Anna Kipnis is a Creative Technlogist at Google Research. She is best known for her previous work at Double Fine Productions, where she contributed as a Gameplay and AI engineer to various unreleased and top secret projects.
Machine Learning presents incredible, completely unexplored possibilities for games, especially for in-game experiences. This talk hopes to inspire you to experiment with ML by showing you concrete examples and provide you with some tools to start you on your way. Anna will present two distinct examples of using ML for games: The first is crafting AI behaviors using mostly English, with the help of Semantic ML language models. One of the most difficult problems in crafting a believable character AI is accounting for every conceivable interaction the player may have with them, and the tedious process of creating the glue between the player action and the content you’ve created for your game. What if you could use Semantic ML to skip that glue step entirely? The second example will tell the story of how Anna’s team was able to get Style Transfer working in real-time, using a convolutional neural network. The look development process takes enormous time and resources. What if game developers had a tool that cut out tedious steps in that process? Lots of visual examples from the prototyping phase will be provided, including what never made it into the GDC floor demo, what worked and what didn’t.