Robustness and Generalization via Generative Adversarial Training

In the Trustworthy AI course, I was invited as a guest instructor to present the “Robustness and Generalization via Generative Adversarial Training” paper after completing the initial sections on generalization and robustness. The objective was to enhance students’ understanding of these concepts by discussing a related research paper.

• Presentation Focus: Key points and concepts from the “GAT” paper, linking them to the course material on generalization and robustness.
• Content: Detailed explanation of the paper’s methodology, experiments, and results. Emphasis on the most important definitions and concepts, including the basic related concepts.
• Interactive Elements: Engaged students through discussions and Q&A sessions, ensuring a thorough understanding of the paper’s contributions.
• Conclusion: Summarized the experiments and results, discussed potential future works, and provided an insightful perspective based on class discussions.

Links:

• paper
• Presentation Slides
• recorded session
• you can find the whole content include the paper, slides and recorded session here.