XAI: EXperiential Learning in AI/ML for Future Innovators

Keynote title

From Perception to Navigation: Multimodal Spatial Intelligence for Autonomous Robots

Abstract

The real world presents many highly dynamic and unstructured environments that challenge conventional robotic perception and navigation systems. This talk presents an integrated research direction combining zero-shot vision–language mapping, biologically inspired navigation, and multimodal traversability estimation. The talk will introduce the ZIVIL framework, which enables robots to construct semantic spatial graphs of complex sites using vision–language models without task-specific training. Next, the talk will introduce LiDAR–camera fusion with self-supervised learning to provide dense and adaptable estimates of terrain traversability. Finally, the talk will cover FLYNN, a robust navigation network inspired by the topology of the fly brain, supporting reliable decision-making in complex environments. Together, these results demonstrate how semantic understanding, biologically inspired computation, and sensor fusion can complement one another to improve autonomous robot operation in unstructured environments.

Keynote title

Knowledge Graphs and AI Data Infrastructure for Community Needs

Abstract

Community needs often cross the boundaries of datasets, disciplines, and organizations. Addressing these needs with artificial intelligence requires infrastructure that connects information, preserves its meaning, and makes supporting evidence accountable. This talk explores how Open Knowledge Networks and knowledge graphs provide a foundation for such infrastructure. It introduces RURAL-KG, an NSF-funded project connecting health, justice, geographic, and treatment-service information to support research on rural resilience. The presentation discusses the technical challenges of integrating heterogeneous data, aligning concepts across domains, maintaining data quality and provenance, and enabling natural-language access through AI. Building on these foundations, it presents the expansion toward Alabama Care Commons, connecting knowledge infrastructure with community partnerships and efforts to strengthen rural referral coordination, health IT readiness, cybersecurity, and provider support. Throughout this progression, community needs guide which information should be connected, which tools should be developed, and how their usefulness should be evaluated. The talk highlights opportunities for researchers, healthcare providers, and community organizations to build and sustain knowledge infrastructure that supports practical, locally relevant applications of AI.

Keynote title

Beyond the AI Buzz: Narrative Intelligence for Real-World National Security

Abstract

Most conversations about AI start with the hype. Ours starts with the problem. AI has made content cheap and fast to produce at scale, and adversaries now use it to shape perception, decisions and behavior. At EdgeTheory, we apply AI to that problem through narrative intelligence. We pair human expertise with AI-assisted analysis to detect, track and explain how narratives spread, who they influence, and what effects they are designed to produce, early enough for leaders to act.

In this talk, we'll share what we've learned deploying AI in practice: where it delivers real value, where it falls short, and why human judgment stays at the center of high-stakes decisions. We'll also look at what's next for industry, research and students, including the EdgeTheory Lab at Mississippi State, where MSU students are already working on real narrative intelligence problems. And as MSU builds its new Corps of Cadets, we'll make the case that tomorrow's leaders need to understand the cognitive domain, and that Mississippi can be a national center for narrative intelligence serving defense and industry.

Keynote title

From Answers to Agents: Building the AI Workforce of Tomorrow

Abstract

AI is moving from tools that answer questions to agents that plan, decide, and act on our behalf. That shift will change how science is done, how companies operate, and what skills the next generation of engineers and researchers will need. In this keynote, Kiran Chinnagangannagari, Co-Founder and Chief Technology and AI Officer at Securin, shares what he has learned putting AI agents into real production work. He covers where they deliver real value today, where they still fail, and why trust, judgment, and human oversight matter more than ever. He also offers practical guidance for students and educators on how to learn, build, and lead in a world where everyone works alongside AI.