Supercomputing for Humanities, Arts, and Social Sciences

High performance computing (HPC) is for humanists, too. And it is about to get easier for Humanities, Arts, and Social Sciences (HASS) scholars to use supercomputing resources in their work.

This is the message Dr. Alan B. Craig is bringing to the Extreme Science and Engineering Discovery Environment (XSEDE) conference this upcoming week in St. Louis, Missouri.

XSEDE is the human infrastructure that provides seamless integration of supercomputing resources, storage, and expertise. Academics in any discipline have free access to XSEDE resources, though applications are required.  

Why would Humanities, Arts, and Social Science (HASS) scholars want to use HPC?

As Craig, Ph.D., University of Illinois at Urbana-Champaign ’05 (right; photo by Stephen N. Kemp), points out, humans are good at some things and not good at others. For instance, humans excel at interpretation. Supercomputers excel at quickly processing data, and they can help with interpretation. When you bring them together, much bigger projects can be imagined.

Coding skills and other high-tech wizardry are not always required to use supercomputers.

People like Craig offer help, training, and tools to make supercomputing easier. It is easy to get started using HPC. But if researchers experience any problems, they can contact the consultants at XSEDE. If they need more specialized or further assistance, they can apply for hands-on support through Extended Collaborative Support Service (ECSS). And, of course, Craig is always happy to answer questions.

Craig has served as the Digital Humanities Specialist in XSEDE’s Novel and Innovative Projects Group since 2012. An expert in virtual reality/augmented reality, Craig was also the Associate Director for Human Computer Interaction at the Institute for Computing in Humanities, Arts, and Social Sciences (I-CHASS) and a Research Scientist at the National Center for Supercomputing Applications (NCSA).

He’s excited about new tools like Gateways, which make it easier for people without much technical skill to use HPC. Gateways are webpages that serve as user-friendly interfaces between scholars and HPC resources.

HASS scholars tend to use text analysis, video analysis, network analysis, image analysis, and GIS methods. As such, Gateways are being built for the first three methods, and the GIS Gateway already exists.

“The nuts and bolts of supercomputers are not something a lot of HASS scholars think about,” Craig says. “But they go to websites all the time. So we created a website. People can bring their data to us and we can show them how to use the tools.”

Supercomputers at NCSA

Craig became an ambassador to HASS fields through a mix of accident and affinity.

In his early days at NCSA, Craig recalls watching a man walk back and forth in the hallway outside of Craig’s office. Craig went to ask how he could help, and the rest is history.

The man in question turned out to be Orville Vernon Burton, Professor of History and Sociology at the University of Illinois (and the future director of I-CHASS). And he was looking for a card reader. (Card readers had been absent from campus for many years and NCSA didn’t have one.)

“When I asked him how I could help him, he showed me the cards and said, ‘I’m trying to find someone who can help me get my data on the supercomputer,’” Craig recalls with a laugh. “And I helped him find a card reader on campus and get his data read. I kind of became the designated help person, which is great because I love helping people.”

He is also interested in the kinds of questions HASS scholars raise. “I like how HASS scholars get at issues beyond computing. They want to think about why we’re doing what we’re doing and not just that we can do these things. There’s a curiosity there that I relate to,” Craig says.

For instance, scientists and humanists alike can use data mining to create relationships between large sets of information. But HASS scholars will often bring an extra perspective. They want to know how the information was collected and to what additional purposes it can be put.

XSEDE’s supercomputing resources can help HASS scholars ask and answer even larger questions. HPC can drastically change the scale of HASS projects. Even more significantly, HPC can broaden how HASS scholars think about the world they are trying to represent.

In his experience, HASS scholars tend to ask large questions about small datasets.

“I’ve talked with scholars who, just as an example, want to use supercomputers to find the voting records for all of southeast Maryland between 1950-1955. They think they’re thinking big, and on a human scale, on the scale of what just one scholar can collect and read through and analyze, they are thinking big. But I ask them, why not include the entire United States? Does it matter what was happening on the other side of the world at this time? If it does, let’s include that too. The scale of HPC changes everything.”

Supercomputers can do more than count voting logs. “People think that computers only do math," Craig says. "It’s a notion that supercomputers can only handle numeric data. But media data actually requires a lot more storage and computing resources, and supercomputers were built to handle that kind of demand.”

Using HPC, HASS scholars have created and analyzed large collections of video,  tracked representations of musical instruments in paintings across time, and virtually experienced 1920s Harlem.

In the process, they have also expanded what it means to be a HASS scholar. “The Gateways can build communities of HASS scholars, as places where people who are interested in the same tools and data sets can work together,” Craig says.

Much work in HASS fields is still produced by single authors. HPC resources can help expand the single scholar model so that collaboration is included in the process of creating new knowledge.

“To solve real-world problems, the scale has to change. The scale of the questions, the scale of the data, and the scale of the research have to be much bigger. We have to think in teams, where, in order to solve a large problem, content people, technical people, visualization people, and project management people come together,” Craig contends.

He makes clear that, just as HASS scholars can learn a lot with and from XSEDE, so too can XSEDE learn a lot from HASS scholars.

“HASS scholars teach us what barriers people are encountering using these systems, and we remove those barriers. They show us how questions can intersect across fields. And probably most importantly, they reveal different ways of representing the world,” Craig says.

For more information about XSEDE or to request assistance with HPC, please contact Dr. Craig at a-craig@illinois.edu.