Center for Artificial Intelligence Research

Research Clusters

CAIR Research Clusters

The Center for AI Research brings together interdisciplinary research clusters focused on building intelligent systems that sense, understand, communicate, reason, and act in real-world environments. Each cluster advances a distinct area of AI research while contributing to CAIR’s broader mission of developing human-centered, trustworthy, and physically grounded artificial intelligence.

Cluster 1: Interactive Physical AI for Rehabilitation, Education, and Alzheimer’s Monitoring

This cluster develops interactive physical AI systems that can understand human cognitive, emotional, and behavioral states and respond through adaptive, multimodal interaction. The goal is to create AI-enabled systems that support rehabilitation, education, healthcare assistance, and cognitive monitoring through natural and human-centered interfaces.

The cluster focuses on combining sensing, human-state modeling, speech, facial expressions, avatars, and embodied interaction to create intelligent systems that can adapt to people in real time. These systems are designed to support applications such as rehabilitation assistance, educational support, Alzheimer’s monitoring, and healthcare engagement.

Current project areas include proposal development for healthcare and education initiatives, interactive AI systems for human-centered support, and foundational implementation work for future AI platforms in partnership-driven research opportunities.

Key Focus Areas

Interactive physical AI systems
Human-state understanding
Rehabilitation and healthcare support
Alzheimer’s and cognitive monitoring
Educational AI systems
Speech, avatar, and facial-expression interaction
Adaptive multimodal response generation

Cluster 2: Multimodal Sensor Fusion, Edge AI, and Robust Physical Perception

This cluster advances AI systems that can perceive and understand complex physical environments using multiple sensing modalities. Its work focuses on sensor fusion, edge intelligence, real-time perception, and robust decision-making under noisy, incomplete, or uncertain conditions.

The cluster develops frameworks that combine sound, spatial sensing, object detection, semantic segmentation, distributed sensors, and edge-AI pipelines. These systems are designed to operate in real-world environments where data may be imperfect, sensors may fail, and fast decisions are required.

Research in this cluster supports applications in physical-context understanding, activity recognition, tunnel detection, distributed monitoring, sound-spatial perception, and resilient sensing for defense, infrastructure, and environmental applications.

Key Focus Areas

Multimodal sensor fusion
Edge AI and real-time inference
Sound-spatial perception
Object detection and semantic segmentation
Sensor synchronization and alignment
Robustness under noisy or missing data
Uncertainty estimation and sensor-failure detection
Distributed sensing and physical-environment monitoring

Cluster 3: Responsible, Ethical, and Compliance-Aware Agentic AI

This cluster develops responsible AI systems that operate safely, transparently, and ethically in real-world environments. Its research focuses on building AI that respects privacy, supports compliance, incorporates human values, and provides accountable decision-making in regulated and sensitive domains.

The cluster explores ethical reasoning, compliance-aware architectures, privacy-preserving AI, human-in-the-loop oversight, auditability, and responsible agentic AI. It also investigates multimodal human-state understanding using signals such as EEG, eye tracking, physiological sensing, behavioral cues, and contextual data.

A major theme of the cluster is EdgeEthics: the development of ethical and context-aware AI systems that can reason responsibly at the edge while maintaining privacy, transparency, and human oversight.

Key Focus Areas

Responsible and ethical AI
Compliance-aware AI systems
Privacy-preserving edge intelligence
Agentic AI governance
Human-in-the-loop oversight
Auditability, traceability, and accountability
EEG, eye tracking, and physiological sensing
Ethical AI for healthcare, education, and regulated environments

Cluster 4: Multimodal Evaluation, Communication, and Human Feedback Systems

This cluster develops AI systems that evaluate human communication and provide structured, explainable feedback for learning, training, and self-improvement. Its work focuses on speech, language, delivery, prosody, engagement, confidence, and audience impact.

The cluster combines multimodal analysis, rubric-based evaluation, large language models, retrieval-based feedback systems, and performance dashboards to help users improve communication over time. These tools can support education, public speaking, professional training, healthcare communication, and faith-based communication contexts.

The goal is to create transparent and human-centered feedback systems that do more than score performance. These systems explain what worked, identify areas for improvement, and provide actionable guidance for growth.

Key Focus Areas

Multimodal communication evaluation
Speech and language analysis
Public-speaking and presentation feedback
Rubric-based assessment
AI-assisted feedback generation
Performance tracking dashboards
Explainable evaluation systems
Human-centered learning and self-improvement

Cluster-Level Impact

Together, these clusters position CAIR as a hub for AI research that connects physical intelligence, human-centered interaction, responsible AI, multimodal perception, and communication feedback. The clusters are designed to produce prototype systems, technical reports, proposal-ready materials, research publications, reusable software infrastructure, and future funded research programs.

Through these efforts, CAIR advances AI systems that are not only intelligent, but also interactive, trustworthy, explainable, adaptive, and ready for real-world deployment.

Page updated

Report abuse