How are Verification and Validation interpreted and what´s the key difference
Verification: The evaluation of whether or not a product, service, or system complies with a regulation, requirement, specification, or imposed condition. It is often an internal process. Contrast with validation.
Validation: The assurance that a product, service, or system meets the needs of the customer and other identified stakeholders. It often involves acceptance and suitability with external customers. Contrast with verification.
The key difference is that Verification takes a look back and checks if the initial requirements have been fullfilled and in contrast Validation checks if development meets the need of the future customers.
How can be delt with ensuring the safety of Attribute Changes by Machine Learning (ML) during the Operation Phase?
Attribute Changes by ML during the operation phase can be checked with the “Virtual Assessment of Automation in Field Operation (VAAFO) Method” —> This testing aims at testing the robustness of the system in a simulated environment with no actuator controll (System is in the Matrix)
How can a System with ML during the operation phase can be protected against unexpected behaviour?
A safe fallback state should be implemented. This state provides:
Basic & known functionallity
Acts like a sleeping reserve
(Restricted leearning funcitonality in predefined corridors
How can parts of a development process be shiftet into the operating phase? (mostly concerned with software)
Open Alpha & Beta Versions for early adopters on a voluntery basis.
This provides fast Feedback and Bug information and spares testing/ user acceptance issues in the finalized product.
How can designs be compared against each other, taking the user groups into account.
In which levels of involment can user data be collected during the operation phase?
Which levels of remaining personal information are possible for collected user data?
Identified data: Data that can unambiguously be associated with a specific person
Pseudonymized data: Data for which all identifiers are substituted by aliases for which the alias assignment is such that it cannot be reversed by reasonable efforts of anyone other than the party that performed them.
Unlinked pseudonymized data: Data for which all identifiers are erased or substituted by aliases for which the assignment function is erased or irreversible, such that the linkage cannot be re-established by reasonable efforts of anyone including the party that performed them.
Anonymized data: Data that is unlinked and which attributes are altered in such a way that there is a reasonable level of confidence that a person cannot be identified, directly or indirectly, by the data alone or in combination with other data.
Aggreagted data: Statistical data that does not contain individual-level entries and is combined from information about enough different persons that individual-level attributes are not identifiable.
Zuletzt geändertvor einem Jahr
