Hypothetically Speaking: Text Generated Redacted Materials
muellerLONG
I have recently learned more about generative text algorithms and am interested in taking different redacted texts relating to the Mueller Report/current events and training a very simple text generator on that data set. Theoretically by doing that one could assume that **eventually** some of the redacted portions of the text would be generated by the algorithm. Maybe. Could we expose some state secrets on accident? Probably not, but interesting to try!?
see the current iteration of the project at: speakinghypothetically.com