Faculty Q+A: Peters’ Important Collaborations with IBM

 

Some of the computer simulations done by Peters with IBM. (Credit: Thomas Peters)

 

By: Eli Freund, Editorial Communications Manager, UConn School of Engineering

Dr. Thomas J. Peters, a professor of computer science and mathematics, who has been with the University of Connecticut for 30 years, has centered his research on computational topology, computer graphics, and scientific visualization, leading to some interesting collaborations.

One of those collaborations, which has been going on for over a decade, has been with IBM, which has brought compelling results. In a Q+A, Peters describes his work and relationship with one of the largest computing companies in the world.

 

  1. When did this collaboration with IBM start? How were you introduced to their University Award Programs?

The collaboration started in 2003 and continues through today, with my first awareness of their university programs in 2005.  I first met my primary IBM collaborator, Dr. K. E. Jordan, when we were co-organizing a SIAM conference, Mathematics for Industry.

  1. Describe your research relating to micelles. What were you trying to figure out and how do the learnings from your research effect the industrial world?

We create novel topological analyses to understand chemical changes that arise from instantaneous changes in shape.  The predictive mathematics is subtle.  Our previously published topological algorithms to predict protein misfolding are being extended to micelles.  The long-term goal is to improve the chemical engineering of micelles for applications for consumers (cosmetics, foods) and industry (petroleum extraction).  Our topological algorithms were previously developed to identify when molecular backbones could self-intersect and fracture, essentially the inverse operation of the step transform, where disjoint micelles merge. 

  1. Why was this work important to IBM?

The micelles are poorly understood and hard to analyze, yet have significant impact in consumer products and industrial applications.  Unilever, which has a significant presence in Connecticut, has considerable interest in this discovery, which IBM can accelerate by innovative simulation algorithms on world-class supercomputers.

  1. In addition to the grants you received from IBM, were there any other major resources you received from them that has been helpful to your research?

I have free access to their OpenPower supercomputers, consistently ranked among the top-five, worldwide.  As an IBM OpenPower Professor, I also provide that access to students.

It is amazing to see their conceptual algorithmic design creativity expand with realization of essentially infinitely many processors, rather than being bound by the limits of a laptop. In 2018-2019, this supercomputer time was conservatively estimated as $250,000, priced at $400 per hour.  The availability of well-curated, reliable data is crucial to advances in data science but is often difficult to obtain.  IBM Research readily provides such simulation data for our computational experiments and analysis validations.  In 2018-2019, this is estimated to have been worth $50,000.  I have enjoyed similar contributions for over a decade.

I was invited to attend IBM-MIT AI Week in Cambridge, MA.  This event was by invitation-only and featured exchanges with world leaders, such as a recent Turing Award winner.  I have already injected the knowledge gained into my research and classes.  Additionally, IBM offers on-line educational resources of their Academic Initiative (educational tutorials), Skills Academy (focused on emerging technologies) and an access program to work on IBM’s Cloud at no cost.  I have already used some of their quantum computing educational materials in my graduate algorithms class, as well as in my own research on `fragile topology’ for modeling quantum superconducting surfaces.

  1. How has IBM been as an industrial partner?

Fantastic. I have been welcomed to their research labs, collaborated freely with their key scientists and fully supported by their administrative staff.  Also, their awards are given in recognition of research accomplishment, not specifically proposed projects, so the money comes with `no strings attached.’  That flexibility is prized for facile response to dynamically changing research conditions. Intellectual property agreements were easily completed and there are no required deliverables.

Award consideration is by nomination, first by my primary collaborator, Dr. Kirk E. Jordan, IBM Distinguished Engineer, Data Centric Solutions; IBM T.J. Watson Research Chief Science Officer; IBM Research UK Member, IBM Academy of Technology.  It was reviewed by a senior executive team, including the IBM Executive Vice President, Dr. John Kelly III, who reports directly to the President and Chairman of IBM Corporation.  It was also reviewed by many of the uppermost technical leaders and executives of IBM, in a rigorous vetting process of 34 major steps.

  1. In your opinion, how important is it for large companies to have programs like IBM’s, which bridge the gap between industry and academia?

I think such relationships are great for both sides, advancing university knowledge and accelerating business goals.  I was formerly a senior scientist at a high-tech Fortune 500 company and advocated strongly for funding interaction with academia.

  1. Beyond your current funding with IBM, do you see this as a long-term partnership, for years to come? If so, what other research would you want to partner with them on?

This has already been a partnership exceeding 15 years, with Faculty Awards, Graduate StudentFellowships, an Open Collaborative Research Award and a Shared University Research Award.  We are planning future work on knot theory, machine learning and quantum computing.

For more information on Peters and his research, please visit: https://tpeters.engr.uconn.edu

DEPARTMENTS

Biomedical Engineering

Chemical & Biomolecular Engineering

Civil & Environmental Engineering

Computer Science & Engineering

Electrical & Computer Engineering

Materials Science & Engineering

Mechanical Engineering

Centers & Institutes