Latest Stories
Vanderbilt University computer science students have developed an accurate, reliable and cost-effective solution for segmenting and labeling television news broadcasts, making them more discoverable and accessible to researchers. This new tool, the first created through the Vanderbilt Cloud Innovation Lab for Libraries and Applied Digital Preservation, powered by AWS, uses artificial intelligence to improve upon manual processes and may reduce costs up to 80 percent.
Launched last October by the Jean and Alexander Heard Libraries and Amazon Web Services, the Vanderbilt Cloud Innovation Lab, or VCIL, provides immersive learning experiences for students interested in using AI and cloud technologies to develop and deploy advancements in digital preservation. It’s one of five such cloud innovation centers in North America powered by AWS and the first housed within an academic library.
The five students in VCIL’s first cohort applied what they’ve learned in the classroom, including computer vision and UX design, to a real-world digital preservation problem—in this case, how to improve the labor-intensive process of arranging the Vanderbilt Television News Archive, one of the largest collections of TV broadcast news in the world. This unique Heard Libraries resource has been recording broadcasts of the U.S. national networks continuously since Aug. 5, 1968, and holds approximately 65,000 hours of content, including 1 million news segments and 500,000 commercial breaks.
The VCIL-developed tool utilizes AI to transcribe video of these broadcasts, break up the content into time-stamped segments, and label each segment by type, such as news report, commercial or teaser. The accuracy of the tool is currently outperforming manually curated data, a statistic of which the team is especially proud, said Erin Geier, the students’ adviser and associate cloud engineer in the Heard Libraries’ Digital Lab, where VCIL is located.
“The students were given the opportunity to experiment and design, innovate and think through the entire process of implementing a cost-effective full-stack project that has immediate community impact,” Geier said. “We will put this proof-of-concept solution into production this summer for the TV News Archive and publish an open-source version that can be used by other institutions with extensive media collections.”
In addition to applying classroom knowledge in a practical context, the first VCIL project gave students the opportunity to develop valuable professional skills. Under the mentorship of AWS and library experts, they worked as a team to conduct rapid prototyping sprints, mirroring technology industry practices. They collaborated directly with stakeholders, building skills in communication, project management and experiential problem solving. And they contributed to an initiative that not only has immediate impact on digital preservation but expands access to our cultural heritage.
“It’s been an amazing experience,” said Rithika Thambireddy, a student cloud developer and rising junior majoring in computer science and mathematics. “One of the things I really enjoyed about the lab was its collaborative nature. I worked alongside four other undergraduate students, and we would come together to discuss the problem, develop our own methodologies, and approach it from different angles. Then we would check in throughout the process to share solutions and ask, ‘What tools did you use? How can we improve efficiency?’
“This experience has taught me to push myself forward, to tap into different resources, and to use AI as a tool to develop theoretical knowledge beyond the classroom,” she said. Thambireddy and her colleagues will represent VCIL and live demonstrate their tool at the AWS Summit in Washington, D.C., June 30–July 1 and continue to share outcomes of their work at showcases this fall and in spring 2027.
For future iterations of their tool, the students plan to expand into automated video clipping and metadata generation, as well as develop a cloud-based collection management tool for archivists. VCIL will continue to focus on scalable, cost-effective and sustainable solutions for libraries while growing its student teams to support a broader portfolio of projects.
“Research libraries are entering a period where large-scale audiovisual collections will require new computational approaches to access and stewardship. What these students have developed demonstrates both the intellectual seriousness of the work and its potential applicability well beyond Vanderbilt,” University Librarian Jon Shaw said.
“The Vanderbilt Cloud Innovation Lab’s first cohort exemplifies what’s possible when students are empowered with AWS technology and the freedom to innovate,” said Noah Eden, global lead for AWS Cloud Innovation Centers. “These students didn’t just learn about AI, they used it to solve a meaningful challenge for Vanderbilt. By open-sourcing their work, they’re enabling other organizations to adopt and leverage their solution. This is exactly the kind of hands-on, experiential learning that prepares the next generation of builders for their careers, and we’re proud to support their work.”