You have determined that you have numerous Machine Learning detections in your environment that are false positives. They are caused by a single binary that was custom written by a vendor for you and that binary is running on many endpoints. What is the best way to prevent these in the future?