Wondering if it be viable to train a classifier on a set of patents that are known to be vague, abused, and pernicious and see what else it digs up.
This works for malware, and simple things like being owned by known patent trolls, who have selected them for their crappiness are easy indicators of what a bad one looks like, a history of litigation, clusters of ones with similarities, etc.
Surely this is somebody's ML driven patent troll startup right now?
This works for malware, and simple things like being owned by known patent trolls, who have selected them for their crappiness are easy indicators of what a bad one looks like, a history of litigation, clusters of ones with similarities, etc.
Surely this is somebody's ML driven patent troll startup right now?