In this paper, we show that personalized levels can be auto- matically generated for platform games. We build on previ- ous work, where models were derived that predicted player experience based on features of level design and on playing styles. These models are constructed using preference learn- ing, based on questionnaires administered to players after playing different levels. The contributions of the current pa- per are (1) more accurate models based on a much larger data set; (2) a mechanism for adapting level design parameters to given players and playing style; (3) evaluation of this adap- tation mechanism using both algorithmic and human players. The results indicate that the adaptation mechanism effectively optimizes level design parameters for particular players.

Role-playing games, and other types of contemporary video games, usually contain a main storyline consisting of several causally related quests. As players have different motivations, tastes and preferences, it can be beneficial to customize game plotlines. In this paper, we present an offline algorithm for adapting human-authored game plotlines for computer role-playing games to suit the unique needs of individual players, thereby customizing gaming experiences and enhancing re-playability. Our approach uses an plan refinement technique based on partial-order planning to (a) optimize the global structure of the plotline according to input from a player model, (b) maintain plotline coherence, and (c) facilitate authorial intent by preserving as much of the original plotline as possible. A theoretical analysis of the authorial leverage and a user study suggest the benefits of this approach.

In 2004, Botea et al. published the HPA* algorithm (Hierarchical Pathfinding A*), which is the first detailed study of hierarchical pathfinding in games. However, HPA* can be optimized. In this paper, we introduce the DHPA* and SHPA* hierarchical pathfinding algorithms, along with a metric for comparing the dynamic performance of pathfinding algorithms in games. We show that DHPA* searches up to an order of magnitude faster than HPA*, but consumes significantly more memory and produces slightly less optimal paths. The SHPA* algorithm searches up to five times faster than HPA* and consumes less memory, but it also produces slightly less optimal paths, and is only fit for static environments. When comparing the algorithms in dynamic environments, our results show that DHPA* often outperforms HPA*. Unlike many other hierarchical pathfinding algorithms, both solutions presented in this paper maintain a read-only terrain representation during search, which simplifies programming and debugging, and improves multithreaded performance.

Pathfinding systems that operate on regular grids are common in the AI literature and often used in real-time video games. Typical speed-up enhancements include reducing the size of the search space using abstraction, and building more informed heuristics. Though effective each of these strategies has shortcomings. For example, pathfinding with abstraction usually involves trading away optimality for speed. Meanwhile, improving on the accuracy of the well known Manhattan heuristic usually requires significant extra memor We present a different kind of speedup technique based on the idea of identifying and eliminating symmetric path segments in 4-connected grid maps (which allow straight but not diagonal movement). Our method identifies rectangular rooms with no obstacles and prunes all interior nodes, leaving only a boundary perimeter. This process eliminates many symmetric path segments and results in grid maps which are often much smaller and consequently much faster to search than the original. We evaluate our technique on a range of different grid maps including a well known set from the popular video game Baldur's Gate II. On our test data, A* can run up to 3.5 times faster on average. We achieve this without using any significant extra memory or sacrificing solution optimality.

A method is presented for converting between Guitar Hero difficulty levels by treating the problem as one of machine translation, with the different difficulties as different languages. The Guitar Hero I and II discs provide aligned corpora with which to train bigram-based language models and translation models. Given an Expert sequence, the model can create sequences of Hard, Medium, or Easy difficulty that retain the feel of the original, while obeying heuristics typical of those difficulties. Training the model requires a single pass through the corpus, while translation is quadratic in the length of the Expert sequence. The method outperforms a recurrent neural network in producing sequences that match the hand-designed levels. The method may make it easier for amateurs to produce content for the Rock Band Network.

A method is presented for training an Input-Output Hidden Markov Model (IOHMM) to identify a player's current goal in an action-adventure game. The goals were Explore, Fight, or Return to Town, which served as the hidden states of the IOHMM. The observation model was trained by directing the player to achieve particular goals and counting actions. When trained on first-time players, training to the specific players did not appear to provide any benefits over a model trained to the experimenter. However, models trained on these players' subsequent trials were significantly better than the models trained to the specific players the first time, and also outperformed the model trained to the experimenter. This suggests that game goal recognition systems are best trained after the players have some time to develop a style of play. Systems for probabilistic reasoning over time could help game designers make games more responsive to players' individual styles and approaches.

In the standard rules of the board game Risk, players take turns selecting or "drafting" the 42 territories on the board until all territories are owned. We present a technique for drafting territories in Risk that combines the Monte Carlo tree search algorithm UCT with an automated evaluation function. Created through supervised machine learning, this function scores outcomes of drafts in order to shorten the length of a UCT simulation. Using this approach, we augment an existing bot for the computer game Lux Delux, a clone of Risk. Our drafting technique is shown to greatly improve performance against the strongest opponents supplied with Lux Delux. The evidence provided indicates that territory drafting is important to overall success in Risk.

Realistic simulations of the movement of infantry in urban environments with 3D models and at interactive rates is of wide and enduring interest. Many video games have attempted it, and military simulations are increasingly doing the same. Previous attempts have fallen short in two areas: properly coordinating movement, and adequate modeling of the detection of hostile targets. Novel algorithms to simulate fireteam movement and visual scanning appropriate to urban environments are described. Measurements of the computational performance of the most expensive components of the approach are provided.

Shepherding with a dog presents an interesting challenge for artificial intelligence, with multiple intelligent systems assessing and interacting with each other in order to achieve a variety of goals. We present a solution to this problem, which consists of a dog AI making use of influence mapping, state machines and A* pathfinding to respond intelligently to real-life shepherding commands issued by a high-level shepherd AI steering the flock of sheep through waypoints on a variety of maps by using pathfinding and influence maps. The role of the AI shepherd can also be taken by a human player (using either a point and click or voice recognition interface) for matches against the artificial shepherd which proved to be a worthy opponent for human testers. The system was evaluated through user testing and provided a high degree of realism and engaging gameplay relying heavily on the workings of the presented AI components.

Chris Jurney (Lead Programmer, Double Fine Productions) Sumit Basu (Microsoft Research) Chris Jurney is a rock and roll experimental game For those who can play an instrument or have a respectable programmer at Double Fine Productions, with 11 singing voice, music can be a wonderful years experience in games and simulation. He has means of creative expression, social engagement, shipped 4 titles in the games industry: Company of and fun. For many others, though, it can be frustrating Heroes, Frontline: Fuel of War, Dawn of War 2, and and inaccessible: even if an inspired youth Brutal Legend. Jurney frequently speaks on the topic has great musical ideas, she may not have the of game AI, having presented at the Game Developers knowledge or ability to get her latest song out from Conference (GDC), GDC China, Columbia her head and into her MP3 player. In this talk, Basu will show three vignettes of how he and his colleagues University, the University of Pennsylvania, and the have used interactive machine learning to New Jersey and Philadelphia chapters of the International extend the creative reach of aspiring musicians: a Game Developers Association (IGDA).