Editor’s Note: This post is written by Zackarey Thoutt. He’s a Udacity Nanodegree program graduate. You may recognize his name, as he’s been all over the media lately. Vice’s Motherboard magazine put it like this: Neural Network Wrote the Next ‘Game of Thrones’ Book Because George R.R. Martin Hasn’t. Zack created that neural network, and in this post, he shares how it all came to be.
~
As I was getting ready to graduate from Udacity’s Deep Learning Foundations Nanodegree program, I began to wonder what to do next. I needed a new project to keep my skills sharp. One episode into Season 7 of Game of Thrones (I’m a huge fan!), the idea hit me—why not train a network to write new chapters for the book we’re all waiting for?
Writing the code for the model and training it only took a few days of work, and after tuning the model’s hyperparameters, I started to get some interesting results. I shared my project with a few friends, and on Udacity’s Slack channel, but that was about it.
Everything changed when I decided to post on HackerNews. It was just a whim, but practically before I realized what was happening, my project was #2 on the HN feed!
After that, the publicity pretty much took off without me having to do much. Someone shared it on the ASOIAF subreddit, a journalist interviewed me and wrote an article for Vice, and since that article was published on Monday, I have been contacted by people from famous podcasts, NVIDIA, the deep learning community, and now, Udacity!
This unexpected (and incredible!) experience was made possible by my Udacity Deep Learning Foundations Nanodegree program, and I think the results of my GOT LSTM network are a testament to both the validity of Udacity as an educational platform, and the growing potential of AI in many aspects of our society.
People have dreamed and speculated about the potential of AI for decades, but I think that technology is finally on the cusp of breaking through the barrier between interesting toy projects and legitimate software that can dramatically increase the efficiency of humankind. My GOT AI is a small taste of what is possible with deep learning. Every day researchers around the world are coming up with new network architectures to better address problems that we face in society. The possibilities of AI going forward are vast and I’m excited to be on the cutting edge of the technology behind the AI movement.
