https://ojs.aaai.org/index.php/AIIDE/issue/feed 2023-10-06T16:32:20-07:00 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/27496 2023-10-06T16:30:20-07:00 Frederic Abraham fm.abraham@student.maastrichtuniversity.nl Matthew Stephenson matthew.stephenson888@gmail.com

This paper investigates the suitability of using Generative Adversarial Networks (GANs) to generate stable structures for the physics-based puzzle game Angry Birds. While previous applications of GANs for level generation have been mostly limited to tile-based representations, this paper explores their suitability for creating stable structures made from multiple smaller blocks. This includes a detailed encoding/decoding process for converting between Angry Birds level descriptions and a suitable grid-based representation, as well as utilizing state-of-the-art GAN architectures and training methods to produce new structure designs. Our results show that GANs can be successfully applied to generate a varied range of complex and stable Angry Birds structures.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27497 2023-10-06T16:30:23-07:00 Adam Amos-Binks aamosbinks@ara.com Bryan S. Weber bryan.weber@csi.cuny.edu

Managing risk with imperfect information is something humans do every day, but we have little insight into the abilities of AI agents to do so. We define two risk management strategies and perform an ability-based evaluation using StarCraft agents. Our evaluation shows that nearly all agents mitigate risks after observing them (react), and many prepare for such risks before their appearance (anticipate). For this evaluation, we apply traditional causal effect inference and causal random forest methods to explain agent behavior. The results highlight different risk management strategies among agents, others strategies that are common to agents, and overall encourage evaluating agent risk management abilities in other AI domains.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27498 2023-10-06T16:30:27-07:00 Michael Beukman mcbeukman@gmail.com Branden Ingram branden.ingram@gmail.com Ireton Liu 2089889@students.wits.ac.za Benjamin Rosman benjamin.rosman1@wits.ac.za

Video game developers are increasingly utilising procedural content generation (PCG) techniques in order to generate more content far quicker than if it were designed. Although promising, much of the successful work to date has been achieved in simple 2D environments or has required significant hand-designed effort. This is due to the difficult nature of defining plausible metrics, fitness functions or reward functions which can quantify the quality of generated levels. Our work aims to avoid this difficulty by utilising minimal human design to build up constraints, and generating diverse levels that maintain these constraints. We achieve this by hierarchically applying the recent WaveFunction collapse (WFC) algorithm. Our approach allows designers to specify larger-scale components, and additional constraints that are difficult to enforce using standard WFC. We empirically demonstrate that our approach does indeed incorporate these higher-level structures, and is more controllable than our baselines. Despite these benefits, our levels do not suffer from a lack of diversity. Finally, we illustrate the scalability and flexibility of our approach by applying it to both 2D and 3D domains.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27499 2023-10-06T16:30:29-07:00 Eugene You Chen Chen eugene@ideaowl.com Adam White amw8@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Evaluating and ranking the difficulty and enjoyment of puzzles is important in game design. Typically, such rankings are constructed manually for each specific game, which can be time consuming, subject to designer bias, and requires extensive play testing. An approach to ranking that generalizes across multiple puzzle games is even more challenging because of their variation in factors like rules and goals. This paper introduces two general approaches to compute puzzle entropy, and uses them to evaluate puzzles that players enjoy. The resulting uncertainty score is equivalent to the number of bits of data necessary to communicate the solution of a puzzle to a player of a given skill level. We apply our new approaches to puzzles from the 2016 game, The Witness. The computed entropy scores largely reproduce the order of a set of puzzles that introduce a new mechanic in the game. The scores are also positively correlated with the user ratings of user-created Witness puzzles, providing evidence that our approach captures notions of puzzle difficulty and enjoyment. Our approach is designed to exploit game-specific knowledge in a general way and thus can extended to provide automatic rankings or curricula in a variety of applications.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27500 2023-10-06T16:30:33-07:00 Logan Fields ldfields@usf.edu John Licato licato@usf.edu

Effectively identifying an individual and predicting their future actions is a material aspect of player analytics, with applications for player engagement and game security. Collectible card games are a fruitful test space for studying player identification, given that their large action spaces allow for flexibility in play styles, thereby facilitating behavioral analysis at the individual, rather than the aggregate, level. Further, once players are identified, modeling the differences between individuals may allow us to preemptively detect patterns that foretell future actions. As such, we use the virtual collectible card game "Legends of Code and Magic" to research both of these topics. Our main contributions to the task are the creation of a comprehensive dataset of Legends of Code and Magic game states and actions, extensive testing of the minimum information and computational methods necessary to identify an individual from their actions, and examination of the transferability of knowledge collected from a group to unknown individuals.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27501 2023-10-06T16:30:35-07:00 Chathura Gamage chathura.gamage@anu.edu.au Vimukthini Pinto vimukthini.inguruwattage@anu.edu.au Matthew Stephenson matthew.stephenson@flinders.edu.au Jochen Renz jochen.renz@anu.edu.au

Performing tasks in a physical environment is a crucial yet challenging problem for AI systems operating in the real world. Physics simulation-based tasks are often employed to facilitate research that addresses this challenge. In this paper, first, we present a systematic approach for defining a physical scenario using a causal sequence of physical interactions between objects. Then, we propose a methodology for generating tasks in a physics-simulating environment using these defined scenarios as inputs. Our approach enables a better understanding of the granular mechanics required for solving physics-based tasks, thereby facilitating accurate evaluation of AI systems' physical reasoning capabilities. We demonstrate our proposed task generation methodology using the physics-based puzzle game Angry Birds and evaluate the generated tasks using a range of metrics, including physical stability, solvability using intended physical interactions, and accidental solvability using unintended solutions. We believe that the tasks generated using our proposed methodology can facilitate a nuanced evaluation of physical reasoning agents, thus paving the way for the development of agents for more sophisticated real-world applications.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27502 2023-10-06T16:30:39-07:00 Jack Kelly jochkell@ucsc.edu Alex Calderwood alexcwd@ucsc.edu Noah Wardrip-Fruin nwardrip@ucsc.edu Michael Mateas mmateas@ucsc.edu

Story generators using language models offer the automatic production of highly fluent narrative content, but they are hard to control and understand, seizing creative tasks that many authors wish to perform themselves. On the other hand, planning-based story generators are highly controllable and easily understood but require story domains that must be laboriously crafted; further, they lack the capacity for fluent language generation. In this paper, we explore hybrid approaches that aim to bridge the gap between language models and narrative planners. First, we demonstrate that language models can be used to author narrative planning domains from natural language stories with minimal human intervention. Second, we explore the reverse, demonstrating that we can use logical story domains and plans to produce stories that respect the narrative commitments of the planner. In doing so, we aim to build a foundation for human-centric authoring tools that facilitate novel creative experiences.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27503 2023-10-06T16:30:41-07:00 Yeo Jin Kim ykim32@ncsu.edu Alex Goslen amgoslen@ncsu.edu Jonathan Rowe jprowe@ncsu.edu Bradford Mott bwmott@ncsu.edu James Lester lester@ncsu.edu

Devising models that reliably recognize player goals is a key challenge in creating player-adaptive games. Player goal recognition is the task of automatically recognizing the intent of a player from a sequence of observed player actions in a game environment. In open-world digital games, players often undertake suboptimal and varied sequences of actions to achieve goals, and the high degree of freedom afforded to players makes it challenging to identify sequential patterns that lead toward specific goals. To address these issues, we present a player goal recognition framework that utilizes a fine-tuned T5 language model, which incorporates our novel attention mechanism called Temporal Contrary Attention (TCA). The T5 language model enables the framework to exploit correlations between observations through non-sequential self-attention within input sequences, while TCA enables the framework to learn to eliminate goal hypotheses by considering counterevidence within a temporal window. We evaluate our approach using game trace data collected from 144 players' interactions with an open-world educational game. Specifically, we investigate the predictive capacity of our approach to recognize player goals as well as player plans represented as abstract actions. Results show that our approach outperforms non-linguistic machine learning approaches as well as T5 without TCA. We discuss the implications of these findings for the design and development of player goal recognition models to create player-adaptive games.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27504 2023-10-06T16:30:44-07:00 Vikram Kumaran vkumara@ncsu.edu Jonathan Rowe jprowe@ncsu.edu Bradford Mott bwmott@ncsu.edu James Lester lester@ncsu.edu

Creating engaging interactive story-based experiences dynamically responding to individual player choices poses significant challenges for narrative-centered games. Recent advances in pre-trained large language models (LLMs) have the potential to revolutionize procedural content generation for narrative-centered games. Historically, interactive narrative generation has specified pivotal events in the storyline, often utilizing planning-based approaches toward achieving narrative coherence and maintaining the story arc. However, manual authorship is typically used to create detail and variety in non-player character (NPC) interaction to specify and instantiate plot events. This paper proposes SCENECRAFT, a narrative scene generation framework that automates NPC interaction crucial to unfolding plot events. SCENECRAFT interprets natural language instructions about scene objectives, NPC traits, location, and narrative variations. It then employs large language models to generate game scenes aligned with authorial intent. It generates branching conversation paths that adapt to player choices while adhering to the author’s interaction goals. LLMs generate interaction scripts, semantically extract character emotions and gestures to align with the script, and convert dialogues into a game scripting language. The generated script can then be played utilizing an existing narrative-centered game framework. Through empirical evaluation using automated and human assessments, we demonstrate SCENECRAFT’s effectiveness in creating narrative experiences based on creativity, adaptability, and alignment with intended author instructions.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27505 2023-10-06T16:30:46-07:00 Eric W. Lang ewlang@cs.utah.edu R. Michael Young young@cs.utah.edu

This paper presents MultiStyle, a multi-agent centralized heuristic search planner that incorporates distinct agent playstyles to generate solution plans where characters express individual preferences while cooperating to reach a goal. We include algorithmic details, an example domain, and multiple different solution plans generated with unique agent playstyle sets. We discuss our intent to incorporate this planner in a tool for game level designers to help them anticipate and understand how teams of players with distinct playstyles may play through their levels. Ultimately, MultiStyle generates solution plans with a novel and increased expressive range by attempting to satisfy sets of action and proposition preferences for each agent.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27506 2023-10-06T16:30:49-07:00 Timothy Merino tm3477@nyu.edu Roman Negri rvn9303@nyu.edu Dipika Rajesh dipika.rajesh@gmail.com M Charity mlc761@nyu.edu Julian Togelius julian@togelius.com

The five-dollar model is a lightweight text-to-image generative architecture that generates low dimensional images or tile maps from an encoded text prompt. This model can successfully generate accurate and aesthetically pleasing content in low dimensional domains, with limited amounts of training data. Despite the small size of both the model and datasets, the generated images or maps are still able to maintain the encoded semantic meaning of the textual prompt. We apply this model to three small datasets: pixel art video game maps, video game sprite images, and down-scaled emoji images and apply novel augmentation strategies to improve the performance of our model on these limited datasets. We evaluate our models' performance using cosine similarity score between text-image pairs generated by the CLIP VIT-B/32 model to demonstrate quality generation.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27507 2023-10-06T16:30:51-07:00 Alan Pedrassoli Chitayat alan.pchitayat@york.ac.uk Florian Block florian.block@york.ac.uk James Walker james.walker@york.ac.uk Anders Drachen anders.drachen@york.ac.uk

Esport games comprise a sizeable fraction of the global games market, and is the fastest growing segment in games. This has given rise to the domain of esports analytics, which uses telemetry data from games to inform players, coaches, broadcasters and other stakeholders. Compared to traditional sports, esport titles change rapidly, in terms of mechanics as well as rules. Due to these frequent changes to the parameters of the game, esport analytics models can have a short life-spam, a problem which is largely ignored within the literature. This paper extracts information from game design (i.e. patch notes) and utilises clustering techniques to propose a new form of character representation. As a case study, a neural network model is trained to predict the number of kills in a Dota 2 match utilising this novel character representation technique. The performance of this model is then evaluated against two distinct baselines, including conventional techniques. Not only did the model significantly outperform the baselines in terms of accuracy (85% AUC), but the model also maintains the accuracy in two newer iterations of the game that introduced one new character and a brand new character type. These changes introduced to the design of the game would typically break conventional techniques that are commonly used within the literature. Therefore, the proposed methodology for representing characters can increase the life-spam of machine learning models as well as contribute to a higher performance when compared to traditional techniques typically employed within the literature.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27508 2023-10-06T16:30:54-07:00 Justus Robertson jjrigm@rit.edu John Heiden jdh5563@rit.edu Rogelio E. Cardona-Rivera rogelio@eae.utah.edu

An interactive narrative is bound by the context of the world where its story takes place. However, most work in interactive narrative generation takes its story world design and mechanics as given, which abdicates a large part of story generation to an external world designer. In this paper, we close the story world design gap with an evolutionary search framework for generating interactive narrative worlds and mechanics. Our framework finds story world designs that accommodate multiple distinct player roles. We evaluate our system with an action agreement ratio analysis that shows worlds generated by our framework provide a greater number of in-role action opportunities compared to story worlds randomly sampled from the generative space.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27509 2023-10-06T16:30:56-07:00 Elinor Rubin-McGregor ellirubimac@gmail.com Brent Harrison harrison@cs.uky.edu Cory Siler cory.siler@uky.edu

In the field of narrative planning, implementing character personality is a challenge that’s been tackled many different ways. Most of these methods do not incorporate any method for personality to shift when characters are put in situations that would, through stress or satisfaction, naturally cause the character to behave differently than usual. Through use of situationally-triggered Personality Exceptions, we can support the generation of a story that prominently features such personality shifts as a narrative tool. This feature is made as generic as possible so that it can be attached onto a wide range of personality models in narrative generators. Through adapting Indexter’s indexes of narrative salience towards tracking internal narrative salience in the characters’ memories, we can accurately pinpoint triggers which are used to activate these personality exceptions in thematically relevant situations.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27510 2023-10-06T16:30:59-07:00 Sasha Volokh volokh@usc.edu William G.J. Halfond halfond@usc.edu

The capability to automatically explore different possible game states and functionality is valuable for the automated test and analysis of computer games. However, automatic exploration requires an exploration agent to be capable of determining and performing the possible actions in game states, for which a model is typically unavailable in games built with traditional game engines. Therefore, existing work on automatic exploration typically either manually defines a game's action space or imprecisely guesses the possible actions. In this paper we propose a program analysis technique compatible with traditional game engines, which automatically analyzes the user input handling logic present in a game to determine a discrete action space corresponding to the possible user inputs, along with the conditions under which the actions are valid, and the relevant user inputs to simulate on the game to perform a chosen action. We implemented a prototype of our approach capable of producing the action spaces of Gym environments for Unity games, then evaluated the exploration performance enabled by our technique for random exploration and exploration via curiosity-driven reinforcement learning agents. Our results show that for most games, our analysis enables exploration performance that matches or exceeds that of manually engineered action spaces, and the analysis is fast enough for real time game play.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27511 2023-10-06T16:31:01-07:00 Stephen G. Ware sgware@gmail.com Lasantha Senanayake lasantha.senanayake@uky.edu Rachelyn Farrell rachelyn.farrell@uky.edu

Narrative planning generates a sequence of actions which must achieve the author's goal for the story and must be composed only of actions that make sense for the characters who take them. A causally necessary action is one that would make the plan impossible to execute if it were left out. We hypothesize that action sequences which are solutions to narrative planning problems are more likely to feature causally necessary actions than those which are not solutions. In this paper, we show that prioritizing sequences with more causally necessary actions can lead to solutions faster in ten benchmark story planning problems.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27535 2023-10-06T16:32:02-07:00 Camila Aliaga camila.aliaga@utalca.cl Cristian Vidal cvidal@utalca.cl Gabriel K. Sepulveda gsepulveda17@alumnos.utalca.cl Nicolas Romero nromero17@alumnos.utalca.cl Fernanda González feravila.fga@gmail.com Nicolas A. Barriga nbarriga@gmail.com

Developing an original video game requires high investment levels, market research, cost-effective solutions, and a quick development process. Game developers usually reach for commercial off-the-shelf components often available in the engine's marketplace to reduce costs. Mixed-initiative authoring tools allow us to combine the thoughtful work of human designers with the productivity gains of automated techniques. However, most commercial AI-assisted Procedural Content Generation tools focus on generating small independent components, and standalone research tools available for generating full game levels with state-of-the-art algorithms usually lack integration with commercial game engines. This article aims to fill this gap between industry and academia. The Level Building Sidekick (LBS) is a mixed-initiative procedural content generation tool built by our research lab in association with four small independent game studios. It has a modular software architecture that enables developers to extend it for their particular projects. The current version has two working modules for building game maps, an early version of a module for populating the level with NPCs or items, and the first stages of a quest editor module. An automated testing module is planned. LBS is distributed as an AI-Assisted videogame-level editor Unity package. Usability testing performed using the ``Think-Aloud'' methodology indicates LBS has the potential to improve game development processes convincingly. However, at this stage, the user interface and the AI recommendations could improve their intuitiveness. As a general comment, the tool is perceived as a substantial contribution to facilitating and shortening development times, compared to only using the base game engine. There is an untapped market for mixed-initiative tools that assist the game designer in creating complete game levels. We expect to fill that market for our partner development studios and provide the community with an open research and development platform in a standard game engine.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27536 2023-10-06T16:32:04-07:00 Aaditya Bhatia aadityabhatia@gmail.com Austin Davis austinleedavis@knights.ucf.edu Soumik Ghosh soumikghosh@knights.ucf.edu Gita Sukthankar gita.sukthankar@ucf.edu

We present an agent development and data collection framework for Generals.io (GIO)--a real-time strategy game with imperfect information in which players attempt to gain control of opponents' starting positions within a 2D grid world. The framework provides event-based communication amongst several modules implemented as microservices, enabling real-time data collection from GIO's streaming data. Its modular design facilitates rapid bot development and testing, while the emphasis on data collection makes it easy to analyze agent performance. We use this framework in a case study of a top-performing GIO agent called Flobot. Our analysis demonstrates that Flobot's performance varies based on its starting position. Based on the analysis performed with our framework, we propose a modification to Flobot's pathfinding algorithm. Statistical tests show that the new algorithm results in a significant reduction in performance variance.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27537 2023-10-06T16:32:07-07:00 James Dameris jdameris@scu.edu Rosaura Hernandez Roman rhernandez2@scu.edu Max Kreminski maxkreminski@gmail.com

The Versu framework is historically notable for its full-featuredness as a suite of tools for creating highly responsive interactive dramas. However, it has also been lost for nearly a decade, and a similarly approachable and flexible simulationist interactive narrative authoring framework has not yet emerged to take its place. We therefore aim to introduce an open-source rational reconstruction of the Versu framework, drawing on publicly available documentation of Versu's design and implementation to assemble a successor system with similar architecture and capabilities. Here, we present the first component of this system: Praxish, a reconstruction of the low-level exclusion logic language atop which the rest of Versu's functionality is based.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27512 2023-10-06T16:31:05-07:00 Wael Al Enezi wael.alenezi@mail.mcgill.ca Clark Verbrugge clump@cs.mcgill.ca

The player's perception of AI behavior significantly influences their overall game experience. This perception is shaped by both interactive encounters and careful observations, particularly in genres like stealth, where gameplay revolves around planning strategies based on AI enemy movement. This paper aims to derive general insights into the player experience concerning two crucial gameplay elements that impact the perception of NPC intelligence. The first element pertains to the actual behavior of opponent NPCs, while the second focuses on the dialogues employed to highlight NPC decision-making. We conducted a user study to assess whether players can discern between complex and simple NPC behavior during gameplay in a specific scenario of a top-down stealth game prototype. We introduced variations in spoken dialogs to determine their effect on player perception. In the end, our findings revealed that when simple dialogs were used, players derived greater enjoyment from a more complex AI behavior. However, using contextual dialog allowed a simple behavior to match a complex behavior in player enjoyment.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27513 2023-10-06T16:31:06-07:00 Wael Al Enezi wael.alenezi@mail.mcgill.ca Clark Verbrugge clump@cs.mcgill.ca

In stealth games, guard patrol behavior constitutes one of the primary challenges players encounter. While most stealth games employ hard-coded guard behaviors, the same approach is not feasible for procedurally generated environments. Previous research has introduced various dynamic guard patrol behaviors; however, there needs to be more play-testing to quantitatively measure their impact on players. This research paper presents a user study to evaluate players' experiences in terms of enjoyment and difficulty when playing against several dynamic patrol behaviors in a stealth game prototype. The study aimed to determine whether players could differentiate between different guard behaviors and assess their impact on player experience. We found that players were generally capable of distinguishing between the various dynamic guard patrol behaviors in terms of difficulty and enjoyment when competing against them. The study sheds light on the nuances of player perception and experience with different guard behaviors, providing valuable insights for game developers seeking to create engaging and challenging stealth gameplay.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27514 2023-10-06T16:31:09-07:00 Rehaf AlJammaz raljamma@ucsc.edu Michael Mateas mmateas@ucsc.edu Noah Wardrip-Fruin nwardrip@ucsc.edu

An ability to morally reason is crucial to the believability of many fictional characters, from Jane Austen’s heroines to the denizens of The Good Place. These works often foreground the complexity of moral questions and the circumstances un- der which different forms of behavior might be justified. Morality is also foregrounded in many games, from Black and White to Mass Effect 3. Yet, most in-game characters judge other characters (or the player) based on a single reputation scale or binary values of right and wrong. There has been little exploration in games of the relationship between char- acter values and beliefs and moral reasoning. In keeping with this year’s conference theme, “Oh the Humanity,” this design postmortem paper describes the design and development of Argument Box, a model of moral argumentation and reason- ing based on Lakoff’s metaphor theory of moral politics. We describe our design approach, iterations, and authoring con- cerns — covering what went right and wrong in our attempts to model morality-based argumentation for believable game characters.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27515 2023-10-06T16:31:11-07:00 Nicholas Baron njbaron@cpp.edu Markus Eger meger@cpp.edu

Grammars are well-suited for the generation of structured content, such as text. Some specialized grammars, such as Shape Grammars, can even be used to generate 3D structures inside a game world like Minecraft. However, the top-down nature of grammars present limitations when it comes to modeling structures that should be connected to or utilize given geometry. In this paper, we describe an extension to an existing grammar model, called Box-Split Grammars, that extends it with the ability to observe existing geometry during the generation process, in order to incorporate it propertly into the generated structures. This modification also requires the addition of back-tracking in order to handle states in which certain geometry was (not) observed. We demonstrate the utility of this extension by showing how it can be used to place support structures for bridges and tunnels in a way that fits within an existing landscape.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27516 2023-10-06T16:31:14-07:00 Manuel Eberhardinger eberhardinger@hdm-stuttgart.de Johannes Maucher maucher@hdm-stuttgart.de Setareh Maghsudi setareh.maghsudi@uni-tuebingen.de

Understanding the interactions of agents trained with deep reinforcement learning is crucial for deploying agents in games or the real world. In the former, unreasonable actions confuse players. In the latter, that effect is even more significant, as unexpected behavior cause accidents with potentially grave and long-lasting consequences for the involved individuals. In this work, we propose using program synthesis to imitate reinforcement learning policies after seeing a trajectory of the action sequence. Programs have the advantage that they are inherently interpretable and verifiable for correctness. We adapt the state-of-the-art program synthesis system DreamCoder for learning concepts in grid-based environments, specifically, a navigation task and two miniature versions of Atari games, Space Invaders and Asterix. By inspecting the generated libraries, we can make inferences about the concepts the black-box agent has learned and better understand the agent's behavior. We achieve the same by visualizing the agent's decision-making process for the imitated sequences. We evaluate our approach with different types of program synthesizers based on a search-only method, a neural-guided search, and a language model fine-tuned on code.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27517 2023-10-06T16:31:16-07:00 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 agents that simulate player behaviour. By clustering player traces and using evolutionary algorithms to evolve parametric agents to best represent those clusters, our pipeline creates persona agents that represent the behavioural space of players. We here propose clustering playtraces based on behaviour, emotional experience and a mixture of both. We implement the pipeline on a test bed game and using 182 collected player traces with both behavioural and emotional information, we demonstrate that our persona agents can generate diverse player-like behaviour both in the level used to evolve them but also in a previously unseen level. We further find that using emotional information leads to better behavioural coverage on both levels. Although on its early stages, our approach offers a new perspective on how game developers and testers can gather insights on player behaviour without having to rely on extensive user testing.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27518 2023-10-06T16:31:19-07:00 Jasper Gerigk jasper.gerigk@mail.utoronto.ca Steve Engels sengels@cs.utoronto.ca

Deck-building games, like Dominion, present an unsolved challenge for game AI research. The complexity arising from card interactions and the relative strength of strategies depending on the game configuration result in computer agents being limited to simple strategies. This paper describes the first application of recent advances in Geometric Deep Learning to deck-building games. We utilize a comprehensive multiset-based game representation and train the policy using a Soft Actor-Critic algorithm adapted to support variable-size sets of actions. The proposed model is the first successful learning-based agent that makes all decisions without relying on heuristics and supports a broader set of game configurations. It exceeds the performance of all previous learning-based approaches and is only outperformed by search-based approaches in certain game configurations. In addition, the paper presents modifications that induce agents to exhibit novel human-like play strategies. Finally, we show that learning strong strategies based on card combinations requires a reinforcement learning algorithm capable of discovering and executing a precise strategy while ignoring simpler suboptimal policies with higher immediate rewards.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27519 2023-10-06T16:31:20-07:00 Robert C. Gray robert.c.gray@drexel.edu Jichen Zhu jichen.zhu@gmail.com Santiago Ontañón santi.ontanon@gmail.com

Experience management (EM) agents in multiplayer serious games face unique challenges and responsibilities regarding the fair treatment of players. One such challenge is the Greedy Bandit Problem that arises when using traditional Multi-Armed Bandits (MABs) as EM agents, which results in some players routinely prioritized while others may be ignored. We will show that this problem can be a cause of player non-adherence in a multiplayer serious game played by human users. To mitigate this effect, we propose a new bandit strategy, the Shapley Bandit, which enforces fairness constraints in its treatment of players based on the Shapley Value. We evaluate our approach via simulation with virtual players, finding that the Shapley Bandit can be effective in providing more uniform treatment of players while incurring only a slight cost in overall performance to a typical greedy approach. Our findings highlight the importance of fair treatment among players as a goal of multiplayer EM agents and discuss how addressing this issue may lead to more effective agent operation overall. The study contributes to the understanding of player modeling and EM in serious games and provides a promising approach for balancing fairness and engagement in multiplayer environments.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27520 2023-10-06T16:31:23-07:00 Emily Halina ehalina@ualberta.ca Matthew Guzdial guzdial@ualberta.ca

Procedural Content Generation (PCG) is the algorithmic generation of content, often applied to games. PCG and PCG via Machine Learning (PCGML) have appeared in published games. However, it can prove difficult to apply these approaches in the early stages of an in-development game. PCG requires expertise in representing designer notions of quality in rules or functions, and PCGML typically requires significant training data, which may not be available early in development. In this paper, we introduce Tree-based Reconstructive Partitioning (TRP), a novel PCGML approach aimed to address this problem. Our results, across two domains, demonstrate that TRP produces levels that are more playable and coherent, and that the approach is more generalizable with less training data. We consider TRP to be a promising new approach that can afford the introduction of PCGML into the early stages of game development without requiring human expertise or significant training data.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27521 2023-10-06T16:31:25-07:00 Branden Ingram branden.ingram@gmail.com Clint Van Alten clint.vanalten@wits.ac.za Richard Klein richard.klein@wits.ac.za Benjamin Rosman benjros@gmail.com

Developing diverse and realistic agents in terms of behaviour and skill is crucial for game developers to enhance player satisfaction and immersion. Traditional game design approaches involve hand-crafted solutions, while learning game-playing agents often focuses on optimizing for a single objective, or play-style. These processes typically lack intuitiveness, fail to resemble realistic behaviour, and do not encompass a diverse spectrum of play-styles at varying levels of skill. To this end, our goal is to learn a set of policies that exhibit diverse behaviours or styles while also demonstrating diversity in skill level. In this paper, we propose a novel pipeline, called PCPG (Play-style-Centric Policy Generation), which combines unsupervised play-style identification and policy learning techniques to generate a diverse set of play-style-centric agents. The agents generated by the pipeline can effectively capture the richness and diversity of gameplay experiences in multiple video game domains, showcasing identifiable and diverse play-styles at varying levels of proficiency.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27522 2023-10-06T16:31:28-07:00 Johor Jara Gonzalez jaragonz@ualberta.ca Seth Cooper se.cooper@northeastern.edu Matthew Guzdial guzdial@ualberta.ca

Automated game design (AGD), the study of automatically generating game rules, has a long history in technical games research. AGD approaches generally rely on approximations of human play, either objective functions or AI agents. Despite this, the majority of these approximators are static, meaning they do not reflect human player's ability to learn and improve in a game. In this paper, we investigate the application of Reinforcement Learning (RL) as an approximator for human play for rule generation. We recreate the classic AGD environment Mechanic Maker in Unity as a new, open-source rule generation framework. Our results demonstrate that RL produces distinct sets of rules from an A* agent baseline, which may be more usable by humans.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27523 2023-10-06T16:31:29-07:00 Johor Jara Gonzalez jaragonz@ualberta.ca Matthew Guzdial guzdial@ualberta.ca

Procedural Content Generation (PCG) and Procedural Content Generation via Machine Learning (PCGML) have been used in prior work for generating levels in various games. This paper introduces Content Augmentation and focuses on the subproblem of level inpainting, which involves reconstructing and extending video game levels. Drawing inspiration from image inpainting, we adapt two techniques from this domain to address our specific use case. We present two approaches for level inpainting: an Autoencoder and a U-net. Through a comprehensive case study, we demonstrate their superior performance compared to a baseline method and discuss their relative merits. Furthermore, we provide a practical demonstration of both approaches for the level inpainting task and offer insights into potential directions for future research.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27524 2023-10-06T16:31:33-07:00 Nic Junius njunius@ucsc.edu Elin Carstensdottir ecarsten@ucsc.edu

Designing AI models for expressive character behavior is a considerable challenge. Such models represent a massive possibility space of individual behaviors and sequences of different character expressions. Iterating on designs of such models is complex because the possibility spaces they afford are challenging to understand in their entirety and map intuitively onto a meaningful experience for a user. Automated playtesting has primarily been focused on the physical spaces of game levels and the ability of AI players to enact personas and complete tasks within those levels. However, core principles of automated playtesting can be applied to expressive models to expose information about their expressive possibility space. We propose a new approach to automated playtesting for AI character behaviors: Expressive Response Curves (ERC). ERC allows us to map specific actions taken by a player to perform a particular expression to understand the affordances of an expressive possibility space. We present a case study applying ERC to Puppitor rulesets. We show that using this method we can compile paths through Puppitor rulesets to map them and further understand the nature of the expressive spaces afforded by the system. We argue that by using ERC, it is possible to give designers more nuanced information and guidance to create better and more expressive AI characters.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27525 2023-10-06T16:31:34-07:00 Abdelrahman Madkour madkour.a@husky.neu.edu Chris Martens c.martens@northeastern.edu Steven Holtzen s.holtzen@northeastern.edu Casper Harteveld c.harteveld@northeastern.edu Stacy Marsella s.marsella@northeastern.edu

Research in procedural content generation (PCG) has recently heralded two major methodologies: machine learning (PCGML) and declarative programming. The former shows promise by automating the specification of quality criteria through latent patterns in data, while the latter offers significant advantages for authorial control. In this paper we propose the use of probabilistic logic as a unifying framework that combines the benefits of both methodologies. We propose a Bayesian formalization of content generators as probability distributions and show how common PCG tasks map naturally to operations on the distribution. Further, through a series of experiments with maze generation, we demonstrate how probabilistic logic semantics allows us to leverage the authorial control of declarative programming and the flexibility of learning from data.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27526 2023-10-06T16:31:37-07:00 Andrew Nash anash@mun.ca Andrew Vardy av@mun.ca Dave Churchill dave.churchill@gmail.com

We present a novel perception model named Herd's Eye View (HEV) that adopts a global perspective derived from multiple agents to boost the decision-making capabilities of reinforcement learning (RL) agents in multi-agent environments, specifically in the context of game AI. The HEV approach utilizes cooperative perception to empower RL agents with a global reasoning ability, enhancing their decision-making. We demonstrate the effectiveness of the HEV within simulated game environments and highlight its superior performance compared to traditional ego-centric perception models. This work contributes to cooperative perception and multi-agent reinforcement learning by offering a more realistic and efficient perspective for global coordination and decision-making within game environments. Moreover, our approach promotes broader AI applications beyond gaming by addressing constraints faced by AI in other fields such as robotics. The code is available at https://github.com/andrewnash/Herds-Eye-View

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27527 2023-10-06T16:31:39-07:00 Jasmine Otto ottojasmine@gmail.com Autumn Chen cchen.intfic@gmail.com Adam M. Smith amsmith@ucsc.edu

Quality-based narratives (QBN) are hypertexts with extensive implicit linked structure. The observation of one passage can have non-obvious long-range implications for the reachability of other passages, which poses an authoring challenge. To help narrative designers address this issue, we produced an interface which visually summarizes all possible playtraces. Our interface leverages answer set programming to produce a query language over possible playtraces, allowing narrative designers to drill down to interesting scenarios. We introduce this interface through the DendryScope tool, which accepts most QBNs written in the Dendry language. We evaluated DendryScope by interviewing four narrative designers as they used the tool to explore Bee, a notable QBN written by Emily Short.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27528 2023-10-06T16:31:42-07:00 Xiangyu Peng xpeng62@gatech.edu Christopher Cui ccui46@gatech.edu Wei Zhou wzhou322@gatech.edu Renee Jia rjia35@gatech.edu Mark Riedl riedl@cc.gatech.edu

Reward design for reinforcement learning agents can be difficult in situations where one not only wants the agent to achieve some effect in the world but where one also cares about how that effect is achieved. For example, we might wish for an agent to adhere to a tacit understanding of commonsense, align itself to a preference for how to behave for purposes of safety, or taking on a particular role in an interactive game. Storytelling is a mode for communicating tacit procedural knowledge. We introduce a technique, Story Shaping, in which a reinforcement learning agent infers tacit knowledge from an exemplar story of how to accomplish a task and intrinsically rewards itself for performing actions that make its current environment adhere to that of the inferred story world. Specifically, Story Shaping infers a knowledge graph representation of the world state from observations, and also infers a knowledge graph from the exemplar story. An intrinsic reward is generated based on the similarity between the agent's inferred world state graph and the inferred story world graph. We conducted experiments in text-based games requiring commonsense reasoning and shaping the behaviors of agents as virtual game characters.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27529 2023-10-06T16:31:44-07:00 Bogdan Pikula pikula@cs.toronto.edu Steve Engels sengels@cs.toronto.edu

Facial expressions are a crucial but challenging aspect of animating in-game characters. They provide vital nonverbal communication cues, but given the high complexity and variability of human faces, the task of capturing the natural diversity and affective complexity of human faces can be a labour-intensive process for animators. This motivates the need for more accurate, realistic and lightweight methods for generating emotional expressions for in-game characters. In this work, we introduce FlexComb, a Facial Landmark-based Expression Combination model, designed to generate a real-time space of realistic facial expression combinations. FlexComb leverages the highly varied CelebV-HQ dataset containing emotions in the wild, and a transformer-based architecture. The central component of the FlexComb system is an emotion recognition model that is trained on the facial dataset, and used to generate a larger dataset of tagged faces. The resulting system generates in-game facial expressions by sampling from this tagged dataset, including expressions that combine emotions in specified amounts. This allows in-game characters to take on variety of realistic facial expressions for a single emotion, which addresses this primary challenge of facial emotion modeling. FlexComb shows potential for expressive facial emotion simulation with applications that include animation, video game development, virtual reality, and human-computer interaction.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27530 2023-10-06T16:31:47-07:00 Debraj Ray debraj1@ualberta.ca Nathan R. Sturtevant nathanst@ualberta.ca

Real-time strategy games require players to respond to short-term challenges (micromanagement) and long-term objectives (macromanagement) simultaneously to win. However, many players excel at one of these skills but not both. This research is motivated by the question of whether the burden of micromanagement can be reduced on human players through delegation of responsibility to autonomous agents. In particular, this research proposes an adversarial navigation architecture that enables units to autonomously navigate through places densely populated with enemies by learning to micromanage itself. Our approach models the adversarial pathfinding problem as a Markov Decision Process (MDP) and trains an agent with reinforcement learning on this MDP. We observed that our approach resulted in the agent taking less damage from adversaries while travelling shorter paths, compared to previous approaches for adversarial navigation. Our approach is also efficient in memory use and computation time. Interestingly, the agent using the proposed approach also outperformed baseline approaches while navigating through environments that are significantly different from the training environments. Furthermore, when the game design is modified, the agent discovers effective alternate strategies considering the updated design without any changes in the learning framework. This property is particularly useful because in game development the design of a game is often updated iteratively.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27531 2023-10-06T16:31:49-07:00 Jungah Son jungah@ucsb.edu Marko Peljhan peljhan@ucsb.edu George Legrady glegrady@ucsb.edu Misha Sra sra@cs.ucsb.edu

This work describes a system to recolor a user’s painting based on the perceived emotional state of the viewer. An automatic palette selection algorithm is used to generate color palettes for a set of emotions. A user can create a painting using one of the generated palettes. To notify the end of the painting, the user clicks on the DONE button. Once the button is pressed, the colors of the user's painting change as the facial expression of the user changes. Facial emotion recognition is used in this process to classify the emotional status of the user’s face.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27532 2023-10-06T16:31:53-07:00 Shuwei Wang shuwei4@ualberta.ca Vadim Bulitko bulitko@ualberta.ca Taoan Huang taoanhua@usc.edu Sven Koenig skoenig@usc.edu Roni Stern roni.stern@gmail.com

Prioritized planning is a popular approach to multi-agent pathfinding. It prioritizes the agents and then repeatedly invokes a single-agent pathfinding algorithm for each agent such that it avoids the paths of higher-priority agents. Performance of prioritized planning depends critically on cleverly ordering the agents. Such an ordering is provided by a priority function. Recent work successfully used machine learning to automatically produce such a priority function given good orderings as the training data. In this paper we explore a different technique for synthesizing priority functions, namely program synthesis in the space of arithmetic formulae. We synthesize priority functions expressed as arithmetic formulae over a set of meaningful problem features via a genetic search in the space induced by a context-free grammar. Furthermore we regularize the fitness function by formula length to synthesize short, human-readable formulae. Such readability is an advantage over previous numeric machine-learning methods and may help explain the importance of features and how to combine them into a good priority function for a given domain. Moreover, our experimental results show that our formula-based priority functions outperform existing machine-learning methods on the standard benchmarks in terms of success rate, run time and solution quality without using more training data.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27533 2023-10-06T16:31:55-07:00 Chujin Zhou zcj2290151272@gmail.com Tiago Machado tiago.la.machado@gmail.com Casper Harteveld c.harteveld@northeastern.edu

We introduce a new reward function direction for intrinsically motivated reinforcement learning to mimic human behavior in the context of computer games. Similar to previous research, we focus on so-called ``curiosity agents'', which are agents whose intrinsic reward is based on the concept of curiosity. We designed our novel intrinsic reward, which we call ``Cautious Curiosity'' (CC) based on (1) a theory that proposes curiosity as a psychological definition called information gap, and (2) a recent study showing that the relationship between curiosity and information gap is an inverted U-curve. In this work, we compared our agent using the classic game Super Mario Bros. with (1) a random agent, (2) an agent based on the Asynchronous Advantage Actor Critic algorithm (A3C), (3) an agent based on the Intrinsic Curiosity Module (ICM), and (4) an average human player. We also asked participants (n = 100) to watch videos of these agents and rate how human-like they are. The main contribution of this work is that we present a reward function that, as perceived by humans, induces an agent to play a computer game similarly to a human, while maintaining its competitiveness and being more believable compared to other agents.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27534 2023-10-06T16:31:58-07:00 Andrew Zhu andrz@seas.upenn.edu Lara Martin laramar@umbc.edu Andrew Head head@seas.upenn.edu Chris Callison-Burch ccb@seas.upenn.edu

The role of a Dungeon Master, or DM, in the game Dungeons & Dragons is to perform multiple tasks simultaneously. The DM must digest information about the game setting and monsters, synthesize scenes to present to other players, and respond to the players' interactions with the scene. Doing all of these tasks while maintaining consistency within the narrative and story world is no small feat of human cognition, making the task tiring and unapproachable to new players. Large language models (LLMs) like GPT-3 and ChatGPT have shown remarkable abilities to generate coherent natural language text. In this paper, we conduct a formative evaluation with DMs to establish the use cases of LLMs in D&D and tabletop gaming generally. We introduce CALYPSO, a system of LLM-powered interfaces that support DMs with information and inspiration specific to their own scenario. CALYPSO distills game context into bite-sized prose and helps brainstorm ideas without distracting the DM from the game. When given access to CALYPSO, DMs reported that it generated high-fidelity text suitable for direct presentation to players, and low-fidelity ideas that the DM could develop further while maintaining their creative agency. We see CALYPSO as exemplifying a paradigm of AI-augmented tools that provide synchronous creative assistance within established game worlds, and tabletop gaming more broadly.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27542 2023-10-06T16:32:20-07:00 Markus Eger titocru+markus@gmail.com Rogelio Enrique Cardona-Rivera titocru+rogelio@gmail.com

The 19th AAAI conference on Artificial Intelligence and Interactive Digital Entertainment was held at the University of Utah, in Salt Lake City, Utah, USA.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27540 2023-10-06T16:32:16-07:00 Colan F. Biemer biemer.c@husky.neu.edu

Procedural level generation can create unseen levels and improve the replayability of games, but there are requirements for a generated level. First, a level must be completable. Second, a level must look and feel like a level that would exist in the game, meaning a random combination of tiles that happens to be completable is not enough. On top of these two requirements, though, is the player experience. If a level is too hard, the player will be frustrated. If too easy, they will be bored. Neither outcome is desirable. A procedural level generation system has to account for the player's skill and generate levels at the correct difficulty. I address this issue by showing how a Markov Decision Process can be used as a director to assemble levels tailored to a player's skill level, but I've only demonstrated that my approach works with surrogate agents. For my thesis, I plan to build on my past work by creating a full roguelike and platformer and running two player studies to validate my approach.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27541 2023-10-06T16:32:17-07:00 Qiao Zhang qzhang490@gatech.edu

With the goal of better understanding Human-machine Teaming (HMT) dynamics and how team competencies that are transportable across contexts can lead to different teaming behaviors and team performances, I propose a series of three studies to explore communication, coordination and adaptation in HMT paradigms. I implement and integrate multiple AI agents and use collaborative games as testing environments to evaluate teaming effects. My work can provide findings to two higher level research questions that are widely studied in HMT: 1) the bidirectional behaviors that human and AI agents may develop when working as a team and, 2) how different types of AI agents can impact the teaming efficiency in human-AI teaming. Besides, my work can also contribute to Human-Computer Interaction and Game AI scholarship with insights into teaming dynamics in Human-AI teaming.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27538 2023-10-06T16:32:10-07:00 Manuel Flurin Hendry manuel.hendry@zhdk.ch Norbert Kottmann norbert.kottmann@zhdk.ch Martin Fröhlich martin.froehlich@zhdk.ch Florian Bruggisser florian.bruggisser@zhdk.ch Marco Quandt marco.quandt@zhdk.ch Stella Speziali stella.speziali@zhdk.ch Valentin Huber valentin.huber@zhdk.ch Chris Salter christopher.salter@zhdk.ch

We present Stanley, a digital sculpture designed to engage audiences with the spontaneous and captivating emotional expressions of an artificial human. A 3D-printed face is brought to life through video projection mapping and a set of machine learning libraries and APIs, enabling real-time, embodied interactions with our virtual character. Stanley’s personality is shaped by traditional acting methods applied to a large language model. By creating human-machine encounters in emotionally salient scenarios, we explore how insights from acting and directing for the stage and the screen can enhance the development of compelling virtual agents. By interacting with Stanley, the audience experiences an entertaining yet unsettling encounter with AI technology, fostering a deeper understanding of machine learning techniques and enabling their critical reflection.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence https://ojs.aaai.org/index.php/AIIDE/article/view/27539 2023-10-06T16:32:12-07:00 Yuqian Sun 10001541@network.rca.ac.uk Zhouyi Li 190210418@stu.hit.edu.cn Ke Fang fang.ke@sz.tsinghua.edu.cn Chang Hee Lee changhee.lee@kaist.ac.kr Ali Asadipour ali.asadipour@rca.ac.uk

Generative AI (GenAI), encompassing image generation and large language models (LLMs), has opened new avenues for gameplay experiences. This paper introduces "1001 Nights", a narrative game centered on GenAI. Drawing inspiration from Wittgenstein's note, "The limits of my language mean the limits of my world", the game exemplifies the concept of language as reality. The protagonist, Shahrzad, possesses a unique power: specific keywords, such as "sword" or "shield", when spoken by others in tales, materialize as tangible weapons, serving as battle equipment against the King. Players guide the LLM-driven King in co-creating narratives, with GPT-4 employing LLM reasoning methods to ensure story consistency. As these narratives progress, the depicted world is dynamically generated and visualized through Stable Diffusion, blurring the boundaries between narrative and in-game reality. This fusion of interactive storytelling combines gameplay paradigms and story together with dynamic content generation. Players not only aim to alter Shahrzad's fate from the original folklore, but also leverage the power of natural language to shape the game's world. With this example, we propose the term "AI-Native games" to categorize innovative games where GenAI is fundamental to the game's novel mechanics and very existence.

2023-10-06T00:00:00-07:00 Copyright (c) 2023 Association for the Advancement of Artificial Intelligence