One of the keynote speakers at the 6th International Conference on "Computer Games; Opportunities and Challenges" is Dr. Julian Togelius, an associate professor in the School of Computer Science and Engineering at New York University (NYU), director of the NYU Game Innovation Lab, and Editor of the IEEE Transactions on Games Journal.

Dr. Togelius is one of the leading researchers in artificial intelligence, evolutionary computing, procedural content production, and computer games fields with an h-index of 62, and will speak at our conference on the machine learning revolution in procedural content generation. The summary of his speech is as follows:

Procedural content generation is a key need in many types of games and virtual worlds. From humble beginnings in the 1980s, PCG methods have advanced rapidly, and as a result, PCG is more widely used than ever. Levels in platform games can be generated, as well as maps in strategy games to quests, characters, textures, items, and planets. Early approaches typically relied on special-purpose algorithms that could not easily be controlled. Academic research on PCG started in earnest a little over a decade ago, coinciding with the rise of indie games and partially catalysed by the success of evolutionary algorithms. Since a few years, another revolution is underway in PCG: the machine learning revolution. I will showcase some recent results in applying machine learning, in particular deep learning, to generating game content. I will also point out some of the limitations of this approach, in particular the training data problem and the constraint modeling problem. I will also point to some ways in which hybridizing with search, optimization, and constraint satisfaction methods can help.