Open porglezomp opened 9 years ago
Interested in seeing how this plays out. It sounds like this might be the new Minstrel.
Reminds me a little bit of what I tried in 2013. Though I got sidetracked by the limitations of the tools I was using and went off in another direction, so I'm looking forward to seeing what you come up with.
I'm joining with https://github.com/porglezomp/nemesis-nanogenmo15.
My idea is to download a bunch of books and run them through semantic processing (with NLTK) to identify the subjects and objects in sentences, and use those to identify interactions between characters. Then I can run some form of machine learning to either generate interactions between characters actually extracted from the stories or to generate new character relationships between random characters. Ideally it will produce a bunch of short stories that are sort of like very low-quality fan fiction.
Just for fun, I've decided that while I'm developing it, I'll only use stories written by women as input. I'm not committing to anything as input to the final product yet; we'll see.