Game of Thrones is undeniably the most watched TV show, at least for this year. The premiere of this show became the most watched in the history of TV. However, the recent hacking of HBO servers ended up in the leaking of scripts and contact details of all the cast of the show. The hackers demanded millions to stop leaking the popular show’s episodes. This is just a little representation of the fan base that drools over the tiniest bit of detail that they can get out of it.
George R. R. Martin is currently dusting finishing touches to the final piece in the Game of Thrones saga, the TV series that won’t come out until 2019. Fans are beginning to get impatient, so one of them came up with the idea of getting a neural network to predict the finale.
Zack Thoutt, a software engineer, is a crazy fan who began to train a recurrent neural network (RNN) to predict how the final season of GoT ‘The Winds of Winter’ will end up. Thoutt just completed an artificial intelligence and deep learning course on Udacity and decided to put it up to some use. He commented on his project,
“I’m a huge fan of Game of Thrones, the books, and the show. I had worked with RNNs a bit in that class and thought I’d give working with the books a shot.”
Artificial Neural Networks replicate the structure of biological neural networks or simply animal brains. The structure continues to modify itself using the obtained information.
Thoutt explains his project as:
“A long short-term memory RNN which is remembering plot points from thousands of words before. It is trying to write a new book. A perfect model would take everything that has happened in the books into account and not write about characters being alive when they died two books ago. The reality, though, is that the model isn’t good enough to do that. If the model were that good, authors might be in trouble. The model is striving to be a new book and to take everything into account, but it makes a lot of mistakes because the technology to train a perfect text generator that can remember complex plots over millions of words doesn’t exist yet.”
The RNN created by Zack has already written 5,376 pages of the first five books of the series, and they were published on GitHub. You can expect this computer generated text to be grammatically incorrect. However, it has managed to make some accurate predictions as well, some of them even matching the theories popular on many GoT forums.
The plot predicted by the neural network may not be the exact copy of the actual one, but this is your chance to dig into a computer written theory of the upcoming series. Computers of the future may begin to write fan-fiction, and this is definitely your first and may be the best look into it. Martin, however, is pretty good with twisting everything away from our expectations, so we just have to wait and see. Meanwhile, you have some fan-fiction to quench your GoT thirst.