In recent years, we have witnessed remarkable breakthroughs in artificial intelligence—fromChatGPT reshaping how humans interact with language to Google Veo 3.1 expanding thefrontier of video generation. As the field evolves, attention is now shifting toward Agentic AI,where systems are no longer merely reactive, but capable of goal-driven, autonomousdecision-making. Yet the next major frontier extends beyond the screen. We are entering the eraof Physical AI, where intelligence moves from the digital world into the physical one. PhysicalAI refers to AI systems that can perceive, reason, and act within real-world
environments—sensing through cameras, microphones, and other sensors, and interacting withthe world via motors, robotic arms, wheels, or other actuators.

In this course, Physical AI serves as the central lens through which all projects and discussionsare contextualized. Students will explore how modern AI models are integrated into real-worldsystems such as robots, smart homes, and autonomous vehicles, and how intelligence,embodiment, and environment jointly shape system behavior.As AI systems increasingly operate around us in everyday life, critical questions emerge: How do these systems function in real-world settings? How do we ensure their reliability, safety, and alignment with human values? And what new challenges arise when AI is deployed in open,
dynamic environments?
If these questions inspire you, this course is designed for you.
Intended Learning Outcomes:
1. Explain the definition, scope, and significance of Physical AI, and distinguish it from purely digital and agentic AI systems.
2. Describe the core concepts, system architectures, and key enabling technologies underlying Physical AI, including sensing, perception, reasoning, planning, and control.
3. Analyze and apply foundational Physical AI techniques to real-world problem settings, taking into account environmental dynamics, embodiment, and system constraints.
4. Identify and evaluate major application domains and representative use cases of Physical AI, such as robotics, smart environments, and autonomous systems.
5. Design, implement, and prototype a Physical AI solution that demonstrates the integrated use of perception, decision-making, and physical action.
Note: The lectures during the week of the 12th–14th will be co-taught with Dr. Chun-yien Chang, a research scientist in the Department of Computer Science at National Yang Ming Chiao Tung University. Dr. Chang will cover ontology engineering and ontology-based evaluation.