Data Science newsletter – June 3, 2021

A smartwatch can signal physiological changes, such as a change in red blood cell count, as well as early signs of dehydration, anemia and illness, according to a new study led by researchers at Stanford Medicine.

The study is among the first to show that smartwatch data correlates with laboratory test results.

Scientists from the lab of Michael Snyder, PhD, professor and chair of genetics, tracked data from smartwatches, blood tests and other tests conducted in a doctor’s office in a small group of study participants. They were curious whether smartwatch readouts, such as heart rate and physical activity, could show physiological changes that are typically revealed through clinical measurements, including blood tests.

The Satcher Health Leadership Institute at Morehouse School of Medicine today launched its Health Equity Tracker, a data platform that highlights the disparate impacts of COVID-19 on marginalized communities.

Built with support from Gilead Sciences, Google.org, the Annie E. Casey Foundation and the CDC Foundation, the tracker is a data visualization tool that displays the scale of COVID-19 cases, deaths and hospitalizations across race and ethnicity, sex and age, from a whole-country view down to the county level.

What’s more, the Health Equity Tracker allows users to view different conditions and determinants that have led to unequal COVID-19 outcomes, including COPD, diabetes, poverty and uninsured rates. It can also be used to compare outcomes in different locations.

Scientists and students with disabilities are often excluded from laboratories — in part because of how they’re designed. Emily Kwong speaks to disabled scientist Krystal Vasquez on how her disability changed her relationship to science, how scientific research can become more accessible, and how STEMM fields need to change to be more welcoming to disabled scientists. [audio, 14:00]

A five-year quest to map the universe and unravel the mysteries of dark energy will officially commence this month as the Dark Energy Spectroscopic Instrument (DESI) begins its main survey to capture and study the light from 35 million galaxies and 2.4 million quasars—the most luminous objects in the universe—across an area of sky in the Northern Hemisphere.

Researchers will use the data collected by DESI to peer back through 12 billion years of the universe’s 13.8 billion-year history. The goal is to unravel more clues to the mystery of dark energy—a hypothetical entity that makes up more than 70 percent of the universe and is hypothesized to be driving the universe’s expansion—and ultimately create the most detailed 3D map of the universe ever made.

DESI is an international science collaboration involving 75 institutions in 13 countries. Among the DESI participants are researchers from the University of Rochester’s cosmology group, a crossdisciplinary group that includes professors, postdoctoral research associates, graduate students, and undergraduates from physics, astronomy, data science, and computer science.

Who hasn’t been stuck seething at an interminable red light with zero cross traffic? When this happened one time too many to Uriel Katz, he co-founded Israel-based, Palo Alto, California-headquartered tech startup NoTraffic in 2017. The company claims its cloud- and artificial-intelligence-based traffic control system can halve rush-hour times in dense urban areas, reduce annual CO2 emissions by a half-billion tons in places like Phoenix/Maricopa County, and slash transportation budgets by 70 percent. That sounded mighty free-lunchy, so I got NoTraffic’s VP of strategic partnerships, Tom Cooper, on the phone.

Here’s how it works: Sensors perceive, identify, and analyze all traffic approaching each intersection, sharing data to the cloud. Here light timing and traffic flow is adjusted continuously, prioritizing commuting patterns, emergency and evacuation traffic, a temporary parade of bicycles—whatever. Judicious allocation of “green time” means no green or walk-signal time gets wasted.

“We have crime-as-a-service, we have AI-as-a-service,” said Philipp Amann, head of strategy at EU law enforcement agency’s Europol’s European Cybercrime Centre. “We’ll have AI-for-crime-as-a-service too.”

Most concerning to cybersecurity officials is deepfake technology — which uses reams of photos and videos to develop uncanny likenesses, or entirely new avatars. The technology has the power to generate pictures and videos that trick people into thinking they’re looking at the real thing, and that’s precisely what cybersecurity experts worry about.

If cybercriminals “manage to come up with ways of assuming your identity or my identity, or create somebody from scratch that doesn’t exist, and they then manage to get through the online verification processes, that’s a huge risk,” Amann said.

So far, about 400 or so colleges plan to require that students who wish to learn in-person receive the Johnson & Johnson, Moderna or Pfizer shots, according to a list compiled by the Chronicle of Higher Education.

Both public and private universities have issued coronavirus vaccine mandates for students, though state colleges in Republican-controlled states have been more likely to eschew such requirements. The American College Health Association, a trade group of college health care providers, recommended colleges require the vaccine for in-person classes where “state law and available resources allow.”

Almost all of the nation’s 4,000 degree-granting colleges, however, are encouraging or helping their students to get their COVID-19 shots. The University of Florida has not issued a vaccine mandate, but has hosted a massive vaccination clinic that aimed to inoculate thousands of students daily. And some colleges are even offering incentives such as cash, university swag or tuition-free courses, according to Inside Higher Ed.

All fish are not created equal, at least when it comes to nutritional benefits.

This truth has important implications for how declining fish biodiversity can affect human nutrition, according to a computer modeling study led by Cornell and Columbia University researchers.

The study, “Declining Diversity of Wild-Caught Species Puts Dietary Nutrient Supplies at Risk,” published May 28 in Science Advances, focused on the Loreto region of the Peruvian Amazon, where inland fisheries provide a critical source of nutrition for the 800,000 inhabitants.

At the same time, the findings apply to fish biodiversity worldwide, as more than 2 billion people depend on fish as their primary source of animal-derived nutrients.

EPFL researchers have developed a device that can continuously measure the blood concentration of propofol – one of the main compounds used in anesthetics – in patients as they’re being operated on. That will help anesthesiologists deliver more personalized doses.

“Because of the rapid pace of innovation in the AI/ML medical device space, and the dramatic increase in AI/ML-related submissions to the Agency, we have been working to develop a regulatory framework tailored to these technologies and that would provide a more scalable approach to their oversight,” FDA spokesperson Jeremy Kahn wrote in an email to MedCity News.

As the FDA drafts up new regulations to evaluate these tools, several big questions loom: How will regulators handle questions about bias and fairness? And what information will patients and their doctors have about these systems?

Getting this right is important. Where a decision by a doctor might only affect a few patients, these algorithms could touch thousands of lives.

The European Commission recently published a proposal for a regulation on artificial intelligence (AI). This is the first document of its kind to attempt to tame the multi-tentacled beast that is artificial intelligence. … Naturally, the European Union does not have jurisdiction over the United Nations, which is governed by international law. The exclusion therefore does not come as a surprise, but does point to a gap in AI regulation. The United Nations therefore needs its own regulation for artificial intelligence, and urgently so.

President Joe Biden’s first budget proposal would give the largest funding boost in nearly two decades to the agency most closely tracking the coronavirus pandemic.

The budget blueprint for fiscal 2022 would include $8.7 billion in discretionary funding for the Centers for Disease Control and Prevention, according to the OMB.

The funding would be used in part for rebuilding “international capacity to detect, prepare for, and respond to emerging global threats,” the OMB said.

168 comments as of June 2, 10 p.m.

IBM and The Grainger College of Engineering at the University of Illinois Urbana-Champaign plan to launch a large-scale collaboration designed to increase access to technology education and skill development, and to combine the strengths of academia and the industrial sector to spur breakthroughs in emerging areas of technology. Specifically, the planned collaboration will focus on the rapidly growing areas of hybrid cloud and AI, quantum information science and technology, accelerated materials discovery, and sustainability to accelerate the discovery of solutions to complex global challenges.

This planned collaboration will be centered in the creation of the new IBM-Illinois Discovery Accelerator Institute, housed within The Grainger College of Engineering at the University of Illinois Urbana-Champaign (UIUC). It will be funded by a ten-year planned research investment from IBM and UIUC, complemented by a major new building project which will house research activities in quantum information, high-performance computing, hybrid cloud and networked environments with support from the State of Illinois, bringing total investments to more than $200 million.

The project was presented at ICLR 2021, the ninth International Conference on Learning Representations, by Reuben [Feinman] who is the primary author on the project, on May 5th. ICLR is a conference focused on the advancement of “representation learning”, primarily through an approach commonly referred to as “deep learning.”

Ultimately, the paper explains how people can learn from rich, general-purpose representations from raw perceptual inputs. Current machine learning approaches do not meet these human standards. Symbolic models, though they have the ability to capture compositional and causal knowledge that allows for flexible generalization, struggle to learn from raw inputs and depend on strong abstraction and simplifying assumptions. Neural network models can learn directly from raw data however they struggle to capture compositional and causal structure and generally must be retrained in order to tackle new tasks.

Online June 10, starting at 12 p.m. Eastern time. “Join us to learn how to apply groundbreaking machine learning technology in your projects, keep growing your skills at the pace of innovation, and boost your career.” [registration required]

Progress in AI relies on researchers’ ability to compare their models’ performance through open, shared benchmarks. Last year, Facebook AI built and released Dynabench, a first-of-its-kind platform that radically rethinks benchmarking in AI, starting with natural language processing (NLP) models. Rather than using static tests, as has been standard in the field, Dynabench uses both humans and NLP models together “in the loop” to create steadily evolving benchmarks that won’t saturate, may be less prone to bias and artifacts, and allow AI researchers to measure performance in ways that are closer to real-world applications.

We’re now announcing a major new capability called Dynaboard, an evaluation-as-a-service platform for conducting comprehensive, standardized evaluations of NLP models. Dynaboard makes it possible for the first time to perform apples-to-apples comparisons dynamically without common issues from bugs in evaluation code, inconsistencies in filtering test data, backwards compatibility, accessibility, and many other reproducibility issues that plague the AI field today.

In this article, you will get a compact overview of MLOps, its stages. You will also get a walkthrough of instances when MLOps is an organizational and communication problem and when it is a tech problem and how to resolve these challenges.