In an endless sea of information, it’s hard to know where to begin a research journey.

If you are doing text-based research at the Hesburgh Libraries, you can pull books from the stacks, pore over indexes and tables of contents and search hundreds or thousands of pages to compile references.

Or, you can introduce yourself to Eric Lease Morgan, digital initiatives librarian in the Navari Family Center for Digital Scholarship (NFCDS) located in the Hesburgh Library.

Navari Family Center for Digital Scholarship (NFCDS)

The NFCDS provides digital scholarship expertise, access to specialized hardware and software, and technology-enriched spaces for Notre Dame faculty, staff and students in any academic discipline at critical points throughout their research and learning journeys.

With partnerships campus-wide, the NFCDS is home to an interdisciplinary team of faculty and staff who offer workshops, consultations, and connections to collaborators in areas including data use and analysis, digital humanities, geographic information systems, research data management, text mining and analysis, virtual reality, and 3D modeling and printing.

As part of the NFCDS, Morgan teaches faculty, staff and students how to utilize text mining, natural language processing, machine learning and generative AI to read at scale. The process allows individuals to search, extract and analyze information from large volumes of text, ranging from a dozen journal articles to hundreds of books. Text mining helps to accelerate research processes by discovering patterns and anomalies in sets of written documents. The process begins by creating a digitized corpus of materials. The content of the corpus is transformed into a data set to make it easier for both people and computers to read and parse.

Morgan’s bio on the Hesburgh Libraries’ website states that his “professional goal is to discover new ways to use computers to provide better library service.” He has explored vended software, designed proprietary solutions and collaborated with internal and external stakeholders to address the Notre Dame community's growing research and academic needs.

“For more than 40 years, Eric has been pushing the envelope of how we can use technology to best support research, teaching and learning,” said Mark Dehmlow, Associate University Librarian for Digital Strategies and Technology.“Text mining and tools are essential research services and yet another way Eric has embraced innovation to enhance the research experience.”

Distant Reader for text mining and natural language processing

To help advance and add depth to scholarly work, Morgan designed the Distant Reader. This web-based tool allows faculty and students to input large volumes of text, analyze and index the data, curate more focused sets of resources and create topic models.

Since its launch, students have utilized Distant Reader to help supplement the reading process and save time. In recent years,

several students have specifically cited the value of Morgan’s workshops and the Distant Reader in their winning essays for the University of Notre Dame Library Research Award.

“He [Morgan] assured me that I had valuable information to share and that in learning to navigate the work using the Distant Reader, I would come to understand the literature and my research more fully,” 2023 Library Research Award winner Tia Williams said about working with Morgan and using Distant Reader to complete her senior thesis.

Williams wanted to write her thesis on the educational landscape in New Orleans following Hurricane Katrina. She worked with Morgan to find articles across several databases, load them into the program and create a data set that could be saved. From there, Williams narrowed the search to only citations using the term ‘Orleans’ in the title, keywords or abstract.

The process narrowed down 2,737 sources into 586 results, which Williams used as a smaller data set to reference as she completed her thesis.

“I was able to see data such as which keywords were most used, how publications surged in certain years, and even how well the distant reader searches were allowing me to grasp what I was trying to through wordmaps Eric formed,” Williams said. “I was reaching the deeper understanding that he explained I would.”

Study Carrels for curating unlimited data

Looking forward, Morgan is working to make it even easier for future researchers to find the information they seek.

One way he is doing this is by curating what he calls “study carrels.” Like an occupied study carrel at the Hesburgh Library which often has stacks of books and printed articles covering the carrel occupant’s research topic, Morgan’s online study carrels contain books and articles in a digital form of plain text data sets.

In the case of Williams, she had to search databases and upload all of her articles to her private Distant Reader account from scratch. While not all licensed content allows for this type of text analysis, Morgan’s current work involves curating collections that would already be available for individuals to use.

“This library of study carrels are akin to traditional library collections, and they will stand the test of time,” Morgan said. “Like traditional library collections, the content and structure of the study carrels make them independent of any specific computer program or operating system. Heck, you don’t even have to have a network connection for them to work.”

Morgan has approximately 4 million items from the public domain in his library, containing about 3,000 collections of study carrels. That number will continue to grow. In fact, HathiTrust — a large-scale collaborative repository of digital content from research libraries — recently granted Morgan access to their largest dataset of public domain volumes, which he plans to add to the Distant Reader catalog.

Amazon Web Services and high-performance computing power

It takes a lot of computing power to curate 4 million items and over 3,000 study carrel collections, but Morgan doesn’t want to stop there. As part of his individual faculty research, he would like to enhance his growing collections in the Distant Reader and ask, “How might artificial intelligence (AI) be applied in libraries to make research collections more useful?”

He has been able to get the computing power he needs and advance his exploratory efforts thanks, in part, to research credits awarded by Amazon Web Services (AWS). According to their website, since 2006, AWS has provided world-leading cloud technologies that enable organizations and individuals to build solutions that transform industries, communities, and lives for the better.

AWS has several programs supporting that mission where students, educators or institutions can apply for cloud credits to further their research. Earlier this year, Morgan applied for and received AWS credits, which he now uses to further his work.

“To explore how AI could be used in libraries, I am applying three relatively new computing techniques to my collection of study carrels: 1) the creation of large-language models, 2) the fine-tuning of large-language models, and 3) augmenting the discovery process using a technique called retrieval-augmented generation,” Morgan said.

“We use many different tools to facilitate the reading process: Why not use computers?” he said. “Now-a-days, the finding of information is not as difficult as it used to be. Thus, research libraries need to innovate and leverage technology to accelerate the access, the reading, and the analysis of information.”