The First Symposium on Machine Learning in Science and Engineering (MLSE’18), jointly organized by faculty from the Georgia Institute of Technology and Carnegie Mellon University, was held in Pittsburgh on June 6-8, 2018. Dana Randall, co-director of the Institute for Data Engineering and Science and Professor in the College of Computing at Georgia Tech, and Newell Washburn, Associate Professor in the Mellon College of Science at Carnegie Mellon University, served as co-chairs of the organizing committee. Randall and Washburn designed the new conference to bridge the diverse research areas that benefit from machine learning. Nearly 400 people attended, including over 40 from Georgia Tech. Among the attendees and session presenters were leaders in academia, government, and industry.
In many disciplines, machine learning has revolutionized some areas, such as in biological and biomedical research, while it remains at an early stage in others. “But data science is increasingly critical to almost every scientific and engineering field,” said Randall. “The goal in creating this new conference was to bring together researchers to highlight ways data science has had a major impact within their fields and allow them to share techniques across disciplines.”
The conference brought together leading and junior researchers in a diversity of research areas who are using machine learning to address fundamental and applied problems. Presentations focused on adapting existing machine learning methods to current research, developing new machine learning algorithms specific to science and engineering, identifying new frontiers of research, and starting new collaborations. The broad agenda was supplemented with short courses on useful state-of-the-art tools and methods, taught by experts in machine learning. Short course topics ranged from an accessible overview of machine learning, to adapting methods for discipline-specific uses.
The symposium featured nine interdisciplinary parallel tracks, or sessions. The focus was on the use of machine learning in science and engineering, rather the usual way of organizing by foundations in computer science, statistics and mathematics. The relevant foundational knowledge was then woven through each track. In this manner, the applications wouldn’t be overshadowed by the foundations or technology itself.
David Sherrill, professor in the School of Chemistry and Biochemistry, said “For researchers in different disciplines, it can be difficult to communicate with one another, which gets in the way of making new connections. It can be very intellectually segregated, and a major aim of this conference was to eliminate as many of those barriers to collaboration as possible. For example, the sessions on molecular properties and design featured speakers from Chemistry, Chemical Engineering, Materials Science, Computational Science, and Pharmacy.”
The event included a diversity of research expertise, participation by under-represented minorities, women, and inclusion of a variety of institutional types beyond the traditional university audience.
Each conference track was organized by leading researchers skilled at applying machine learning to their discipline. The topical themes for MLSE’18 included:
- Biomedical Engineering & Health Informatics, organized by Newell Washburn and May D. Wang
- Chemical Engineering, organized by Zachary Ulissi and AJ Medford
- Chemistry, organized by David Yaron and David Sherrill
- Civil and Environmental Engineering, organized by Mario Berges and James Tsai
- Electrical and Computer Engineering, organized by Radu Marculescu and Justin Romberg
- Engineering and Public Policy, organized by Alex Davis, Kaye Husbands Fealing and Omar Asensio
- Materials Science & Engineering, organized by Elizabeth Holm and Surya Kalidindi
- Mechanical Engineering, organized by Albert Presto and Thomas Kurfess
- Physics, organized by Manfred Paulini and Deirdre Shoemaker
Students were given an opportunity to learn about data-driven approaches to science and engineering research and to present their own results at a poster session. The National Science Foundation provided merit-based travel awards for 29 student applicants from 12 universities.
The organizers of MLSE’18 plan to publish proceedings in a special issue of the Journal of Computational Chemistry, with David Sherrill serving as editor.
Other co-organizers from both institutions aided the conference planning:
- From Georgia Tech: Justin Romberg (College of Engineering), David Sherrill (College of Sciences), and Deirdre Shoemaker (College of Sciences)
- From Carnegie Mellon: Elizabeth Holm (Materials Science and Engineering), Rachel Mandelbaum (Physics), Diana Marculescu (Electrical and Computer Engineering), Barnabas Poczos (Machine Learning), and Aarti Singh (Machine Learning)
MLSE’18 benefited from several sponsors. The National Science Foundation and the National Institute of Standards and Technology funded student travel support. Other sponsors included Google, IBM, Intel, Citrine, Covestro, Johnson Controls International, ExxonMobil, Briston-Myers Squibb, Open-Eye Scientific, the Journal of Chemical Physics, the Provost’s Office at Carnegie Mellon, and the Institute for Data Engineering and Science at Georgia Tech.
Next year, the conference will be held at Georgia Tech and hosted by the Institute for Data Engineering and Science (IDEaS). The event organizers plan to expand the program to include two additional workshops, including a new Women in Data Science Workshop, and a session dedicated to the future vision of machine learning in research.
For further event details, view the online program.
Event Co-chairs Dana Randall of Georgia Tech, and Newell Washburn of Carnegie Mellon University, address the conference participants.
Surya Kalidindi, Professor in the George W. Woodruff School of Mechanical Enginering at Georgia Tech, presents “Data Analytics for Mining Process-Structure-Property Linkages for Hierarchical Materials" at MLSE'18.
Kaye Husbands Fealing, Chair of the School of Public Policy at Georgia Tech, presents “Assessing Scientific Outcomes from Federal Funding of Food Safety Research Using Unstructured Data Techniques” at MLSE'18.
Students from the University of Maryland present a poster on the use of machine learning in diagnosing Parkinson's Disease at MLSE'18.
Xiaowei Yue, a Ph.D. student in the H. Milton Stewart School of Industrial and Systems Engineering at Georgia Tech, presents a poster at MLSE'18.