Special Topics in Computer Engineering

The objective of the course is to understand key concepts and techniques required for the explainable AI(XAI) and DL acceleration system. Lecture will cover core methodology for XAI in a variety of detailed areas, such as visual explanation, attention mechanisms, human-AI collaboration, explainable reinforcement learning (RL), explainable graph neural networks (GNNs), etc. We also introduce the acceleration methodology for training/inference of explainable models in the perspective of system frameworks.