We note two potential potential ML attack issues (issue 439 and issue 440). However they are rather heavy VERIS changes. Barring data we probably don't want to make them. A lighter change may be to add a 'plus.ML' to check if an attack was against a machine learning system of some type. (We'd likely want to put it in VCDB rather than core VERIS.) This would give us a place to collect data about how often ML models are being attacked which could then tell us when to pursue issues 439 and 440.
We note two potential potential ML attack issues (issue 439 and issue 440). However they are rather heavy VERIS changes. Barring data we probably don't want to make them. A lighter change may be to add a 'plus.ML' to check if an attack was against a machine learning system of some type. (We'd likely want to put it in VCDB rather than core VERIS.) This would give us a place to collect data about how often ML models are being attacked which could then tell us when to pursue issues 439 and 440.