Our society and environment increasingly incorporate interaction with digital media and virtual realities, whether it is via a video game, internet browsing, or a mobile phone. User interaction becomes a key component, and systems need to dynamically adapt to the user, instead of simply streaming "canned" static content. This is where our labs come into play:
Our research area is interactive intelligence, i.e., how various components of a virtual world can cleverly adjust to new situations in real time.
This starts at the level of virtual characters, which need to adapt their course of action dependent on the user. For example, a monster in a video game might need to reason on which actions to take to get through a door that was blocked by the player. There is much more than this however, which is why we call such characters virtual actors. They do not only need to act in a goal-driven way that makes sense to the user, but also display convincing emotions and personality, fill a specific character role and actively drive the happenings ahead. The need for virtual actors applies of course to a much wider area than computer gaming, e.g., imagine a virtual assistant, teacher, tour guide, or even friend.
Interactive intelligence does not stop at the character level however. Much more further-reaching tasks require an intelligent adaptation to user actions. For example, as a scene in an interactive setting cannot be fully scripted in advance, there needs to be an automated intelligent mechanism for scene presentation (an intelligent camera) so that the user can experience all important elements. On a higher level, even a whole story line can be re-planned online so that a highly captivating experience is still guaranteed for the user, no matter how the user decides to influence the direction of the story. In short, this leads to procedural content generation and automated optimization of user fun and experience.
For more information on some of our application areas in entertainment, please have a look at this brief overview paper (3 pages) on the future of game AI.
Our core expertise area is the mechanisms for implementing interactive intelligence. However, as it is apparent from above, many other disciplines are involved in realizing our goals. This includes arts, psychology and screenwriting. We are thus pursuing a highly interdisciplinary approach, working in close cooperation with teams from other disciplines.
Our labs will have teams in many different projects and scenarios, ranging from interactive entertainment to new ways of teaching and virtual tour guides. The different teams will, however, always contribute to our core research and software for interactive intelligence, which serves in turn as base platform for all teams.
Instead of ivory-tower research, our direction is to work on real-world scenarios. We thus actively cooperate with industry, and use realistic test beds that point us to effects and properties that were beyond our initial system expectations.
At the center of our research is so-called goal-oriented action planning technology. This technology serves to automatically select and execute basic actions to pursue given goals. We are stepwise improving on various features of this planning technology and adapting it to the needs of interactive media.
Other keywords: intelligent agents, planning & scheduling, plan recognition, machine learning & statistics, probability theory, case-based reasoning, knowledge representation, decision theory, search control, constraint programming, local search
The labs are directed by Alexander Nareyek, who holds an assistant professorship at the National University of Singapore, and also heads the university's multi-disciplinary Games Lab. He received his diploma and Ph.D. from the TU Berlin/Germany, held positions at GMD-FIRST/Germany, Carnegie Mellon University/USA, and the Cork Constraint Computation Centre/Ireland, and served as CEO and CTO for Digital Drama Studios/Czech Republic. He is one of the leading figures in the field of artificial intelligence for gaming and serves on numerous academic and industrial committees. For the International Game Developers Association (IGDA), he is responsible for matters regarding artificial intelligence, and serves as chairperson of the IGDA's Artificial Intelligence Interface Standards Committee. (See his personal homepage for more information.)