’s Data Team examined data collected over the past 10 years from Explore Courses to analyze enrollment trends in tech and tech ethics classes. The number of tech classes offered at Stanford has increased and enrollment in tech classes has also increased. However, the number of tech ethics courses offered by the university and the enrollment figures in these classes have not kept up.

Every day, companies use data in innovative ways to inform their decisions–in everything from targeted marketing to autonomous driving vehicles. Data scientists are the people who analyze, review and present that valuable information to predict outcomes, and provide data-enabled insights critical to making important decisions. Clarkson University now offers a master of science degree in applied data science, one of today’s most in-demand fields.

“By renaming our program to applied data science, we now more accurately reflect the focus of what students learn in courses on the management, analysis, interpretation, and manipulation of large, complex data sets,” said Operations and Information Systems Professor Boris Jukic, director of the program. “This degree offers a multi-disciplinary education that differentiates Clarkson graduates in the global job marketplace. The skills they gain here allow them to critically evaluate and address real-world problems from a data-enabled perspective.”

Founded in 2020, FloWaste has the potential to solve a multibillion-dollar global problem. The technology uses 3D imaging and machine learning to capture volumetric food data, categorising and quantifying the food detected.

“We attach 3D cameras to kitchen waste bins and gather information as to what items are being wasted, when and how much,” Mc Donnell explained. “This data can then be used to optimise the kitchen going forward, bringing insights to management to reduce waste and unnecessary costs.”

The genesis of the idea came from a user-centred design project Mc Donnell completed at Trinity College Dublin back in 2018.

There is substantial evidence of the potential for class bias in the use of standardized tests to evaluate college applicants, yet little comparable inquiry considers the written essays typically required of applicants to selective US colleges and universities. We utilize a corpus of 240,000 admissions essays submitted by 60,000 applicants to the University of California in November 2016 to measure the relationship between the content of application essays, reported household income, and standardized test scores (SAT) at scale. We quantify essay content using correlated topic modeling (CTM) and the Linguistic Inquiry and Word Count (LIWC) software package. Results show that essays have a stronger correlation to reported household income than SAT scores. Essay content also explains much of the variance in SAT scores, suggesting that essays encode some of the same information as the SAT, though this relationship attenuates as household income increases. Efforts to realize more equitable college admissions protocols can be informed by attending to how social class is encoded in non-numerical components of applications.

New variants of SARS-CoV-2 are rapidly spreading through the U.S. These coronavirus variants are being studied because they may be more transmissible, cause more severe disease, or reduce the efficacy of current vaccines. Our live tracker keeps you updated with the key information you need to know.

Decades before COVID-19 gripped the world, basic science researchers were laying the groundwork for lifesaving medical and technological breakthroughs now central to the world’s pandemic response. With support from the Kavli Foundation and the Simons Foundation, the Science Philanthropy Alliance has enlisted a team of science writers to explore these science origin stories. Each month, the COVID-19 Basic Science Prequels will unpack the people, history, and serendipitous discovery behind topics that now dominate our daily lives.

The EU has approved an artificial intelligence (AI) technology that can identify people at risk of a fatal heart attack, years before it strikes.

The CaRi-Heart technology, developed by British Heart Foundation (BFH) spinout company Caristo Diagnostics, utilises coronary CT angiogram (CTTA) scans already performed in clinical practice.

It uses AI and deep-learning technology to produce a fat attenuation index score (FAI-Score), which accurately measures inflammation of blood vessels in and around the heart.

When SESP senior Bijal Mehta hosted a meeting on equity and ethics in technology, McCormick junior Mason Secky-Koebel was the only person to show up. That’s when they started talking about artificial intelligence ethics.

Three months later, Secky-Koebel and Mehta founded the Responsible Artificial Intelligence Student Organization to raise awareness around the ethical issues of AI.

Many industries use algorithms to predict certain outcomes in healthcare, hiring opportunities and other important decisions, Mehta said. Prejudices often get “baked into the data,” she said, and are then reproduced in algorithms that may determine who will be best fit for a job or needs healthcare.

Online April 15, starting at 1:30 p.m. Title: Everything you always wanted to know about AI and AI For Good* but were afraid to ask [registration required]

In machine learning, models are commonly estimated via empirical risk minimization (ERM), a principle that considers minimizing the average empirical loss on observed data. Unfortunately, minimizing the average loss alone in ERM has known drawbacks—potentially resulting in models that are susceptible to outliers, unfair to subgroups in the data, or brittle to shifts in distribution. Previous works have thus proposed numerous bespoke solutions for these specific problems.

In contrast, in this post, we describe our work in tilted empirical risk minimization (TERM), which provides a unified view on the deficiencies of ERM (Figure 1). TERM considers a modification to ERM that can be used for diverse applications such as enforcing fairness between subgroups, mitigating the effect of outliers, and addressing class imbalance—all in one unified framework.