Adversarial Examples and Human-ML Alignment

The course focuses on understanding adversarial examples and aligning human perception with machine learning models. The learning outcomes include gaining insights into deep networks, interpreting adversarial perturbations, and exploring the consequences on interpretability, training modifications, and robustness tradeoffs. The course teaches skills in analyzing correlations in data, conducting "counterfactual" analysis with robust models, and identifying non-robust features leading to adversarial examples. The teaching method involves theoretical discussions and practical experiments. This course is intended for individuals interested in machine learning, deep learning, and the alignment of human perception with ML models.