This report presents a tool developed for the analysis and visualisation of Rolling Horizon Evolutionary Algorithms, featuring a GUI which allows integration within the General Video Game AI Framework. Users are able to easily customize the parameters of the agent between runs and observe an in-depth analysis of its performance through various visual information extracted from gameplay data, live while playing the game. This visualisation aims to inform a deeper analysis into algorithm behaviour, in an attempt to justify why they make the decisions they do and improve their performance based on this knowledge.
Modelling a player's understanding of NPC movements can be useful for adapting gameplay to different play styles. For stealth games, what a player knows or suspects of enemy movements is important to how they will navigate towards a solution. In this work, we build a uniform abstraction of potential player path knowledge based on their partial observations. We use this representation to compute different path estimates according to different player expectations. We augment our work with a user study that validates what kinds of NPC behaviour a player may expect, and develop a tool that can build and explore appropriate (expected) paths. We find that players prefer short simple paths over long or complex paths with looping or backtracking behaviour.
3D construction sandbox games such as Minecraft have provided new opportunities for people to express their creativity. However, individual players have few tools to help them learn about architectural style or how to improve the structure they are building. Ideally, players could utilize tools that capitalize on the large numbers of 3D models built by others to offer guidance for their particular project. We trained a neural network to classify a large collection of Minecraft models from various websites in terms of style (Ancient, Asian, Medieval, or Modern). We present experimental results demonstrating that our model can classify the user-indicated style of a structure with 55% accuracy. We further demonstrate use of this model to highlight nearest neighbors to a specific query structure. We have integrated these tools into a Minecraft Mod that allows players to classify their structure's style and view nearest neighbors in real-time.
Wiggins, Joseph B. (University of Florida) | Kulkarni, Mayank (University of Florida) | Min, Wookhee (North Carolina State University) | Mott, Bradford (North Carolina State University) | Boyer, Kristy Elizabeth (University of Florida) | Wiebe, Eric (North Carolina State University) | Lester, James (North Carolina State University)
Player affect is a central consideration in the design of game-based learning environments. Affective indicators such as facial expressions exhibited during gameplay may support building more robust player models and adaptation modules. In game-based learning, predicting player mental demand and engagement from player affect is a particularly promising approach to helping create more effective gameplay. This paper reports on a predictive player-modeling approach that observes player affect during early interactions with a game-based learning environment and predicts selfreports of mental demand and engagement at the conclusion of gameplay sessions. The findings show that automatically detected facial expressions such as those associated with joy, disgust, sadness, and surprise are significant predictors of players' self-reported engagement and mental demand at the end of gameplay interactions. The results suggest that it is possible to create affect-based predictive player models that can enable proactively tailored gameplay by anticipating player mental demand and engagement.
In this paper we tackle a problem of tile-based combat in the turn-based strategy (space 4X) video game Children of the Galaxy (CotG). We propose an improved version of Monte Carlo tree search (MCTS) called MCTS considering hit points (MCTS HP). We show MCTS HP is superior to Portfolio greedy search (PGS), MCTS and NOKAV reactive agent in small to medium combat scenarios. MCTS HP performance is shown to be stable when compared to PGS, while it is also more time-efficient than regular MCTS. In smaller scenarios, the performance of MCTS HP with 100 millisecond time limit is comparable to MCTS with 2 seconds time limit. This fact is crucial for CotG as the combat outcome assessment is precursor to many strategical decisions in CotG game. Finally, if we fix the amount of search time given to the combat agent, we show that different techniques dominate different scales of combat situations. As the result, if searchbased techniques are to be deployed in commercial products, a combat agent will need to be implemented with portfolio of techniques it can choose from given the complexity of situation it is dealing with to smooth gameplay experience for human players.
Authoring in the context of Interactive Storytelling (IS) is inherently difficult, and there is a need for authoring tools that both enable and assist authors in the creation of new content. In this paper, we discuss our approach for creating an AI-assisted authoring tool via the concept of mixed-initiative systems. We introduce our tool, Mimisbrunnur, which uses this concept to assist authors in the creation of story content. We explain how the tool functions and introduce its fundamental components, including Natural Language Processing, a Suggestion Generator, and three authoring modules.
First we compare the believability of agent plans taken from the spaces of valid classical plans, intentional plans, and belief plans. We show that the plans that make the most sense to humans are those in the overlapping regions of the intentionality and belief spaces. Second, we validate the model's approach to representing anticipation, where characters form plans that involve actions they expect other characters to take. Using a short interactive scenario we demonstrate that players not only find it believable when NPCs anticipate their actions, but sometimes actively anticipate the actions of NPCs in a way that is consistent with the model.
Modeling player engagement is a key challenge in games. However, the gameplay signatures of engaged players can be highly context-sensitive, varying based on where the game is used or what population of players is using it. Traditionally, models of player engagement are investigated in a particular context, and it is unclear how effectively these models generalize to other settings and populations. In this work, we investigate a Bayesian hierarchical linear model for multi-task learning to devise a model of player engagement from a pair of datasets that were gathered in two complementary contexts: a Classroom Study with middle school students and a Laboratory Study with undergraduate students. Both groups of players used similar versions of Crystal Island, an educational interactive narrative game for science learning. Results indicate that the Bayesian hierarchical model outperforms both pooled and context-specific models in cross-validation measures of predicting player motivation from in-game behaviors, particularly for the smaller Classroom Study group. Further, we find that the posterior distributions of model parameters indicate that the coefficient for a measure of gameplay performance significantly differs between groups. Drawing upon their capacity to share information across groups, hierarchical Bayesian methods provide an effective approach for modeling player engagement with data from similar, but different, contexts.
A significant amount of work has advocated that Learning from Demonstration (LfD) is a promising approach to allow end-users to create behaviors for in-game characters without requiring programming. However, one major problem with this approach is that many LfD algorithms require large amounts of training data, and thus are not practical for learning from human demonstrators. In this paper, we focus on LfD with limited training data, and specifically on the problem of active LfD where the demonstrators are human. We present the results of a user study in comparing SALT, a new active LfD approach, versus a previous state-of-the-art Active LfD algorithm, showing that SALT significantly outperforms it when learning from a limited amount of data in the context of learning to play a puzzle video game.