The tool currently assumes deterministic behavior of the FUT. However, there are many cases where systems act probabilistically, which could cause issues (in particularly, boundary lost errors).
Some things to consider
Multiple executions per sample
Confidence measures instead of true/false classifications
The tool currently assumes deterministic behavior of the FUT. However, there are many cases where systems act probabilistically, which could cause issues (in particularly, boundary lost errors).
Some things to consider