Skip to main content

NCSA Aids Researchers with AI on Multiple Fronts

A digital rendering of carbon chains hovering over a circuit board background. The image is meant to convey the use of AI or supercomputing resources in the field of materials science.

NCSA is the caretaker and manager of the Delta supercomputer, but that means more than just making sure the machine is running. The Center recently worked with the Argonne National Laboratory (ANL) on their research using AI to identify materials capable of passively capturing carbon. Delta’s powerful AI capabilities were used in this collaborative effort, which also involved researchers from the Beckman Institute at the University of Illinois Urbana Champaign (UIUC), the University of Illinois at Chicago (UIC) and the University of Chicago (UChicago), but NCSA’s involvement didn’t end there.

The research team, led in part by Eliu Huerta, formerly of NCSA and currently working at Argonne as a computational scientist, wanted to find a carbon-absorbing material that was cost-effective and sustainable with a low environmental impact. If it sounds like finding such a thing is a needle in a haystack, you aren’t far off the mark. Fortunately, supercomputers are the perfect tool for finding those theoretical needles. In this case, the needle of choice is metal-organic frameworks (MOFs), porous materials capable of absorbing carbon dioxide (CO2).

Utilizing the Polaris supercomputer at the Argonne Leadership Computing Facility (ALCF), the team was able to create as many variations of MOFs as needed to find that special combination with all the desirable attributes. In fact, they found far more than one needle. They found 120,000 possible MOFs that would work in under 30 minutes.

An image of NCSA's Delta supercomputer. The supercomputer is wrapped in a colorful design, the word Delta shifts from violets to oranges and yellows, the colors are in triangular shapes and the A in Delta is the mathematical symbol for Delta.

NCSA’s Delta is a GPU-based supercomputer and will soon be joined by DeltaAI.

They took the most viable MOFs and turned to Delta’s capabilities to run molecular dynamics simulations. These would allow the team to choose only the most promising MOFs – those that fit all the ideal qualities of a sustainable carbon-capturing material. These simulations can take a lot of time to run, especially when you have so many MOFs to test, but Delta’s GPU-based architecture means these simulations can be done much more quickly than the typical CPU counterpart. 

The usage of AI was also instrumental in this research. Hyun Park is the lead author of the resulting publication in Nature Communications Chemistry. Hyun, a graduate assistant at the Beckman Institute, explains, “I believe the key to making successful molecule targets, whether they are MOFs or biomolecules, [is a] tight marriage of [a] diverse set of old and new techniques. … In addition to contributing my expertise in biophysics and AI, I am glad I had hands-on experience working with different scientists on MOF project, merging many fields of science.”

NCSA has long understood the importance of using AI in research like this. The Center has a tradition of providing training and expertise to researchers new to using supercomputers. The NCSA Delta team saw the potential of including researchers in their training who had found success in using Delta and offered to take Park and Huerta’s efforts one step further. Working with Huerta and Park, NCSA supported and hosted training sessions called “AI for Science using Delta.” Presenters in this series of training sessions included experts from NCSA and Park. The sessions were recorded and can be viewed here.

As part of the NCSA training program, we strive to facilitate the seamless integration of AI into scientific exploration, fostering intellectual growth and empowering researchers to maximize the potential of high-performance computing.

–Bruno Abreu, NCSA/IQUIST research scientist and quantum computing co-lead

Bruno Abreu headshot

Using what he’d learned from his work on the MOFs, Park has been able to effectively train other researchers on how to best take advantage of AI in their work using a real-world successful research project as an example.

“I hope that participants in my sessions gained valuable insights to carry forward,” Hyun said. “There are many impactful ways High-Performance Computing (HPC) helped with my research, and I wanted to share with others what I found most useful. There are three main takeaways from my sessions that I think should be shared with researchers: First, I emphasize the accessibility for researchers to elevate their projects through HPC, facilitating scalability in both data volume and computational speed. Secondly, HPC offers a platform for researchers to disseminate their work consistently within a robust computational framework. Thirdly, the sessions shed light on the integration of Artificial Intelligence into molecular sciences leveraging the power of HPC. My goal is that my workshops and lectures contribute to the intellectual growth of as many curious minds as possible.”

Experts at NCSA want to do more than just assist researchers with powerful resources. Part of the Center’s mission is to help researchers find new ways to adapt to these resources. The field of AI is expanding, but for many researchers, it’s a brand-new tool. By demystifying AI through training, NCSA hopes that scientists will be able to do more with the machines they’re given access to.

A picture of Greg Bauer

Part of Delta’s mission is to introduce tomorrow’s researchers to new methods that accelerate their work and provide insight sooner. The ‘AI for Science’ series is just one example of how NCSA is committed to bringing these new insights and methods to the broader research community.

–Greg Bauer, senior technical advisor and co-PI of the Delta project, NCSA

“As part of the NCSA training program, we strive to facilitate the seamless integration of AI into scientific exploration, fostering intellectual growth and empowering researchers to maximize the potential of high-performance computing,” said Bruno Abreu, NCSA/IQUIST research scientist and quantum computing co-lead. “Our training sessions, like ‘AI for Science using Delta,’ aim to bridge the gap between researchers and the transformative capabilities of AI, providing hands-on experience and showcasing innovative research in fields like molecular and materials science, biophysics, chemistry, astrophysics and many others. Delta and DeltaAI are systems designed to take the most out of big data and AI methods, and we want our researchers to take full advantage of that. We find that training sessions that incorporate domain science in the context of using these advanced cyberinfrastructure resources are a great way to provide meaningful learning experiences.”

You can read more about the details of this research in Argonne’s story here: Argonne scientists use AI to identify new materials for carbon capture and in Beckman’s story here: Beckman scientists use AI to identify new materials for carbon capture


NCSA combines next-generation processor architectures and NVIDIA graphics processors with forward-looking user interfaces and file systems to create Delta, a powerful computing and data analysis resource that is part of the national cyberinfrastructure ecosystem through ACCESS. The project partners with the Science Gateways Community Institute to empower broad communities of researchers to easily access Delta and with the University of Illinois Division of Disability Resources & Educational Services and the School of Information Sciences to explore and reduce barriers to access. Delta is funded through NSF OAC 2005572.

Back to top