TU Wien:Security, Privacy and Explainability in Machine Learning/2022-06-30 Exam 2022S
Zur Navigation springen
Zur Suche springen
Exam on 2022-06-30[Bearbeiten | Quelltext bearbeiten]
6 Open questions offline:
- LORE: What are the two components of the explanation? Which algorithm is used?
- Zero Shot Learning: What are the mappings? How can it be used for explainability?
- CIA triangle: Which attacks can be performed on each angle. Are they performed during training or testing time?
- Name approaches to federate two different ML models. Which are the challenges and differences?
- What is Fingerprinting and what is Watermarking? In which contexts were they mentioned in the lecture? What are the differences? When is it used?
- Name two approaches for generating adversarial examples. What are defenses for them and what are shortcomings?