Week 08 - Reconfiguring Semiconductor R&D Paradigm with AI

Topic Overview:

This topic explores how AI can reconfigure semiconductor R&D by leveraging authentic, continuous, process-context-rich data generated from a pilot line. Focusing on process R&D, design modeling, and factory facilities, it presents a closed-loop AI framework where atomic-scale simulation supports device and system design, design guides fabrication, and pilot-line data continuously calibrates upstream models. This approach highlights the unique value of pilot-line-driven R&D institutions in bridging fundamental modeling, engineering development, and real-time manufacturing intelligence.

Key Topics:

1. Introduce AI-driven process R&D using machine learning inter-atomic potentials for wide-bandgap semiconductor materials such as SiC, GaN, and Ga₂O₃, enabling high-accuracy, low-cost atomic-scale simulation.
2. Explain how AI can be embedded across device, packaging, and circuit design to build a multi-scale collaborative design chain from materials to systems.
3. Discuss how pilot-line data, equipment signals, and characterisation results can power online-learning AI agents for real-time decision support and continuous model optimisation.

Teaching Format

Lecture and case studies