The Broken Patent System, and How Machine Learning Can Help


Seth Fletcher of Scientific American recently wrote, "Last month molecular geneticist and crusader for intellectual property reform Richard Jefferson wrote a Forum column for Scientific American in which he explained the concept of 'innovation cartography'—the idea that mapping the vast and inscrutable world of patents will enable more players to create more innovations, some of which might end up doing a lot of good. Since then, Jefferson and his colleagues have launched a new version of The Lens, an open resource for searching and analyzing patents issued worldwide; published a paper in Nature Biotechnology on using The Lens to explore global gene patents; and announced a new round of funding from the Bill & Melinda Gates Foundation. I suspected Jefferson had more to say than he could fit into that Forum piece, so I gave him a call."


Jefferson told Fletcher, "The shameful reality of the U.S. patent system is it is not obliged to register change in ownership. That is a shocking lack of transparency. With The Lens, one of the things you can do is try to extract truth of ownership by looking at patent families. Other countries do register change of ownership. So if I see that a certain family of patents in the U.S. is registered to the University of Michigan, but in Finland and Germany they’re registered to [the controversial patent aggregator] Intellectual Ventures, then it’s possible that Intellectual Ventures also has that patent in the U.S. but has never registered the change of ownership. Part of the goal of the Lens is to shine light on patterns of ownership and use. With machine learning and data you can find shell companies. 'Wait a minute, these companies all have the same PO box in Nevada, and the owner is the wife of a lawyer for Intellectual Ventures'.”


Read more here.


Image: Courtesy Flickr/ brunosan