Building Resilient Educational AI Through Multimodal Context Fusion Presented at AAAI 2026 Summer Symposium

Einbrain Lab is pleased to share that Donggil Song successfully presented our work at the AAAI 2026 Summer Symposium in Seoul, South Korea.

Presentation Details
Title: Building Resilient Educational AI Through Multimodal Context Fusion
Authors: Donggil Song & Anne Lippert
Event: AAAI 2026 Summer Symposium
Location: Seoul, South Korea
Date: June 2026

This presentation focused on how educational AI systems can become more resilient when they are grounded in multimodal context. Rather than relying only on text-based prompts or isolated learner inputs, the work explores how AI can use task context, learner actions, environmental state, and interaction history to provide more stable and meaningful support.

The project was demonstrated through a virtual reality algebra learning system, where learners interact with mathematical objects and receive AI-guided support. By connecting the AI teammate to observable actions and system state, the approach is designed to reduce ungrounded responses, support learner exploration, and maintain useful guidance even when interaction conditions change.

At the symposium, the discussion emphasized a broader design principle for educational AI: resilience is not only about model capability, but also about how systems are architected to remain useful, transparent, and adaptable in real learning environments. Multimodal context fusion offers one practical path toward AI systems that can better handle noisy inputs, shifting learner strategies, latency, and resource constraints.

We are grateful for the opportunity to share this work with the AAAI community and to exchange ideas with researchers working across artificial intelligence, learning technologies, human-AI collaboration, and resilience. We also appreciate the support of the National Science Foundation, which has helped make this line of research possible.

This work continues Einbrain Lab's mission to engineer XR solutions with AI for advanced technical training and innovation, with a focus on systems that are not only intelligent but also dependable, human-centered, and useful in practice.