1/11
Looks like no tags are added yet.
Name | Mastery | Learn | Test | Matching | Spaced | Call with Kai | Chat |
|---|
No analytics yet
Send a link to your students to track their progress
Rights of the data subject
Right to access their data
Right to be forgotten
Right to be informed
Right to explainability (shaky)
How do we honor rights?
Comply with legal restrictions, anonymize data, follow guidelines based on governing body, and access control.
Anonymize data
Minimization, pseudonymization, suppression, etc.
Achieve anonymity (k-anonymity, l-diversity)
Guidelines based on governing body
HIPAA breach rules, GDPR, US patchwork
Acess control
Principle of least privilege, scoped access, IAS
What if we want to build something?
Decide software licensing
Desing data collection and storage to comply with guidelines (Github minimization, protecting keys)
Humans vs AI in creating things
Access
Production team: who has it and what should be protected or differ by team
Client side: how do we structure this from an organizational standpoint, who are the clients
Final output
How explainable is that model and how explainable should it be?
Tradeoffs between function and explainability (speed vs security vs explanation)
what happens when something goes wrong
Forensic audit, lawsuit, model registration,
Explainability vs. usability
lawsuits, company audits
When is a simple model better
When a more complex model would take too much time, if the complex model doesn’t provide additional accuracy.
Assignment categories