Projects

At Kennesaw State University, the newly established Saga Labs are preparing an undergraduate research project known as OWLLLM, which stands for "Optimizing Word Learning via Large Language Models." This innovative project introduces junior researchers to cutting-edge Large Language Model (LLM) technology. This year's focus is on leveraging the high impact practice (HIP) of using high-frequency vocabulary to facilitate modern language learning.

The effort is funded by the Office of Undergraduate Research and the Norman J Radow College of Humanities and Social Sciences. As AI and natural language processing technologies evolve at a rapid pace, our team is delving into how these advancements can be harnessed to improve language learning outcomes and experiences.

Our diverse team of undergraduate researchers, with backgrounds in linguistics and literature, are engaged in a comprehensive research process. Their responsibilities include conducting in-depth background research on AI/LLMs, reviewing existing literature on AI/LLMs applied to language learning, brainstorming and prototyping AI applications for language interactions, and, time permitting, user testing and evaluation of these prototypes. The project culminates in the analysis and presentation of findings, including preparation for conference presentations.

Participants in this project are not only delving into a cutting-edge research area but also gaining a wealth of experience and skills. These range from project management and software development to language assessment, critical thinking, communication, and collaboration. Students will emerge with a deepened understanding of LLM/AI terminology, insights from literature reviews, and the ability to articulate the study’s contributions to the field. Moreover, they will hone skills in designing studies, developing hypotheses, writing research papers, and presenting their research to audiences.

The research lab receives support in the form of stipends from Kennesaw State University's First-Year Scholars Program, as well as travel and research support through the Norman J. Radow College of Humanities and Social Sciences. Our lab is most grateful for the generous sponsorship and mentorship from industry leaders such as Microsoft, OpenAI, Luminint, and IBM.

Our project, "Artificial Intelligence (AI) and Large Language Models (LLM) as Learning Supports for Second Language Acquisition (SLA)," represents a groundbreaking fusion of contemporary linguistics and AI technology. Situated at the cutting edge of educational innovation, this initiative is a proud part of Microsoft Research's Accelerate Foundation Models Research (AFMR) program. We are committed to reshaping the approach to second language acquisition, and we extend our heartfelt thanks to Microsoft and OpenAI for their substantial support in this endeavor.

By leveraging the transformative potential of AI and LLMs, our interdisciplinary research team seeks to redefine the paradigms of language learning. Our approach intertwines the intricacies of human language processing with the sophisticated capabilities of AI, aiming to unlock new methodologies that are not only more effective but also aligned with individual learning styles and cultural contexts.

Our project endeavors to:

1. Enhance Linguistic Understanding: Utilize the advanced capabilities of LLMs to provide nuanced language learning experiences, tailored to diverse learner profiles.
2. Bridge Cultural Divides: By integrating AI into language learning, we aim to foster greater intercultural communication, understanding, and empathy.
3. Facilitate Accessible Learning: Leveraging Microsoft Azure OpenAI's robust platform, we strive to democratize language learning, making it accessible and equitable for learners across the globe.
4. Advance Scientific Inquiry: Our research contributes to the broader goals of AFMR by pushing the boundaries of what's possible in natural language processing and cognitive linguistics.

In collaboration with a global network of scholars and technologists, our project not only represents a step forward in educational technology but also embodies a commitment to the harmonious integration of AI with human values, sustainability, and creativity. Join us as we embark on this exhilarating journey to reshape the future of language learning through the lens of AI and Large Language Models.

The DEFA Film Library at the University of Massachusetts Amherst, a premier archive and research center for films from the former German Democratic Republic (GDR), is embarking on an exciting new collaboration with the Saga. This partnership aims to enhance the accessibility and reach of East German cinema through state-of-the-art subtitling and translation services.

The DEFA Film Library is one-of-a-kind. The library houses an extensive collection of film materials, including 35mm and 16mm prints, DVDs, books, periodicals, and articles. This diverse collection provides a rich resource for researchers, teachers, producers, and film programmers, as well as the general public interested in the cinematic history and culture of East Germany. In addition to its archival and research functions, the DEFA Film Library also offers a range of services, including making films available for instructional purposes, and training doctoral students in language and literature to deliver industry-grade subtitling for foreign language media.

In 2024, our lab pioneered a process to create highly accurate subtitles and enhanced accessibility features for the DEFA Film Library’s extensive collection of DDR media content. This initiative is part of our broader mission to make a significant impact in areas where our capabilities are most needed. Our goal is to support the DEFA Film Library's mission by making its rich repository of East German films more accessible to a wider audience. By providing precise and culturally nuanced subtitles, we aim to open up these cinematic treasures to researchers, educators, students, and film enthusiasts who may not be fluent in German. This will not only enhance the educational value of the library’s collection but also contribute to the global appreciation and understanding of East German film culture.

This partnership represents an intersection of technology, language, and cultural preservation, and we believe it will yield significant benefits for both our institutions and the wider communities of film scholars and German cinema fans.

In an innovative fusion of technology and education, the Saga team, in collaboration with Union Grove Middle School in Henry County, Georgia, is pioneering a transformative literacy project. This initiative leverages the power of artificial intelligence to revolutionize reading comprehension and engagement for students competing in the prestigious Helen Ruffin Reading Bowl.

The cornerstone of this project is an advanced AI assistant, meticulously developed by our specialized research lab. Utilizing the cognitive services of Microsoft Azure, this conversational AI agent is designed to offer precise, reliable responses to student inquiries about the diverse range of novels in their reading list. This tool is expertly crafted to deepen literary understanding while safeguarding against the inaccuracies typically associated with large language models.

Our collective ambition extends beyond the confines of Henry County. We envisage this AI-driven approach to literacy as a model that can be replicated across Georgia and potentially nationwide. Our aim is to kindle a passion for reading among young learners, enhance their comprehension skills, and ultimately, improve literacy rates across various age groups and educational backgrounds. For more information about literacy initiatives such as this one, please explore the work of the Barbara Bush Foundation, whose dedication to literacy and education culminated in a literacy X-Prize a few years ago.

The Saga team is eager to explore potential synergies with the literacy coaches, librarians, and K-12 educators. If you wish to explore a partnership that advances both education and technology for societal betterment, please contact us.

In 2023-24 Saga invited its first applications for no-cost high quality transcription in support of qualitative research. In each case, Saga consults with PIs about their particular use cases and provides flexible tooling and AI-drive transcription generations in a timely manner. These projects leverage Saga's special capabilities in the transcription of audio recordings with multiple speakers, and the parsing of English spoken by non-native speakers or native speakers of global varieties of English. Saga is proud to partner with Drs. Shuva and Choi of the CUNY system in this pilot endeavor.

Nafiz Zaman Shuva, Ph.D.
Assistant Professor
Graduate School of Library and Information Studies
Queens College
City University of New York (CUNY)

Yeju "Chloe" Choi, Ph.D.
Assistant Professor of Sociology
John Jay College of Criminal Justice
City University of New York (CUNY)