Research

NuRI Lab studies how robots learn reusable skills, construct predictive world models, navigate dynamic environments, and manipulate objects purposefully alongside people.

Research Topics

Neural structures connecting to the learning system of an intelligent robot

Robot Learning

We study how robots acquire reusable skills from demonstrations, interaction, and feedback. Our work combines imitation learning, reinforcement learning, language-conditioned policies, human-inspired learning, and safe adaptation in real-world environments.

A robot using a spatial world model to evaluate possible actions in an everyday environment

World Models

We develop predictive representations that allow robots to understand how environments may change and anticipate the consequences of their actions. Our research connects spatial reasoning, future prediction, 3D scene representation, causal reasoning, and planning.

A mobile robot navigating through a three-dimensional reconstruction of an indoor environment

Robot Navigation

We build language-guided and socially aware navigation systems for dynamic environments. Our work combines visual navigation, semantic mapping, spatial memory, 3D Gaussian Splatting, and multimodal scene understanding.

A robot learning a multi-step manipulation task with human guidance

Robot Manipulation

We study perception-driven grasping, object-centric 3D reconstruction, and long-horizon task execution. Our research integrates task and motion planning, skill chaining, reusable policies, and interaction for robust real-world manipulation.

Projects

AI digital performer generation for virtual production
Active · 2026–2031

[AI Star Fellowship] Research on AI Digital Performer Generation and Control Intelligence for Virtual Production

The project creates intelligent digital performers that understand immersive environments, respond to user needs, and become more helpful, natural, and reliable through continuous interaction and feedback.

Funded by the Ministry of Science and ICT (MSIT) through IITP.

Past Projects

Physical AI based manufacturing technology validation

[Physical AI PoC] Physical AI-based PoC Platform for Advanced-Manufacturing Technology Validation

A scalable verification platform integrating perception, reasoning, and action for data-driven, automated, and human-aligned manufacturing technology validation.

Funded by the Ministry of Science and ICT (MSIT) through NIPA · 2025.08–2025.12.

Brain-inspired AI and human-like intelligence

[Brain AI] Brain-Inspired AI with Human-Like Intelligence

Research on developmental cognition, computational neuroscience, and brain-based artificial intelligence toward machines with incrementally growing cognitive abilities.

Funded by the Ministry of Science and ICT (MSIT) through IITP · Research participation: 2019.04–2022.12.

Safe and socially acceptable robot learning

[SW Star Lab] Robot Learning: Efficient, Safe, and Socially-Acceptable Machine Learning

Research on data-efficient, safe, and socially acceptable robot learning for autonomous systems operating and interacting alongside people in dynamic environments.

Funded by the Ministry of Science and ICT (MSIT) through IITP · Research participation: 2019.04–2022.12.