Machine learning in games has always been a controversial subject. While academics and enthusiasts simply love the idea of applying ML to games, currently it's rarely used in industry. The reasons for this is partly a question of technology, and partly methodology — both of which can require different skills compared no classical game AI. However, the temptations of unlocking the door to designers creating behaviors are hard to deny!
For these reasons, hearing experiences and case studies from developers who've built such technology is particularly interesting. I recently republished an interview and analysis of Artificial Contender. The Technical Director of this product, Iskander Umarov, also recently gave a presentation on the subject where he answered all questions about the technology.
This is a presentation about behavior capture in general and how TruSoft approaches the problem. In this session, you'll be introduced to the concept of learning by example, then discover the different technologies that can be applied to this domain. In the second half, you'll see how this kind of technology can be applied to commercial sports game, showing learning by imitation used in practice.