Data Science newsletter – September 28, 2021

Newsletter features journalism, research papers and tools/software for September 28, 2021

 

Apple looks to digital biomarkers for features detecting depression, cognitive decline

MobiHealthNews, Laura Lovett


from

Apple is looking to use digital biomarkers to help detect depression and early-stage cognitive decline, according to a new report out of The Wall Street Journal.

The end goal, according to the Journal, is to create a new Apple feature that would tell users if there was a potential mental illness. Data collected regarding a users’ mobility, physical activity, sleep pattern could be included in the algorithm.

The report notes that the new effort could be tied to collaborations that the Silicon Valley titan already has established with UCLA researchers and pharma company Biogen.


MIT’s toolkit lets anyone design their own muscle-sensing wearables

Engadget, S. Dent


from

MIT has unveiled a new toolkit that lets users design health-sensing devices that can detect how muscles move. The university’s Science and and Artificial Intelligence Laboratory (CSAIL) created the kit using something called “electrical impedance tomography” (EIT), that measures internal conductivity to gauge whether muscles are activated or relaxed. The research could allow for wearables that monitor distracted driving, hand gestures or muscle movements for physical rehabilitation.

In a paper, the researchers wrote that EIT sensing usually requires expensive hardware setups and complex algorithms to decipher the data. The advent of 3D printing, inexpensive electronics and open-source EIT image libraries has made it feasible for more users, but designing a wearable setup is still a challenge.


Can Wearable Sweat Lactate Sensors Contribute to Sports Physiology?

ACS Sensors journal


from

The rise of wearable sensors to measure lactate content in human sweat during sports activities has attracted the attention of physiologists given the potential of these “analytical tools” to provide real-time information. Beyond the assessment of the sensing technology per se, which, in fact, has not rigorously been validated yet in controlled conditions, there are many open questions about the true usefulness of such wearable sensors in real scenarios. On the one hand, the evidence for the origin of sweat lactate (e.g., via the sweat gland, derivation from blood, or other alternative mechanisms), its high concentration (1–25 mM or even higher) compared to levels in the blood, and the possible correlation between different biofluids (particularly blood) is rather contradictory and generates vivid debate in the field. On the other hand, it is important to point out that accurate detection of sweat lactate is highly dependent on the procedure used to collect and/or reach the fluid, and this can likely explain the large discrepancies reported in the literature. In brief, this paper provides our vision of the current state of the field and a thoughtful evaluation of the possible reasons for present controversies, together with an analysis of the impact of wearable sweat lactate sensors in the physiological context. Finally, although there is not yet overwhelming scientific evidence to provide an unequivocal answer to whether wearable sweat lactate sensors can contribute to sports physiology, we still understand the importance to bring this challenging question up-front to create awareness and guidance in the development, validation, and implementation of wearable sensors. [full text]


Toward a smarter electronic health record

MIT News


from

Electronic health records have been widely adopted with the hope they would save time and improve the quality of patient care. But due to fragmented interfaces and tedious data entry procedures, physicians often spend more time navigating these systems than they do interacting with patients.

Researchers at MIT and the Beth Israel Deaconess Medical Center are combining machine learning and human-computer interaction to create a better electronic health record (EHR). They developed MedKnowts, a system that unifies the processes of looking up medical records and documenting patient information into a single, interactive interface.

Driven by artificial intelligence, this “smart” EHR automatically displays customized, patient-specific medical records when a clinician needs them. MedKnowts also provides autocomplete for clinical terms and auto-populates fields with patient information to help doctors work more efficiently.


A blast into a clean energy future? Scientists tout nuclear fusion breakthrough

The San Diego Union-Tribune, Rob Nikolewski


from

With a powerful zap from nearly 200 laser beams that lasted less than 9 billionth of a second, scientists in Northern California believe they have achieved a major milestone in their research to advance nuclear fusion.

News of last month’s experiment has circulated quickly around the scientific community and stirred optimism that hydrogen fusion may someday provide an abundant and clean source of energy around the globe.

“People have been working at this for decades trying to achieve this,” said Annie Kritcher, a physicist at the Lawrence Livermore National Laboratory and the lead designer for the experiment. “I think it has extremely energized the whole community.”


Datasheets for Datasets help ML engineers notice and understand ethical issues in training data

ACM CSCW, Karen Boyd


from

Bias, privacy, and harmful use are just a few of the potential ethical problems accompanying training data for machine learning algorithms. Along with researchers, ML engineers themselves are concerned and see the need for technical and work practice interventions that defend against the quickly-evolving and often-hidden ethical problems in training data.

In this paper, I tested one such intervention, called Datasheets for Datasets. Datasheets are files accompanying datasets, designed in part to help ML engineers notice potential ethical issues in unfamiliar training data by documenting the dataset’s context. Unlike similar interventions, Datasheets have three important features: they focus on training data (rather than already trained models), they are general purpose (and can be used for any data type and any ML technique), and they are written in plain language. This means that 1) they intervene in training data (and fairly early on in the algorithm development process) 2) they can be used by a wide variety of firms and taught in introductory ML classes to students who will go on to work just about anywhere and 3) they can be understood and evaluated by non-expert stakeholders, like managers, users, citizens, auditors, and policy-makers.

I wanted to know whether Datasheets for Datasets achieve one of their most ambitious goals: do they help engineers notice potential ethical issues in data they have never seen before? A


University of Oxford Launches Podium Analytics Institute for Youth Sports Medicine and Technology

Science X


from

Oxford University has been selected as the home of the new Podium Analytics Institute for Youth Sports Medicine and Technology. This will be the world’s first academic Institute focused on young athletes’ safety and lifelong health and will combine Oxford’s longstanding tradition in sports and education with the very best of science, medicine, and technology.

The new Institute will shift the traditional emphasis of research into sports injury—which is predominantly adult-centric and based upon treatment—by concentrating on younger athletes, 11-18 years old, and will focus on prevention rather than cure.


Tree Thinking

Places, Shannon Mattern


from

As trees become data points, they are all too readily cast as easy fixes for profound problems. Trees as tools of carbon capture, tall timber as an instrument for sustainable construction, green barriers as sound buffers along roadways: sylvan solutions to systemic snafus. The media scholar Jennifer Gabrys argues that such approaches are efforts to frame (and tame) hard problems — wicked problems — in computational terms. Forest data sets in particular, she writes, tend to “present the problem of environmental change through … metrics that in turn legitimate specific technological interventions to meet targets for averting environmental catastrophe.” 5 In other words, these technological tools promote techno-solutionist responses to problems that are simultaneously ecological, cultural, social, economic, and political. 6

It’s easier to plant a tree — and to allow a generative design dashboard to tell you precisely where to plant it — than it is to change our individual and collective consumption habits or to muster the political will to eliminate fossil fuels. In 2019, a research team at ETH Zürich mapped the potential global tree canopy and discovered that the world could accommodate an additional 0.9 billion hectares of canopy cover, which could store over 200 gigatons of carbon; as a member of the team told The Guardian, “This new quantitative solution shows [forest] restoration isn’t just one of our climate change solutions, it is overwhelmingly the top one.”


Geisinger awarded $5 million to develop diagnostic tool for genetic disorders

Geisinger news releases


from

A team of Geisinger researchers has been awarded a $5 million grant from the National Institute of Health’s National Human Genome Research Institute to develop a tool that will allow healthcare providers to diagnose a genetic basis for select medical conditions in real time.

Determining that a medical condition has a genetic basis can have a significant impact on the course of treatment. The proposed High Impact Phenotype Identification System (HIPIS) will shorten the time between onset of symptoms and discovery of a genetic basis for 13 medical conditions, improving patient care and outcomes.


Lab team one of 10 awarded $26 million DOE grant for data science

Los Alamos National Laboratory, News Releases


from

Principal investigator Kipton Barros of Theoretical Division will lead a multidisciplinary team of researchers on a three-year, $2.4 million project in partnership with the California Institute of Technology (Caltech) to use data science — including artificial intelligence and machine learning (AI/ML) — to advance the understanding of chemical and materials systems.


Three Universities Team for NSF-Funded ‘ACES’ Reconfigurable Supercomputer Prototype

HPC Wire


from

As Moore’s law slows, HPC developers are increasingly looking for speed gains in specialized code and specialized hardware – but this specialization, in turn, can make testing and deploying code trickier than ever. Now, researchers from Texas A&M University, the University of Illinois at Urbana-Champaign and the University of Texas at Austin have teamed, with NSF funding, to build a $5 million prototype supercomputer (“ACES”) with a dynamically configurable smörgåsbord of hardware, aiming to support developers as hardware needs grow ever more diverse.


New LawTech Center To Focus on Data, Online Policy

University of Virginia School of Law


from

The University of Virginia School of Law is launching a new scholarly center that will focus on pressing questions in law and technology, with Professor Danielle Citron serving as inaugural director.

The LawTech Center will address policy issues, legal texts as data and the use of tech in the legal profession.

“While some of the topics may be focused on matters at the heart of public debate — like the legal responsibilities of online platforms, privacy and security legislation, or algorithms used in criminal sentencing — others may be dedicated to educating students, academics, lawyers and the public about the laws and policies governing networked technologies,” Citron said. “We also have a tradition at UVA of using empirical methods to examine legal texts and teaching about how lawyers and the legal system use technology.”


How white scholars are colonizing research on health disparities

STAT, Usha Lee McFarling


from

Fueled by the massive health disparities exposed by the coronavirus pandemic and the racial reckoning that followed the murder of George Floyd, health equity research is now in vogue. Journals are clamoring for it, the media is covering it, and the National Institutes of Health, after publicly apologizing for giving the field short shrift, recently announced it would unleash nearly $100 million for research on the topic.

This would seem to be great news. But a STAT investigation shows a disturbing trend: a gold rush mentality where researchers with little or no background or training in health equity research, often white and already well-funded, are rushing in to scoop up grants and publish papers. STAT has documented dozens of cases where white researchers are building on the work of, or picking the brains of, Black and brown researchers without citing them or offering to include them on grants or as co-authors.


The smart toilet era is here! Are you ready to share your analprint with big tech?

The Guardian, Emile Saner


from

For the past 10 years, Sonia Grego has been thinking about toilets – and more specifically what we deposit into them. “We are laser-focused on the analysis of stool,” says the Duke University research professor, with all the unselfconsciousness of someone used to talking about bodily functions. “We think there is an incredible untapped opportunity for health data. And this information is not tapped because of the universal aversion to having anything to do with your stool.”

As the co-founder of Coprata, Grego is working on a toilet that uses sensors and artificial intelligence to analyse waste; she hopes to have an early model for a pilot study ready within nine months. “The toilet that you have in your home has not functionally changed in its design since it was first introduced,” she says, in the second half of the 19th century. There are, of course, now loos with genital-washing capabilities, or heated seats, but this is basic compared with what Grego is envisaging. “All other aspects of your life – your electricity, your communication, even your doorbell – have enhanced capabilities.”


Weighing wastewater’s worth as a COVID-19 monitoring tool

Chemical & Engineering News, Celia Henry Arnaud


from

More than a year into surveilling sewage for outbreaks, scientists have a better idea of what the approach can—and can’t—tell us about the health of communities


Digging Deeper: Wastewater Testing and Air Monitoring

University of California-Davis, UC Davis News


from

Heather Bischel, an assistant professor of civil and environmental engineering, and David Coil, a project scientist with the UC Davis Genome Center, co-lead an initiative with Healthy Davis Together, or HDT, to collect regular samples from wastewater on campus and around the city of Davis. They also ran a pilot program earlier this year to swab stand-alone HEPA air filters inside five elementary schools in Davis.


Snake Venom Identification via Fluorescent Discrimination

Analytical Chemistry; Fei Chen, Meng Qin, Wei Liu, Fan Wang, Wanjie Ren, Huihua Xu, and Fengyu Li


from

The identification and discrimination of snake venom are highly desired for timely clinical treatment. However, the complex components in snake venom make it a great challenge to achieve rapid and accurate identification. Inspired by the organism’s taste sensing system, a fluorescent sensor array that could differentiate snake venoms was fabricated. The interaction of snake venoms with different fluorescent dyes in the sensor array gave rich information, based on which efficient detection of complex snake venom was achieved. The main six proteins of snake venom in the same concentration, different concentrations, and their mixtures were identified with 100% accuracy. Furthermore, seven snake venoms belonging to different snake families were discriminated in PBS buffer and human plasma. Interferents of bovine serum albumin (BSA), thrombin, and transferrin (TRF) demonstrated the practicability of the fluorescent sensor array. This strategy of a multiresponse sensor array provides an effective method for accurate and rapid venom toxicology analysis, benefiting early and timely clinical diagnosis and treatment.


Georgia Tech experts: Online learning doesn’t have to be inferior

AJC.com, The Atlanta Journal-Constitution, Maureen Downey


from

Two prominent Georgia Tech computer scientists maintain online classes can be as effective as face-to-face instruction and urge expansion of opportunities that liberate students from the confines of a classroom and an 8 a.m. to 3 p.m. daily schedule.

The ambitious vision of online education in the new book “The Distributed Classroom” by David A. Joyner and Charles Isbell could be a tough sell. Along with hearing from lots of K-12 parents this year about their children’s disappointing virtual experiences during the COVID-19 pandemic, many parents in Georgia Tech forums bemoaned the migration to remote learning and contended their college kids were essentially teaching themselves.


Why government technologists love the Domino’s pizza tracker

StateScoop, Benjamin Freed


from

As a native Chicagoan, Scott Jensen has some opinions on pizza. The former director of the Rhode Island Department of Labor and Training grew up partial to a local purveyor of deep pan and crispy thin-crust pies.

But one night in spring 2020, flummoxed by the COVID-19 pandemic’s overwhelming of the unemployment insurance system he ran — and far from the Midwest — Jensen ordered from a brand more associated with stoned college kids: Domino’s. At the time, about 200,000 Rhode Islanders had lost their jobs due to the early waves of the pandemic, and as in nearly every other state, the unemployment insurance system was breaking under the stress of record-busting demand and limited processing capacity.

Jensen didn’t order Domino’s for the test-lab-perfected pizzas, wings or cheesy bread. He did it to test out the 18,000-location-strong chain’s delivery tracker. He wanted to see if it might offer a hint of how to fix Rhode Island’s faltering unemployment benefits at the peak of the crisis.


Events



The 5th annual #CSforALLSummit is in less than a month!

Twitter, CSforAll


from

“Join us virtually on October 19 and 20 as we announce the 2021 Commitments made by our member community!”


Deadlines



NFL Big Data Bowl 2022

“Your challenge is to generate actionable, practical, and novel insights from player tracking data that corresponds to special teams play.” Deadline for submissions is July 7.

Careers


Tenured and tenure track faculty positions

Assistant Professor in Statistical and Data Sciences: Biostatistics



Smith College, Program in Statistical and Data Sciences; Northampton, MA

Assistant/Associate Professor of Technology Policy



Duke University, Sanford School of Public Policy; Durham, NC
Full-time, non-tenured academic positions

SHOW Laboratory Manager



University of Wisconsin, School of Medicine and Public Health; Madison, WI

Leave a Comment

Your email address will not be published.