Computational Intelligence and Games 

About: Computational Intelligence and Games is an academic conference. The conference publishes majorly in the area(s): Game mechanics & Game design. Over the lifetime, 993 publications have been published by the conference receiving 19594 citations.

ViZDoom: A Doom-based AI research platform for visual reinforcement learning

Abstract: The recent advances in deep neural networks have led to effective vision-based reinforcement learning methods that have been employed to obtain human-level controllers in Atari 2600 games from pixel data. Atari 2600 games, however, do not resemble real-world tasks since they involve non-realistic 2D environments and the third-person perspective. Here, we propose a novel test-bed platform for reinforcement learning research from raw visual information which employs the first-person perspective in a semi-realistic 3D world. The software, called ViZDoom, is based on the classical first-person shooter video game, Doom. It allows developing bots that play the game using the screen buffer. ViZDoom is lightweight, fast, and highly customizable via a convenient mechanism of user scenarios. In the experimental part, we test the environment by trying to learn bots for two scenarios: a basic move-and-shoot task and a more complex maze-navigation problem. Using convolutional deep neural networks with Q-learning and experience replay, for both scenarios, we were able to train competent bots, which exhibit human-like behaviors. The results confirm the utility of ViZDoom as an AI research platform and imply that visual reinforcement learning in 3D realistic first-person perspective environments is feasible.

Player modeling using self-organization in Tomb Raider: Underworld

Abstract: We present a study focused on constructing models of players for the major commercial title Tomb Raider: Underworld (TRU). Emergent self-organizing maps are trained on high-level playing behavior data obtained from 1365 players that completed the TRU game. The unsupervised learning approach utilized reveals four types of players which are analyzed within the context of the game. The proposed approach automates, in part, the traditional user and play testing procedures followed in the game industry since it can inform game developers, in detail, if the players play the game as intended by the game design. Subsequently, player models can assist the tailoring of game mechanics in real-time for the needs of the player type identified.

Towards automatic personalised content creation for racing games

Abstract: Evolutionary algorithms are commonly used to create high-performing strategies or agents for computer games. In this paper, we instead choose to evolve the racing tracks in a car racing game. An evolvable track representation is devised, and a multiobjective evolutionary algorithm maximises the entertainment value of the track relative to a particular human player. This requires a way to create accurate models of players' driving styles, as well as a tentative definition of when a racing track is fun, both of which are provided. We believe this approach opens up interesting new research questions and is potentially applicable to commercial racing games.

An experiment in automatic game design

Abstract: This paper presents a first attempt at evolving the rules for a game. In contrast to almost every other paper that applies computational intelligence techniques to games, we are not generating behaviours, strategies or environments for any particular game; we are starting without a game and generating the game itself. We explain the rationale for doing this and survey the theories of entertainment and curiosity that underly our fitness function, and present the details of a simple proof-of-concept experiment.

A data mining approach to strategy prediction

Abstract: We present a data mining approach to opponent modeling in strategy games. Expert gameplay is learned by applying machine learning techniques to large collections of game logs. This approach enables domain independent algorithms to acquire domain knowledge and perform opponent modeling. Machine learning algorithms are applied to the task of detecting an opponent's strategy before it is executed and predicting when an opponent will perform strategic actions. Our approach involves encoding game logs as a feature vector representation, where each feature describes when a unit or building type is first produced. We compare our representation to a state lattice representation in perfect and imperfect information environments and the results show that our representation has higher predictive capabilities and is more tolerant of noise. We also discuss how to incorporate our data mining approach into a full game playing agent.