NIST 2026 SUMMIT

Prof. Minhee Jun delivered a selected oral presentation at the 2026 NIST University Research Summit, held at the National Institute of Standards and Technology in Gaithersburg, Maryland, from June 30 to July 1, 2026.

Her presentation, titled “Toward Trustworthy Physical AI: Human Demonstrations, Task-Centric Digital Twins, and TEVV for Robot Learning,” introduced a research vision for developing trustworthy Physical AI systems through privacy-aware human demonstrations, task-centric digital twins, and rigorous testing and evaluation.

The proposed framework explores the use of smart-glasses recordings and existing video repositories to capture human task demonstrations and transform them into simulation-ready digital environments. Privacy-preserving processing and information-masking techniques are incorporated to protect sensitive information while maintaining the usefulness of the collected data for AI training and evaluation.

The research also emphasizes Test, Evaluation, Verification, and Validation (TEVV) through reproducible datasets, benchmark development, and systematic performance-assessment methods. These capabilities could support safer, more reliable, and more scalable robot-learning systems while reducing dependence on costly real-world experimentation.

Prof. Jun’s presentation also highlighted potential opportunities for collaboration with NIST in digital twin technologies, evaluation science, benchmarking, and standards development for trustworthy Physical AI.

Her participation in the Summit reflects CAIR’s continued commitment to advancing responsible artificial intelligence, robotics, digital twins, and scientifically grounded approaches to AI evaluation.