https://ojs.aaai.org/index.php/AIIDE/issue/feed 2024-11-15T00:00:00-08:00 Managing Editor publications19@aaai.org Open Journal Systems <p>The AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment (AIIDE) is intended to be the definitive point of interaction between entertainment software developers interested in AI and academic and industrial AI researchers. Sponsored by the Association for the Advancement of Artificial Intelligence (AAAI), the conference proceedings reflects the work of &nbsp;both the research and commercial communities, and promotes AI research and practice in the context of interactive digital entertainment systems with an emphasis on commercial computer and video games.</p> https://ojs.aaai.org/index.php/AIIDE/article/view/31891 2024-11-14T22:13:25-08:00 Rogelio Enrique Cardona-Rivera rogelio@example.com Seth Cooper sethcopper@example.com

The 20th AAAI conference on Artificial Intelligence and Interactive Digital Entertainment was held at the University of Kentucky, in Lexington, Kentucky, USA.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31874 2024-11-14T22:12:08-08:00 Mahsa Bazzaz bazzaz.ma@northeastern.edu Seth Cooper se.cooper@northeastern.edu

Procedurally generated levels created by machine learning models can be unsolvable without further editing. Various methods have been developed to automatically repair these levels by enforcing hard constraints during the post-processing step. However, as levels increase in size, these constraint-based repairs become increasingly slow. This paper proposes using explainability methods to identify specific regions of a level that contribute to its unsolvability. By assigning higher weights to these regions, constraint-based solvers can prioritize these problematic areas, enabling more efficient repairs. Our results, tested across three games, demonstrate that this approach can help to repair procedurally generated levels faster.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31875 2024-11-14T22:12:12-08:00 Mahsa Bazzaz bazzaz.ma@northeastern.edu Seth Cooper se.cooper@northeastern.edu

Generative Adversarial Networks (GANs) are unsupervised models designed to learn and replicate a target distribution. The vanilla versions of these models can be extended to more controllable models. Conditional Generative Adversarial Networks (CGANs) extend vanilla GANs by conditioning both the generator and discriminator on some additional information (labels). Controllable models based on complementary learning, such as Rumi-GAN, have been introduced. Rumi-GANs leverage negative examples to enhance the generator's ability to learn positive examples. We evaluate the performance of two controllable GAN variants, CGAN and Rumi-GAN, in generating game levels targeting specific constraints of interest: playability and controllability. This evaluation is conducted under two scenarios: with and without the inclusion of negative examples. The goal is to determine whether incorporating negative examples helps the GAN models avoid generating undesirable outputs. Our findings highlight the strengths and weaknesses of each method in enforcing the generation of specific conditions when generating outputs based on given positive and negative examples.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31876 2024-11-14T22:12:16-08:00 Steph Buongiorno sbuongiorno@smu.edu Lawrence Klinkert jklinkert@mail.smu.edu Zixin Zhuang zixinzhuang@smu.edu Tanishq Chawla tchawla@mail.smu.edu Corey Clark coreyc@smu.edu

Large language models (LLMs) offer unprecedented flexibility in procedural generation, enabling the creation of dynamic video game storylines that evolve with user input. A critical aspect of realizing this potential is allowing players and developers to provide dynamic or free-form text to drive generation. Ingesting free-form text for a video game poses challenges, however, as it can prompt the LLM to generate content beyond the intended narrative scope. In response to this challenge, this research introduces Procedural Artificial Narrative using Generative AI (PANGeA) for leveraging LLMs to create narrative content for turn-based, role-playing games (RPGs). PANGeA is an approach comprised of components including a memory system, validation system, a Unity game engine plug-in, and a server with a RESTful interface that enables connecting PANGeA components with any game engine as well as accessing local and private LLMs. PANGeA procedurally generates level data like setting, key items, non-playable characters (NPCs)), and dialogue based on a set of configuration and design rules provided by the game designer. This process is supported by a novel validation system for handling free-form text input during game development and gameplay, which aligns LLM generation with the narrative. It does this by evoking the LLM's capabilities to dynamically evaluate the text input against game rules that reinforce the designer's initial criteria. To enrich player-NPC interactions, PANGeA uses the Big Five Personality model to shape NPC responses. To explore its broad application, PANGeA is evaluated across two studies. First, this research presents a narrative test scenario of the prototype game, Dark Shadows, which was developed using PANGeA within the Unity game engine. This is followed by an ablation study that tests PANGeA's performance across 10 different role-playing game scenarios–from western to science fiction–and across three model sizes: Llama-3 (8B), GPT-3.5, and GPT-4. These evaluations demonstrate that PANGeA's NPCs can hold dynamic, narrative-consistent conversations that, without the memory system, would exceed the LLM's context length. In addition, the results demonstrate PANGeA's validation system not only aligns LLM responses with the game narrative but also improves the performance of Llama-3 (8B), enabling it to perform comparably to large-scale foundational models like GPT-4. With the validation system, Llama-3 (8B)'s performance improved from 28% accuracy to 98%, and GPT-4's from 71% to 99%. These findings indicate PANGeA can help game designers generate narrative-consistent content while leveraging LLMs of different sizes, suitable for various devices.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31877 2024-11-14T22:12:20-08:00 Mahdi Farrokhi Maleki mahdi.farrokhimaleki@ucalgary.ca Richard Zhao richard.zhao1@ucalgary.ca

Procedural Content Generation (PCG) is defined as the automatic creation of game content using algorithms. PCG has a long history in both the game industry and the academic world. It can increase player engagement and ease the work of game designers. While recent advances in deep learning approaches in PCG have enabled researchers and practitioners to create more sophisticated content, it is the arrival of Large Language Models (LLMs) that truly disrupted the trajectory of PCG advancement. This survey explores the differences between various algorithms used for PCG, including search-based methods, machine learning-based methods, other frequently used methods (e.g., noise functions), and the newcomer, LLMs. We also provide a detailed discussion on combined methods. Furthermore, we compare these methods based on the type of content they generate and the publication dates of their respective papers. Finally, we identify gaps in the existing academic work and suggest possible directions for future research.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31878 2024-11-14T22:12:24-08:00 Pedro M. Fernandes pedro.miguel.rocha.fernandes@tecnico.ulisboa.pt Manuel Lopes manuel.lopes@tecnico.ulisboa.pt Rui Prada rui.prada@tecnico.ulisboa.pt

This paper proposes a novel pipeline for generating game levels that elicit predefined emotional experiences from players. Our approach uses evolutionary algorithms alongside data-driven persona agents, predictive emotional models, a PCG parametric level generator, and a newly defined language for the clear and computable definition of player emotional experiences: ExpREx (Experience Regular Expressions). Using these components, we evolve game levels to match the player experience goals specified using the ExpREx language, aiming to create levels that evoke specific emotional experiences for different subsets of players. The efficacy of our method was validated through a user study involving 101 participants, whose continuous annotations of emotional experience were collected and analyzed to assess the congruence between the actual emotional responses elicited and those targeted by our pipeline. We found that 93.73% of the ExpREx goals targeted were also reported by the user study subjects.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31879 2024-11-14T22:12:30-08:00 Sara Karimi sara.karimi@king.com Sahar Asadi sahar.asadi@king.com Amir H. Payberah payberah@kth.se

Solving complex long-horizon tasks through Reinforcement Learning (RL) from scratch presents challenges related to efficient exploration. Two common approaches to reduce complexity and enhance exploration efficiency are (i) integrating learning-from-demonstration techniques with online RL, where the prior knowledge acquired from demonstrations is used to guide exploration, refine representations, or tailor reward functions, and (ii) using representation learning to facilitate state abstraction. In this study, we present Skill-Conditioned Online REinforcement Learning (SCORE), a novel approach that leverages these two strategies and utilizes skills acquired from an unstructured demonstrations dataset in a policy gradient RL algorithm. This integration enriches the algorithm with informative input representations, improving downstream task learning and exploration efficiency. We evaluate our method on long-horizon robotic and navigation tasks and game environments, demonstrating enhancements in online RL performance compared to the baselines. Furthermore, we show our approach’s generalization capabilities and analyze its effectiveness through an ablation study.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31880 2024-11-14T22:12:34-08:00 Yazeed Mahmoud yazeed1@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Designing and tuning a game is a complex creative process, with a variety of tools have been designed for assisting human designers in this process. But, many existing tools either entirely take the design away from humans, or they only assist in generating content - they cannot be used to create an entire game from scratch. This paper explores how Exhaustive Procedural Content Generation (EPCG) can be integrated as a co-creative partner in the design process. We describe how EPCG was used alongside human designers to create a novel Hexagon Tangram puzzle, from initial conception to puzzle curriculum. EPCG is used to answer design questions at each step of the design, which enable human designers to make more informed design choices.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31881 2024-11-14T22:12:39-08:00 Dave Mobley dave.mobley@uky.edu Adrienne Corwin adriennecorwin@gmail.com Brent Harrison harrison@cs.uky.edu

This work investigates how natural language task descriptions can accelerate reinforcement learning in games. Recognizing that human descriptions often imply a hierarchical task structure, we propose a method to extract this hierarchy and convert it into "options" – policies for solving subtasks. These options are generated by grounding natural language descriptions into environment states, which are then used as task boundaries to learn option policies either by leveraging prior successful traces or from human created walkthroughs. We evaluate our approach in both a simpler grid-world environment and the more complex text-based game Zork, comparing option-based agents against standard Q-learning and random agents. Our results demonstrate the effectiveness of incorporating natural language task knowledge for faster and more efficient reinforcement learning across different environments and Q-learning algorithms, including tabular Q-learning and Deep Q-Networks.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31882 2024-11-14T22:12:43-08:00 Teresa Saller teresa_saller@freenet.de Ramon Lawrence ramon.lawrence@ubc.ca Vadim Bulitko bulitko@ualberta.ca

Efficient grid-based path finding is important in video games especially for larger maps and when moving many agents. Algorithms based on abstraction have an order of magnitude faster search time performance than A* at the cost of a small amount of memory and increased suboptimality. Some approaches also compute and store paths to improve search performance. This paper evaluates new and existing algorithm variants for abstraction and path computation and investigates their performance for video game path finding with map changes. The results show that abstraction has significant advantages over A* and can be implemented efficiently for changing maps. Computing, storing, and reusing paths also has benefits especially when several searches can be performed before the map changes.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31883 2024-11-14T22:12:47-08:00 Fiona Shyne shyne.f@northeastern.edu Seth Cooper se.cooper@northeastern.edu

Creating well-balanced combat encounters can be a difficult task for Game Managers (GMs) in tabletop games such as Dungeons and Dragons (DnD). This work uses a simulation environment to generate new sets of DnD encounters that can be optimized for both difficulty and balance among player contributions. Encounters are evaluated using simulated games that can either be run probabilistically (using dice rolls) or with deterministic expected outcomes. While the expected approach allows game outcomes to be simulated substantially faster and is a good estimate of difficulty, it is a less reliable measure of balance. A genetic algorithm was used to generate encounters that meet the desired difficulty and where all players are needed for success.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31884 2024-11-14T22:12:52-08:00 Megan Sumner msumner@ualberta.ca Vardan Saini vardan1@ualberta.ca Matthew Guzdial guzdial@ualberta.ca

Game development is a highly technical practice that traditionally requires programming skills. This serves as a barrier to entry for would-be developers or those hoping to use games as part of their creative expression. While there have been prior game development tools focused on accessibility, they generally still require programming, or have major limitations in terms of the kinds of games they can make. In this paper we introduce Mechanic Maker, a tool for creating a wide-range of game mechanics without programming. It instead relies on a backend symbolic learning system to synthesize game mechanics from examples. We conducted a user study to evaluate the benefits of the tool for participants with a variety of programming and game development experience. Our results demonstrated that participants' ability to use the tool was unrelated to programming ability. We conclude that tools like ours could help democratize game development, making the practice accessible regardless of programming skills.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31885 2024-11-14T22:12:56-08:00 Kristen K. Yu kkyu@ualberta.ca Matthew Guzdial guzdial@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Modern commercial games are designed for mass appeal, not for individual players, but there is a unique opportunity in video games to better fit the individual through adapting elements in games. In this paper, we focus on AI Directors, systems which can dynamically modify a game, that attempt to personalize the player experience to a player's preference. AI Directors have in the past provided inconclusive results, so their effect on player experience is currently unknown. In this paper, we take three AI Directors and directly compare them in a human subject study to test their effectiveness on quest selection. Our results show that a non-random AI Director provides a better player experience than a random AI Director.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31890 2024-11-14T22:13:19-08:00 Elinor Rubin-McGregor ellirubimac@gmail.com

Adding character depth in narrative planners is a complex subject with many approaches. One common strategy used to add depth to characters is to imbue them with personality traits. One attribute that few personality models consider, however, is how character behavior should change when the character is exposed to reminders of powerful psychological influences. Through creating and refining a new digital construct known as Personality Exceptions, I will provide a tool that makes it much easier for personality systems to support such conditions. I intend to make a modular tool that researchers can attach to existing personality models to enhance the character depth of their characters and, thus, improve the overall quality of computational narratives.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31886 2024-11-14T22:13:02-08:00 Seth Cooper se.cooper@northeastern.edu

2D tile-based levels are a common format for video game research, particularly in procedural content generation. Often, tiles are represented as text characters. Here, we describe level2image, a utility that provides a flexible means for converting such levels into a variety of image formats. It can use text tiles or substitute image tiles or backgrounds. There is support for various geometry overlays, such as paths through the level, areas of interest, or boundaries. The utility is a Python script, intended to reduce duplicated work creating such converters within the game research community.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31887 2024-11-14T22:13:06-08:00 Sage Deo sagevdeo@gmail.com Jonathan Chung jonathanchung1113@gmail.com Joshua McCoy jamccoy@ucdavis.edu

Incremental story sifters analyze an in-progress simulation to extract interesting narrative content or find a set of events that have the potential to become more narratively interesting if followed up in a certain way. There has been some investigation on the potential for an incremental story sifter to suggest future narrative events to human authors, but the technique of guiding a simulation using story sifting, without any human interference, remains completely unexplored. Thus, we present Shepherd, a drama manager powered by incremental story sifting, which guides otherwise completely autonomous characters toward making narratively interesting choices.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31888 2024-11-14T22:13:10-08:00 Yazeed Mahmoud yazeed1@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

This paper describes the artifact associated with a paper on the use of AI tools for the creation of a Hexagram Tangram Puzzle currently under review at AIIDE. We describe our motivation for creating the puzzle, the layout and construction of the puzzle, and what is publicly available with the artifact.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31889 2024-11-14T22:13:14-08:00 Kristen K. Yu kkyu@ualberta.ca Matthew Guzdial guzdial@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Open datasets are the foundation of many types of academic research. One field that requires datasets is data-driven player modeling, but there is a lack of variety in existing datasets. We introduce the FarmQuest Player Telemetry (FPT) dataset, a new playthrough dataset of a cozy farming game. We envision this dataset will be used for reproducing prior results, evaluating initial ideas, and evaluating the generalizability of new and existing algorithms.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31861 2024-11-14T22:11:07-08:00 Toby Best t.j.best@qmul.ac.uk Simon Lucas simon.lucas@qmul.ac.uk Raluca Gaina r.d.gaina@qmul.ac.uk

The Game Master is a player role synonymous with many tabletop games. The asymmetric gameplay of the role provides different opportunities compared to other players, and can be both cooperative and competitive with the other players in the same game. Though complex environments for exploring human and Artificial Intelligence collaboration exist, few focus on the Game Master role's semi-cooperative play. Here, we propose a new complex environment based on the board game `Descent: Journeys in the Dark (Second Edition)', as part of the Tabletop Games Framework, showcasing one-versus-many play, tactical combat, and large, dynamic action and state spaces. We include baseline AI player performance of Monte Carlo Tree Search agents in this game, finding them to be well-adept at considering multiple possible end-game conditions compared to the greedy One Step Look Ahead agents. In-depth analysis reveals interesting behaviours and Hero synergies, with the aim of informing the design of games and AI models to enhance human experience in semi-cooperative environments.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31862 2024-11-14T22:11:12-08:00 Seth Cooper se.cooper@northeastern.edu Mahsa Bazzaz bazzaz.ma@northeastern.edu

Procedurally generated game levels should be completable. The representation used for levels and game mechanics impacts the types of games for which different techniques can be applied. Previous work used a constraint solving approach to simultaneously generate levels with example playthroughs, showing they can be completed using the game's mechanics. However, that work used 2D grid-based rewrite rules. In this work, we extend previous approaches by representing levels as more general graphs, and game mechanics as rewrites on node and edge labels of subgraphs. Using this approach, graph-based levels with playthroughs are generated. We describe the approach and demonstrate its application in some games with graph-based levels.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31863 2024-11-14T22:11:16-08:00 Kaylah Facey kaylah.facey@gmail.com Seth Cooper se.cooper@northeastern.edu

WaveFunctionCollapse (WFC) is a constraint-satisfaction-based approach to procedural content generation via machine learning (PCGML). It is relatively easy to implement and requires very little training data, making it a popular approach. Generated game levels are guaranteed to look locally similar totheir example tilemaps; however, local adjacency rules often fail to capture global solvability rules, potentially making many such levels unplayable. Existing approaches to improving the solvability of WFC-generated levels typically require adding additional game-specific information in the form of global constraints, substantially increasing the complexity and time required for setup. The purpose of this work is to explore whether using level solutions as training data can allow WFC to learn solvability constraints and game mechanics. We have implemented a novel space-time approach that uses three-dimensional space-time blocks representing solutions to 2D levels as both input and output. Experiments using this method show that space-time WFC is capable of demonstrating localized game mechanics and creating small playable levels with given solutions. However, levels are slow to generate, and some high-level constraints are still not captured.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31864 2024-11-14T22:11:20-08:00 Mira Fisher mira.fisher@uky.edu Stephen Ware sgware@gmail.com

Contemporary automated planning research emphasizes the use of domain knowledge abstractions like heuristics to improve search efficiency. Transformative automated abstraction techniques which decompose or otherwise reformulate the problem have a limited presence, owing to poor performance in key metrics like plan length and time efficiency. In this paper, we argue for a reexamination of these transformative techniques in the context of narrative planning, where classical metrics are less appropriate. We propose a model for automating abstraction by decomposing a planning problem into subproblems which serve as abstract features of the problem. We demonstrate the application of this approach on a low-level problem and discuss key features of the resulting abstract problem. Plans in the abstract problem are shorter, representing summaries of low-level plans, but can be directly translated into low-level plans for the original problem.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31865 2024-11-14T22:11:24-08:00 Dave Goel dgoel1@ualberta.ca Athar Mahmoudi-Nejad athar1@ualberta.ca Matthew Guzdial guzdial@ualberta.ca

Player Experience Modelling (PEM) is the study of AI techniques applied to modelling a player's experience within a video game. PEM development can be labour-intensive, requiring expert hand-authoring or specialized data collection. In this work, we propose a novel PEM development approach, approximating player experience from gameplay video. We evaluate this approach predicting affect in the game Angry Birds via a human subject study. We validate that our PEM can strongly correlate with self-reported and sensor measures of affect, demonstrating the potential of this approach.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31866 2024-11-14T22:11:30-08:00 Ian Horswill ian@northwestern.edu Samuel Hill samuelhill2022@u.northwestern.edu

Logic programming and rule-based systems are often chosen for tasks such as social simulation because their use of declarative rules and predicates map well to rules of social engagement. Unfortunately, they are often quite slow, due in part to its heavy use of pointer chasing, dynamic allocation, garbage collection, and runtime type-checking, making it difficult to use for large numbers of characters or high-frequency updates. For appropriate tasks, bottom-up execution of logic programs can provide the declarativity of logic programming without its performance issues. We argue that large-scale character simulations are a “sweet spot” for bottom-up LP. We present a language, TED, that combines declarativity with excellent performance. TED can be used with any game engine supporting C#. In head-to-head comparisons TED code was 2-3 orders of magnitude faster than Prolog, 2-5 times more compact than C#, and only 25% slower than C#. It is used in both the research game Voix de la Ville and the upcoming commercial game Rise of Industry 2.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31867 2024-11-14T22:11:34-08:00 Lawrence J. Klinkert jklinkert@smu.edu Steph Buongiorno sbuongiorno@smu.edu Corey Clark coreyc@smu.edu

To enhance immersion and engagement in video games, the design of Affective Non-Player Characters (NPCs) is a key focus for researchers and practitioners. Affective Computing frameworks improve Non-player characters (NPC) by providing personalities, emotions, and social relations. Large Language Models (LLMs) bring the promise to dynamically enhance character design when coupled with these frameworks, but further research is needed to validate the models truly represent human qualities. In this research, a comprehensive analysis investigates the capabilities of LLMs to generate content that aligns with human personality, using the Big Five and human responses from the International Personality Item Pool (IPIP) questionnaire. Our goal is to benchmark the performance of various LLMs, including frontier models and local models, against an extensive dataset comprising over 50,000 human surveys of self-reported personality tests to determine whether LLMs can replicate human-like decision-making with personality-driven prompts. A range of personality profiles were used to cluster the test results from the human survey dataset. Our methodology involved prompting LLMs with self-evaluated test items for each personality profile, comparing their outputs to human baseline responses, and evaluating the accuracy and consistency. Our findings show that some local models had 0% alignment of any personality profiles when compared to the human dataset, while the frontier models, in some cases, had 100% alignment. The results indicate that NPCs can successfully emulate human-like personality traits using LLMs, as demonstrated by benchmarking the LLM's output against human data. This foundational work serves as a methodology for game developers and researchers to test and evaluate LLMs, ensuring they accurately represent the desired human personalities and can be expanded for further validation.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31868 2024-11-14T22:11:38-08:00 Vikram Kumaran vkumara@ncsu.edu Jonathan Rowe jprowe@ncsu.edu James Lester lester@ncsu.edu

Interactive narrative in games utilize a combination of dynamic adaptability and predefined story elements to support player agency and enhance player engagement. However, crafting such narratives requires significant manual authoring and coding effort to translate scripts to playable game levels. Advances in pretrained large language models (LLMs) have introduced the opportunity to procedurally generate narratives. This paper presents NarrativeGenie, a framework to generate narrative beats as a cohesive, partially ordered sequence of events that shapes narrative progressions from brief natural language instructions. By leveraging LLMs for reasoning and generation, NarrativeGenie, translates a designer’s story overview into a partially ordered event graph to enable player-driven narrative beat sequencing. Our findings indicate that NarrativeGenie can provide an easy and effective way for designers to generate an interactive game episode with narrative events that align with the intended story arc while at the same time granting players agency in their game experience. We extend our framework to dynamically direct the narrative flow by adapting real-time narrative interactions based on the current game state and player actions. Results demonstrate that NarrativeGenie generates narratives that are coherent and aligned with the designer’s vision.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31869 2024-11-14T22:11:43-08:00 Tim Merino tm3477@nyu.edu Sam Earle sam.earle@nyu.edu Ryan Sudhakaran ryan@jesterlabs.ai Shyam Sudhakaran shyam@jesterlabs.ai Julian Togelius julian@togelius.com

The Connections puzzle is a word association game published daily by The New York Times (NYT). In this game, players are asked to find groups of four words that are connected by a common theme. While solving a given Connections puzzle requires both semantic knowledge and abstract reasoning, generating novel puzzles additionally requires a form of metacognition: generators must be able to accurately model the downstream reasoning of potential solvers. In this paper, we investigate the ability of the GPT family of Large Language Models (LLMs) to generate challenging and creative word games for human players. We start with an analysis of the word game Connections and the unique challenges it poses as a Procedural Content Generation (PCG) domain. We then propose a method for generating Connections puzzles using LLMs by adapting a Tree of Thoughts (ToT) prompting approach. We evaluate this method by conducting a user study, asking human players to compare AI- against human-generated Connections puzzles. Our findings show that LLMs are capable puzzle creators, and can generate diverse sets of enjoyable, challenging, and creative Connections puzzles as judged by human users.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31870 2024-11-14T22:11:48-08:00 Darryl Roman darryl.roman@ucf.edu Noah Ari noah.ari@ucf.edu Johnathan Mell johnathanmell@me.com

Team-based games are a keystone pillar of the gaming industry. Sadly, the understanding of team dynamics—and the recommendations for both human and AI-based teammates—are based on a rudimentary understanding of human-AI teaming. We propose a superior metric, which provides information about team effectiveness in an efficient and easily-replicable manner. Without an accurate and effective metric for team evaluation, it is nigh-impossible to provide feedback to players and game designers to improve team balance. We provide such a metric. The Harmony Index, a novel algorithm using real-world data, provides simpler and more accurate actionable directives to improve game design across MOBAs and other game genres. We prove its predictive power in a separate analysis and make recommendations for its use in assessing team effectiveness as well as its future use in additional domains.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31871 2024-11-14T22:11:53-08:00 Kimiya Saadat kimiya.saadat@ucalgary.ca Richard Zhao richard.zhao1@ucalgary.ca

Playing two-player games using reinforcement learning and self-play can be challenging due to the complexity of two-player environments and the possible instability in the training process. We propose that a reinforcement learning algorithm can train more efficiently and achieve improved performance in a two-player game if it leverages the knowledge from the single-player version of the same game. This study examines the proposed idea in ten different Atari 2600 environments using the Atari 2600 RAM as the input state. We discuss the advantages of using transfer learning from a single-player training process over training in a two-player setting from scratch, and demonstrate our results in a few measures such as training time and average total reward. We also discuss a method of calculating RAM complexity and its relationship to performance.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31872 2024-11-14T22:11:58-08:00 Junwen Shen junwen5@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Metrics for problem difficulty are used by many puzzle generation algorithms, as well as by adaptive algorithms that want to provide players with the puzzles at the correct level of difficulty. A recently proposed general metric, puzzle entropy, combines an analysis of game mechanics with a model of player knowledge in the form of inference rules to predict problem difficulty. The entropy of a puzzle is the amount of information required, given a player’s knowledge about the puzzle, to describe a solution to a puzzle. This paper generalizes the concepts of puzzle entropy and solution information, providing a better foundation for the previous work and creating new algorithms, Minimum Solution Information and Total Solution Information. While functionally very similar to past work, the new algorithm allows knowledge about a puzzle to be represented as a policy, something that can be learned more easily. We then evaluate the impact of policies, inference rules, and player knowledge in the 2016 game, The Witness.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/31873 2024-11-14T22:12:02-08:00 Fiona Shyne shyne.f@northeastern.edu Kaylah Facey facey.k@northeastern.edu Seth Cooper se.cooper@northeastern.edu

Determining if and how the difficulty of algorithmic puzzle solvers is related to the difficulty and enjoyment for human players is a challenging task. In this work, we explored this relationship using logic grid puzzles. We used an algorithmic solver to estimate the difficulty of the puzzles by capturing the number of ``solver loops'' through the algorithm. This characteristic was used to generate and evaluate a set of puzzles of varying algorithmic difficulty using constrained MAP-Elites. Then, we ran a user study to gather information on the player experience of these puzzles. We tested the relationship between solver loops and player experience on generated puzzles and found that the number of solver loops is statistically significantly correlated with subjective perception of difficulty and borderline statistically significantly correlated with puzzle correctness.

2024-11-15T00:00:00-08:00 Copyright (c) 2024 Association for the Advancement of Artificial Intelligence