Mining the Data in Science

A MAJOR CONTRIBUTOR TO GLOBAL WARMING, CONCRETE PRODUCTION HAS NOTHING TO DO WITH GLAUCOMA, ONE OF THE LEADING CAUSES OF IRREVERSIBLE BLINDNESS IN THE WORLD—except perhaps at the University of Miami, where two faculty members hope to show how data science and computing can address critical problems in their respective fields.

An assistant professor of civil, architectural, and environmental engineering, Luis Ruiz Pestana aims to use machine learning to create the first computer model that simulates how concrete deteriorates over time. His ultimate goal: Develop a more durable microstructure for concrete, enabling bridges, buildings, and highways to last centuries, rather than decades.

An assistant professor of ophthalmology at Bascom Palmer Eye Institute, Dr. Swarup Swaminathan is using advanced statistical modeling to comb thousands of patients’ records for clues that will predict which individuals are at greatest risk for rapidly progressing glaucoma—so sight-saving interventions can begin before it’s too late.

Both of their ideas have paradigm-shifting potential but, for now, are budding experiments supported by new interdisciplinary grants from the University’s Institute for Data Science and Computing, or IDSC (pronounced i-disk). Formally launched early last year—just before the novel coronavirus pandemic shut down most of the world—IDSC evolved from one of the University’s most successful experiments, the Center for Computational Science (CCS), with the ambitious mission to transform the University into a global epicenter of data science through research, education, ethics, and workforce training.

“Data is everywhere. Everybody creates it, everybody uses it, and every day it grows in volume, velocity, variety, and veracity,” says Nick Tsinoremas, vice provost for research and computing and the founding director of both CCS and IDSC. “The question is, how can we use data as an asset? How can we extract information and gain insights from complex data sets to solve complex problems? And what are our responsibilities? How do we ensure the data is secure? That we are using it ethically? That we do good with it, that we make a difference in our community—and the world?”

When Tsinoremas, an international leader in computational genomics and bioinformatics, launched CCS in 2007, it was hoped the center would become the hub for the high-performance computing and software engineering needed to elevate the University’s problem-solving research. But nobody knew if it would work. After all, the institution had no advanced computing cyber infrastructure, no culture of sharing resources across disciplines or campuses.

“Everyone was putting their own computing power in their closets or under their desks,” recalls IDSC deputy director Ben Kirtman, professor and director of the Cooperative Institute for Marine and Atmospheric Studies at the Rosenstiel School of Marine and Atmospheric Science, who joined the University and CCS as program director for climate and environmental hazards. “So the whole concept of bringing together resources to produce something greater than the sum of the parts was a new idea. We didn’t know if faculty would embrace that.”

Thirteen years later, thanks to the University’s vision and investments in one of academia’s largest centralized cyber infrastructures and expertise in software applications that support research and data-driven discoveries, CCS’s successor is poised to catapult the University into the center of the data revolution and help propel Miami’s emergence as an international tech hub.

Even amid the COVID-19 pandemic, IDSC launched its new grant program to pair researchers who have big ideas with data scientists, spearheaded education initiatives to meet local workforce needs and promote an understanding of data science among students and the public, and established numerous academic and industry partnerships that are advancing real-time solutions to real-world issues.

And, just as IDSC marked its first anniversary in February, the institute secured a combined $12 million in endowed funds from two philanthropic titans—the John S. and James L. Knight Foundation and Phillip and Patricia Frost—that will enable the institute to attract some of the best and brightest innovators and scholars in artificial intelligence and machine learning as its first full-time faculty.

“Think of this as high-test fuel that will top off our tank and allow us to go further and go faster,” says Jeffrey Duerk, the University’s executive vice president for academic affairs and provost. “One of the most important opportunities for the world right now is to harness the power of data science to understand complicated problems and find great solutions. This accelerates our ability to do that. While the internet powered the last tech revolution, data science, machine learning, and AI will drive the next one. Our partnerships with the Frosts and the Knight Foundation provide us new opportunities to lead this new revolution.”

Recognizing that retaining and attracting talent is key to advancing Miami’s burgeoning technology sector, the Knight Foundation committed a total of $6 million in new and redirected funds to help establish six endowed faculty chairs at IDSC, the second of a planned group of affiliated research enterprises that will fall under the umbrella of the Frost Institutes for Science and Engineering. The Knight gift unlocked an additional $6 million in matching funds from the Frosts, the University’s longtime benefactors who launched the Frost Institutes in 2017 with a $100 million gift.

Now recruiting the initial endowed faculty chairs, Tsinoremas says they will likely include leading experts in smart homes, smart cities, and digital health, which are key areas of IDSC research, along with programs in earth sciences, data ethics, data visualization, communication, and design. And their addition, he says, will not only draw more renowned technological expertise to the University and the community but also help the institution meet its goal of infusing data science throughout the curriculum and ensure that every student, from music to math majors, graduates with a degree of data-savviness—if not a new master’s degree in data science.

In collaboration with various schools and colleges, CCS and IDSC spearheaded the creation of the University’s new master’s degree in data science, which has tracks in technical data science, data visualization, smart cities, and marine and atmospheric science. Launched last fall with 15 students, the master’s degree program attracted more than 50 applicants for this fall.

Also new this fall: the first introductory course in data science for first-year students. Called Data Science for the World, it was developed under the guidance of Mitsunori Ogihara, a professor in the Department of Computer Science and program director for big data analytics and data mining. Now leading IDSC’s workforce development and education initiatives, he has long believed that every college graduate should know how to do basic computer programming and data analysis.

“Today, that is a fundamental skill of a college graduate,” Ogihara says, “because the more data you have, the better-informed decisions you can make.”

But the most vital role of the IDSC chairs will be to bring new ideas and insights to the highly skilled and collaborative data scientists already at the heart of the institute and catalyze research with industry and government partners that will generate a new wave of data-informed practices and solutions to real-world problems—something that is already well underway.

A prime example is IDSC’s collaboration with General Electric Global Research to develop smart technology that will promote healthy aging at home—improving quality of life while reducing health care costs. Among the ideas being explored are apps that would remind seniors to take their medications or sensors that would check their vital signs every day and alert them to potential dangers—like a sudden step down into their garage or high levels of pollen or pollutants outside their home.

“The need for these things was obviously accelerated by the pandemic, but it was also driven by the desire of many aging individuals to not move to assisted living facilities or nursing homes, and, instead, to age gracefully in the comfort of their own homes,” notes Yelena Yesha, visiting distinguished professor and IDSC’s chief innovation officer who, among many initiatives that capitalize on real-time data, is also collaborating on a blockchain project to detect and track fake news by identifying the source in real time.

Today Tsinoremas is confident that the University’s unparalleled infrastructure, which allows realtime analysis, will attract more top talent who can help drive data science research, applications, and training to new heights. Just during the past few years, the University installed Triton, one of the fastest supercomputers in the nation that, customized for the University by IBM, can process artificial intelligence and machine-learning workloads in real time.

The University is also the first to deploy AT&T’s 5G+ and multi-access edge computing technology, which will deliver more data from the internet to wireless devices at a faster pace. And, it has invested nearly $5 million in the University of Miami Laboratory for Integrative Knowledge, a key element of the University’s Roadmap to Our Next Century aimed at nurturing the cross-campus collaborations the University envisioned when it recruited Tsinoremas to launch its fledgling “experiment” 13 years ago.

Now, with IDSC’s own first round of interdisciplinary grants, early-career researchers like Ruiz Pestana and Swaminathan, who both joined the faculty in 2019, have the opportunity to use IDSC’s powerful computation and analytic resources to test their ideas for transforming the production of concrete or identifying patients with aggressive glaucoma—and are eager to become skilled data scientists themselves.

“All of these things that we are funding have potential to be transformative science breakthroughs. Some are going to hit. Some are going to miss. That’s the nature of science,” says Kirtman, who now leads IDSC’s Atmosphere, Ocean, and Earth Science program and predicts that it won’t be long before everybody will depend on data science to do their work. “It wouldn’t surprise me if there is a button, an app on your phone, that you’ll be able to ask certain kinds of questions and is going to use data science to produce the answers.”