@RetractionWatch Thorp both reveres peer-reviewed literature and ignores it to suit publisher interests. There are 20 years of papers to contradict his view of the replication crisis. It isn't a small number of actors at fault, and nobody thinks fraud is the main driver. https://t.co/NQNpKgwVjH
I've just discovered that @alexbyrnes also automated this extraction problem in 2014, using CV techniques and hand-built schemas describing different doc types. Insane work. And additional validated training data! https://t.co/GOv0ITa667