What can Deep Learning do for Serious Games?

What can Deep Learning do for Serious Games?

Roger Smith, President, Model Benders LLCHEALTHCARE

Click for Speaker Bio

AI has been a part of games, serious games, and training simulations for decades. AI techniques like finite state machines, knowledge-based systems, and constraint models are used to create intelligent opponents and support characters in most games. But these AI techniques have not typically been very useful in other phases of the game-based learning process – specifically for constructing complex scenarios data sets or analyzing voluminous post-play data to measure performance quality.
In this session we will explore the applications of “the new AI”, deep learning and machine learning techniques, to the process of automating the creation and modification of large scenario data sets that are created before the game/sim is run. We will also apply these new techniques to the analysis of the data that is generated by players during the game mission. Large, multi-player games can generate gigabytes worth of data that challenges the ability of human analysts to identify important metrics and patterns. Deep Learning algorithms excel at processing massive amounts of data and identifying multi-dimensional patterns hidden within it. The availability of these algorithms opens the door to a deeper understanding of human performance and greater automation across the entire game-based learning process.

Attendee Benefits:

Attendees will be introduced to deep learning and machine learning algorithms and where these can be applied in the game-based learning process – specifically, autoencoders, convolutional neural networks, generative adversarial networks, transformers, self-organizing maps, support vector machines, clustering, and K-nearest neighbor. We will also discuss how to get better results by joining “old AI” (symbolic) and “new AI” (connectionist) methods together into complimentary models.

Thu 12:00 am - 12:00 am