Trustworthy AI

This course introduced the basics of adversarial robustness, reinforcement learning, AI safety, digital twin/metaverse/certification, generalization, and social good (privacy and fairness).

Learning Outcomes:

1. Adversarial machine learning: poisoning, evasion, FGSM, robust physical attack, randomization, robust AI.
2. Model based Reinforcement learning - Neural networks, Gaussian processes.
3. Trustworthy reinforcement learning.
4. Applied AI for real-world problems.

Programming Language: Pytorch, Python