Turning to Machine Learning for Molecular, Materials Research

NSF-funded workshop hosted by Lehigh’s Institute for Data, Intelligent Systems, and Computation (I-DISC) is first in a series on applying data science techniques across disciplines.

Lehigh student Bowen Li presents

Lehigh student Bowen Li presenting his poster on "Balancing exploration and exploitation to design compact training sets for data-driven molecular property models."

The workshop was over. Rows of brown bag lunches were lined up and ready to be taken from a conference table covered in a black tablecloth. A bus was waiting outside.

Still, participants at the event titled “Foundational & Applied Data Science for Molecular and Material Science & Engineering” lingered, talking in small groups in Iacocca Hall’s Wood Dining Room on Lehigh’s Mountaintop Campus. It was exactly the scene the workshop was meant to generate.

“There was a common thread here of machine learning and data science, but the workshop brought together a diverse range of fields, which gave people the opportunity to engage with those they wouldn’t ordinarily encounter,” says Srinivas Rangarajan, an assistant professor of chemical and biomolecular engineering at the P.C. Rossin College of Engineering and Applied Science. “And that was the goal of the event—to bring together top experts from a range of disciplines to share the latest techniques as well as the challenges in machine learning.”

Alex Dunn

Alex Dunn, currently a postdoc at Lawrence Berkeley National Lab, presenting his talk titled "Software Tools for Accelerating Materials Discovery with Machine Learning."

The three-day event started May 22 and was the first in a series of conferences and lectures funded by an NSF TRIPODS-X grant awarded to Lehigh’s Institute for Data, Intelligent Systems, and Computation (I-DISC). Future topics will include applications of machine learning and big-data science in chemical processes, autonomous robotics, supply chain optimization and cognitive neuroscience.

Late in 2018, the NSF unveiled its support for I-DISC to organize these open workshops, part of a broader funding announcement around $8.5 million in grants for 19 collaborative projects involving 23 U.S. universities. Two of the grants were awarded to Lehigh.

The inaugural workshop brought together computer scientists, applied mathematicians, material scientists, chemists, biologists, and chemical, industrial and bioengineers. Through a series of presentations, poster sessions, and networking opportunities, attendees focused on recent developments and application of data science algorithms and tools.

Speakers came from academia, national laboratories and industry. Paulette Clancy, head of the chemical and biomolecular engineering department at Johns Hopkins University, gave a plenary talk on “Merging Physical Science and Machine Learning to Tackle Complexity and Combinatorics in Materials Processing.”

Paulette Clancy

Paulette Clancy, professor and head of Chemical & Biomolecular Engineering at Johns Hopkins University, gave a plenary talk on "Merging Physical Science and Machine Learning to Tackle Complexity and Combinatorics in Materials Processing."

By all accounts, the event, organized by Rangarajan, fellow Lehigh engineering faculty members Jeetain Mittal and Joshua Agar, and Payel Das of IBM Thomas J Watson Research Center, was a success.

“Attendees were impressed with the beauty of Lehigh’s campus, the spectacular setting for the conference, the list of speakers, and the diversity and range of topics,” says Mayuresh V. Kothare, chair of the Department of Chemical and Biomolecular Engineering. “It’s an example of how I-DISC is promoting new knowledge by creating new networks of professionals in these complex domains.”

The second workshop in the series, scheduled for early Fall 2019, will focus on machine learning topics related to robotics, automated control, and dynamical systems.

Story by Christine Fennessey

Photos by Kate Duffy Photography

About I-DISC

The Institute for Data, Intelligent Systems, and Computation (I-DISC) at Lehigh is devoted to the study of problems that involve massive amounts of data and/or large-scale computations, and developing the science that enables the extraction of useful and actionable information across disciplines and research fields. The analysis of complex and massive datasets and the development of sophisticated computational models are essential to our understanding and prediction of complex phenomena and systems associated with personalized medicine, healthcare delivery, transportation systems, social networks, the human brain, global climate, and international economic development. I-DISC builds upon the foundation of Lehigh research expertise in areas such as machine learning, optimization, probabilistic modeling, data-driven decision making, high-performance and data-intensive computing, statistical signal and image processing, data representation and management, modeling and simulation, robotics and computer vision, business and management technology, and privacy and security. Teams of researchers combine fundamental data and computational approaches with those focused on critical applications, utilizing core high-performance computing and cyber–physical systems facilities, creating a fertile space for collaboration with industrial, academic, and governmental partners to attack some of the most pressing problems in technology and society.


Related Stories

A solar panel field

Mapping the Energy Transport Mechanism of Chalcogenide Perovskite for Solar Energy Use

Lehigh engineers  have characterized the thermal energy conversion mechanism in the lattice of an advanced nanomaterial called chalcogenide perovskite and demonstrated its 'tunability'―important for its potential use in solar energy generation.

How airplane noise affects fetal health

Muzhe Yang: How Airplane Noise Affects Fetal Health

Yang examines the unintended consequences of an air traffic modernization project on babies’ birth weight.

Lehigh professor Hannah Dailey

Engineering Orthopaedic Care

Hannah Dailey ’02 ’06G ’09G seeks to bridge the gap in clinicians’ ability to predict how quickly bones heal.