An investigation of challenges encountered when specifying training data and runtime monitors for safety critical ML applications

Open in new window