Recent Research from the NSF AI Institute for Engaged Learning (EngageAI)

• In Event: Next-Generation Educational AI Research From Three NSF-Funded National AI Institutes

Abstract

Objectives
The mission of the NSF AI Institute for Engaged Learning (EngageAI) is to produce transformative advances in STEM learning and teaching with AI-driven narrative-centered learning environments. The Institute conducts research on narrative-centered learning technologies, collaborative narrative-centered learning, and multimodal analysis of engaged learning to create deeply compelling story-based learning experiences. Woven throughout the Institute’s activities is a strong focus on ethics, with an emphasis on creating AI-augmented learning that is deeply informed by considerations of fairness, accountability, transparency, trust, and privacy. The Institute’s vision is informed by connections with diverse stakeholders to ensure that the learning environments it creates promote diversity, equity, and inclusion.

Perspectives
Our formulation of narrative-centered learning is guided by problem-based learning, anchored instruction, cognitive apprenticeship, learning-by-teaching, and other approaches to collaborative inquiry that ground learning in meaningful tasks.

Methods
Our research has three complementary thrusts. First, our learning environments dynamically tailor interactive narratives to serve the problem-solving needs and interests of students in particular learning contexts through adaptation of plot, storyworld elements, and character conversations to best support learning. These capabilities draw on emerging advances in large language models and machine learning to create interactive narrative planning and generation technologies (Authors, 2024). They are complemented by research on narrative-centered learning authoring tools (Authors, in press), which leverage generative AI to customize narrative-centered learning environments to meet the specific needs of students, classrooms, curricula, and research programs.

Second, we conduct design-based research to create adaptive collaborative narrative-centered learning environments to analyze and support students’ collaborative inquiry and problem-solving processes during narrative-centered learning (Authors, 2024). Our adaptive collaborative learning framework is driven by advances in natural language understanding, including multi-party dialogue analysis.

Third, we are developing multimodal vision-language modeling systems to support video summarization and query-based retrieval capabilities to enable researchers and teachers to ask questions about video-audio data from student learning with AI-enabled narrative-centered learning environments. Rapid innovation in multimodal AI integrating natural language processing, computer vision, and machine learning enables increasingly powerful capabilities in multimodal analyses of collaborative narrative-centered learning (Authors, 2023).

References
Authors. (2024). Online reinforcement learning-based pedagogical planning for narrative-centered learning environments. Proceedings of the AAAI Conference on Artificial Intelligence, (pp. 23191-23199).
Authors. (2024). Examining coordinated computer-based fixed and adaptive scaffolds in collaborative problem-solving environments. Proceedings of the Seventeenth International Conference on Computer-Supported Collaborative Learning, (pp. 43-50).
Authors. (in press). Procedural level generation in educational games from natural language instruction. IEEE Transactions on Games.
Authors. (2023). Self-chained image-language model for video localization and question answering. Advances in Neural Information Processing Systems, (pp. 76749-76771).

Author

• James C. Lester, North Carolina State University