This paper investigates the performance of Large Language Models (LLMs) as autonomous agents in the Hearthstone digital collectible card game. Although traditional numerical agents have demonstrated strong results in competitive game environments, the reasoning capabilities of LLM-based agents remain largely unexplored in this context. To address this gap, we developed an LLM-driven Hearthstone agent using the Sabberstone framework to evaluate several models, including GPT-4o, GPT-4o-mini, o3-mini, and GPT-5-mini, across multiple decks and prompting strategies. Our experiments compare their win rates against established numerical agents and analyze the impact of different prompting techniques, such as Chain-of-Thought (CoT), Reverse Chain-of-Thought (RCoT), ReAct, and Directed Acyclic Graph (DAG) prompting.

Artificial Intelligence (AI) characters in games, particularly non-player characters (NPC), offer exciting new possibilities for enhancing immersion. While these characters can respond to players' questions, another challenge remains: Are they capable of engaging in complex reasoning while also retaining and effectively utilizing information from past interactions? Using the memory stream of conversational agents, our aim is to explore the reasoning capabilities of various LLM, from large server-based models like ChatGPT-3.5 to small and local models like Llama 2 7b and Mistral 7B. Our objective is to evaluate AI performance in gamified environments, starting with simple reasoning tasks suited to a game context within social settings, and progressively advancing to a role-playing game that simulates a job-hiring process. The results suggest that both larger server-based models and smaller models exhibit reasoning capabilities in socially gamified scenarios. However, especially for small local models, more research is needed to enhance the quality of their performance.

Effective social communication relies on understanding emotions, interpreting social cues, and maintaining meaningful conversations. To support the development of these skills, we propose a serious game featuring Conversational Artificial Intelligence agents that enable users to engage in questdriven conversations across various scenarios. Users can design custom scenarios and set specific conversation goals as quests to complete one-on-one conversations with a virtual agent who can express basic emotions. To provide structured guidance, the system offers predefined scenarios such as a date simulation, job interview, and Who Am I? game. In addition, it features Theory of Mind training to help people improve their ability to detect sarcasm, irony, and navigate everyday interactions more effectively. We design and evaluate our system to ensure a highly effective and engaging experience. Built on GPT-4o, our approach was evaluated using the Bot Usability Scale, with 20 users successfully recruited online. The results show that the agent's functionality was well received, showcasing high usability and effective communication. However, there is room for improvement in contextual relevance. Although LLM maintained coherent conversations, we faced challenges with long-term context retention. Furthermore, although the avatar's emotional expressions were generally accurate, further refinement is needed to better align them with the context.

In this study, we explore the reasoning capabilities of Large Language Models (LLMs) within the context of the social communication game Werewolf, aiming to evaluate their performance in managing complex system states commonly found in computer games.Our agent architecture gathers data, refines them into detailed information, and plans actions based on this knowledge.To demonstrate the feasibility of using LLM based agents in computer games, we developed a simulation and evaluation tool using the Unity game engine.This software enables users to experiment with various LLMs and agent architectures and to measure model performance within the application.For evaluation, we tested three models: GPT-3.5 Turbo, Mistral-7B-OpenOrca, and Nous-Hermes-Llama2-13B.The results show that even smaller models can perform reasonably well in Werewolf.However, their error rate is significantly higher, highlighting the need for additional software modules or fine-tuning to improve their accuracy.

Autism Spectrum Disorder (ASD) can profoundly affect reciprocal social communication, resulting in substantial and challenging impairments. One aspect is that for people with ASD conversations in everyday life are challenging due to difficulties in understanding social cues, interpreting emotions, and maintaining social verbal exchanges. To address these challenges and enhance social skills, we propose the development of a learning game centered around social interaction and conversation, featuring Artificial Intelligence agents. Our initial step involves seven expert interviews to gain insight into the requirements for empathetic and conversational agents in the field of improving social skills for people with ASD in a gamified environment. We have identified two distinct use cases: (1) Conversation partners to discuss real-life issues and (2) Training partners to experience various scenarios to improve social skills. In the latter case, users will receive quests for interacting with the agent. Additionally, the agent can assign quests to the user, prompting specific conversations in real life and providing rewards for successful completion of quests.

In this paper, we present a systematic review of the applications of (1) Extended Reality (XR), (2) Augmented Reality (AR), and (3) Virtual Reality (VR) technologies to enhance emotion recognition and emotion expression in people with Autism Spectrum Disorder (ASD). ASD can affect various abilities, and poses challenges to the recognition of emotions in others, which is often referred to as “social blindness”. Treating this condition typically requires intensive one-on-one or small-group therapy sessions, which can be costly and limited in terms of availability. With the growing number of diagnoses of ASD, concerns have risen regarding a potential “lost generation” that may face difficulties in fulfilling its potential. Through this comprehensive review, we aim to provide an overview of innovative approaches that use XR technologies to improve the learning experience of individuals with ASD.

We propose an outdoor localization system using a particle filter. In our approach, a textured, geo-registered model of the outdoor environment is used as a reference to estimate the pose of a smartphone. The device position and the orientation obtained from a Global Positioning System (GPS) receiver and an inertial measurement unit (IMU) are used as a first estimation of the true pose. Then, multiple pose hypotheses are randomly distributed about the GPS/IMU measurement and use to produce renderings of the virtual model. With vision-based methods, the rendered images are compared with the image received from the smartphone, and the matching scores are used to update the particle filter. The outcome of our system improves the camera pose estimate in real time without user assistance.

Augmented reality (AR) allows for an on-site presentation of information that is registered to the physical environment. Applications from civil engineering, which require users to process complex information, are among those which can benefit particularly highly from such a presentation. In this paper, we will describe how to use AR to support monitoring and documentation of construction site progress. For these tasks, the responsible staff usually requires fast and comprehensible access to progress information to enable comparison to the as-built status as well as to as-planned data. Instead of tediously searching and mapping related information to the actual construction site environment, our AR system allows for the access of information right where it is needed. This is achieved by superimposing progress as well as as-planned information onto the user's view of the physical environment. For this purpose, we present an approach that uses aerial 3-D reconstruction to automatically capture progress information and a mobile AR client for on-site visualization. Within this paper, we will describe in greater detail how to capture 3-D, how to register the AR system within the physical outdoor environment, how to visualize progress information in a comprehensible way in an AR overlay, and how to interact with this kind of information. By implementing such an AR system, we are able to provide an overview about the possibilities and future applications of AR in the construction industry.