lectures on game AI

The problem of creating believable game AI poses numerous challenges for computational intelligence research. A particular challenge consists in creating human-like behaving game bots by means of applying machine learning to game-play data recorded by human players. In this paper, we propose a novel, biologically inspired approach to behavior learning for video games. Our model is based on the idea of movement primitives and we use Archetypal Analysis to determine elementary movements from data in order to represent any player action in terms of convex combinations of archetypal motions. Given these representations, we use supervised learning in order to create a system that is able to synthesize appropriate motion behavior during a game. We apply our model to teach a first person shooter game bot how to navigate in a game environment. Our results indicate that the model is able to simulate human-like behavior at lower computational costs than previous approaches.

Most computer games that are set in simulated 3D worlds require the player to act according to his or her current location. For the programming of convincing artificial game agents this poses the problem of providing them with a sense of where they are. However, due to the growing size and complexity of games, manual annotations that capture all essential areas and aspects may soon become infeasible. In this paper, we therefore propose a data-driven approach to the semantic clustering of in-game locations. Following the paradigm of gamebot programming by human demonstration, we apply the technique of spectral clus- tering to automatically derived waypoint maps. First results underline that this approach indeed provides spatial partitions of game maps that make sense from a gamers point of view.

Present day video games offer interesting perspec-tive for computer science research and eduction. As the first generation in history that grew up playing computer games is gradually moving from being students to being teachers, we are now witnessing a growing number of scientific publications and university courses dealing with games. In this contribution, we will report on both these aspects. By means of the problem of behaviour modelling for game characters, we will outline the po-tential of modern commercial games for research in subsymbolic machine learning. Using the network traffic of multiplayer games as input for our algorithms, we will present results obtained with different neural network architectures. Finally, we also shall point out the benefits of video games in teaching machine learning materials and report on first encouraging experiences with computer games in undergraduate education.

Superior weapon handling is the road to success in first person shooter games. On the expert level of gameplay, this demands more than just accurate tracking and targeting of opponents. It also requires knowledge about the behavior and characteristics of different weapons in the player's arsenal. Given enough practice, human players usually master both skills. Usual gamebots, in contrast, can only track and target but lack the sense for the appropriateness of a weapon in a given situation. As a consequence, their performance and behavior appears non-natural. In this paper, we propose to make use of machine learning techniques to realize aiming behavior that is adapted to the spatio-temporal context. We will present a mixture of experts architecture of neural networks which is specialized in the use of several weapons. Experimental result show that this paradigm indeed is able to produce more human-like aiming behavior.

As it strives to imitate observably successful actions, imitation learning allows for a quick acquisition of proven behaviors. Recent work from psychology and robotics suggests that Bayesian probability theory provides a mathematical framework for imitation learning. In this paper, we investigate the use of Bayesian imitation learning in realizing more life-like computer game characters. Following our general strategy of analyzing the network traffic of multi-player online games, we will present experiments in automatic imitation of behaviors contained in human generated data. Our results show that the Bayesian framework indeed leads to game agent behavior that appears very much human-like.

Traditionally, the programming of bot behaviors for com-mercial computer games applies rule-based approaches. But even complex or fuzzyfied automatons cannot really chal-lenge experienced players. This contribution examines whether bot programming can be treated as a pattern recog-nition problem and whether behaviors can be learned from recorded games. First, we sketch a technical computing in-terface to a commercial game that allows rapid prototyping of classifiers for bot programming. Then we discuss the use of self organizing maps to represent manifolds of high dimen-sional game data and how multilayer perceptrons can model local characteristics of such manifolds. Finally, some exper-iments in elementary behavior learning are presented.

Evaluating the spatial behavior of players allows for comparing design intent with emergent behavior. However, spatial analytics for game development is still in its infancy and current analysis mostly relies on aggregate visualizations such as heatmaps. In this paper, we propose the use of advanced spatial clustering techniques to evaluate player behavior. In particular, we consider the use of DEDICOM and DESICOM, two techniques that operate on asymmetric spatial similarity matrices and can simultaneously uncover preferred locations and likely transitions between them. Our results highlight the ability of asymmetric techniques to partition game maps into meaningful areas and to retain information about player movements between these areas.

Analyzing telemetry data of player behavior in computer games is a topic of increasing interest for industry and research, alike. When applied to game telemetry data, pattern recognition and statistical analysis provide valuable business intelligence tools for game development. An important problem in this area is to characterize how player engagement in a game evolves over time. Reliable models are of pivotal interest since they allow for assessing the long-term success of game products and can provide estimates of how long players may be expected to keep actively playing a game. In this paper, we introduce methods from random process theory into game data mining in order to draw inferences about player engagement. Given large samples (over 250,000 players) of behavioral telemetry data from five different action-adventure and shooter games, we extract information as to how long individual players have played these games and apply techniques from lifetime analysis to identify common patterns. In all five cases, we find that the Weibull distribution gives a good account of the statistics of total playing times. This implies that an average player's interest in playing one of the games considered evolves according to a non-homogeneous Poisson process. Therefore, given data on the initial playtime behavior of the players of a game, it becomes possible to predict when they stop playing.