This paper introduces CharMoral, a dataset designed to analyze the moral evolution of characters in long-form narratives. CharMoral, built from 1,337 movie synopses, includes annotations for character actions, context, and morality labels. To automatically construct CharMoral, we propose a four-stage framework, utilizing Large Language Models, to automatically classify actions as moral or immoral based on context. Human evaluations and various experiments confirm the framework’s effectiveness in moral reasoning tasks in multiple genres. Our code and the CharMoral dataset are publicly available at https://github.com/BaeSuyoung/CharMoral.

Understanding how urban parks are utilized and perceived by the public is crucial for effective urban planning and management. This study introduces a novel dataset derived from Instagram, using 42,187 images tagged with #Seoul and #Park hashtags from 2017 to 2023. These images were filtered using InternLM-XComposer2, a Multimodal Large Language Model (MLLM), to confirm they depicted park scenes. GPT-4 then annotated the filtered images, resulting in 29,866 valid image annotations of physical elements, human activities, animals, and emotions. The dataset is publicly available at https://huggingface.co/datasets/RedBall/seoul-urban-park-analysis-by-llm.

Emotion significantly influences human behavior and decision-making processes.We propose a labeling methodology grounded in Plutchik's Wheel of Emotions theory for emotion classification.Furthermore, we employ a Mixture of Experts (MoE) architecture to evaluate the efficacy of this labeling approach, by identifying the specific emotions that each expert learns to classify.Experimental results reveal that our methodology improves the performance of emotion classification.

While controllable text generation has received attention due to the recent advances in large-scale pre-trained language models, there is a lack of research that focuses on story-specific controllability. To address this, we present Story Control via Supervised Contrastive learning model (SCSC), to create a story conditioned on genre. For this, we design a supervised contrastive objective combined with log-likelihood objective, to capture the intrinsic differences among the stories in different genres. The results of our automated evaluation and user study demonstrate that the proposed method is effective in genre-controlled story generation.

Narrative is an essential factor that makes games more enjoyable. However, predicting the story quality has been challenging for decades. In this paper, we propose to use contextual word embedding models such as BERT and ELMo, for predicting a story's success in terms of quality and popularity by using the story text only. Since deep learning models generally require extensive data, we conducted experiments to test the efficacy of our proposed model by leveraging the movie plot summaries. We present the results of the evaluations and conclude with discussions.

Conflict is essential for the development of complex plots or more in-depth character design. Therefore, many literary works have employed a conflict-centered narrative structure. However, it is challenging to incorporate conflicts in a planning-based story generation because a planning algorithm constructs a plan that avoids conflicts, which can occur when one action negates the precondition of the other character's action. To address this issue, this paper presents an intelligent storytelling system that can automatically generate conflicts and their resolutions using the interleaved causal relationships between different characters in the story. We describe an example to show that the proposed algorithm can successfully generate a story that contains inter-personal conflicts and is presented in the Virtual Environment. We conclude with a discussion and future study.

Automated story generation has received much attention in the last few decades, as storytelling enhances the user experience in the game, education, and training domains. In this paper, we demonstrate an intelligent story generation system which plans a story using an AI planner. It then displays the story as a 3D animation using the Unity Game engine. The central idea underlying our system lies in the use of a database as a communication channel among sub-modules. The demonstration showcases the complete pipeline process of story generation and visualization where the story changes as the user interacts with the system.

In this study, we propose a system design for interactive narrative generation which uses database for information exchange among system sub-modules. We present database schema design and a simple example to show how our approach works. We conclude with our future work.

While games are a popular social media for children, there is a real risk that these children are exposed to potential sexual assault. A number of studies have already addressed this issue, however, the data used in previous research did not properly represent the real chats found in multiplayer online games. To address this issue, we obtained real chat data from MovieStarPlanet, a massively multiplayer online game for children. The research described in this paper aimed to detect predatory behaviors in the chats using machine learning methods. In order to achieve a high accuracy on this task, extensive preprocessing was necessary. We describe three different strategies for data selection and preprocessing, and extensively compare the performance of different learning algorithms on the different data sets and features.

Interactive storytelling has been receiving a growing attention from AI and game communities and a number of computational approaches have shown promises in generating stories for games. However, there has been little research on stories evoking specific cognitive and affective responses. The goal of the work we describe here is to develop a system that produces a narrative designed specifically to arouse suspense from its reader. Our approach attempts to create stories that manipulate the reader's suspense level by elaborating on the story structure that can influence the reader's narrative comprehension at a specific point in her reading. Adapting theories developed by cognitive psychologists, our approach uses a plan-based model of narrative comprehension to determine the final content of the story in order to manipulate the reader's suspense. In this paper, we describe our system implementation and empirical evaluations to test the efficacy of this system.

We present our research towards the design of a computational framework capable of modelling the formation and evolution of global patterns (i.e. group structures) in a population of social individuals. The framework is intended to be used in collaborative environments, e.g. social serious games and computer simulations of artificial societies. The theoretical basis of our research, together with current state of the art and future work, are briefly introduced.

We present an ontology-based reasoning approach for saving energy in a smart home setting where a mobile phone can serve as a generic sensor which can collect the inhabitant's contextual data. The paper details an ontology that describes the smart home domain and a prototype to test the system. Finally, we conclude with lessons learned from our work in developing an energy-aware smart home prototype and suggestions for future work.

Focalization, perspective-taking in stories, restricts narrative information to the eye of a character, allowing the audience limited perception for a particular effect. This paper briefly describes the work in narrative theories on focalization and our computational model of focalization in narrative generation using an Artificial Intelligence planning-based approach. Keys to the story generation with focalization are the use of distinctive knowledge (i.e., distinctive plan libraries) and the concept of focalizing factors for each focal character in the story.

Conflict is an unavoidable feature of life, but the development of conflict resolution management skills can facilitate the parties involved in resolving their conflicts in a positive manner. The goal of our research is to develop a serious game in which children may experiment with conflict resolution strategies and learn how to work towards positive conflict outcomes. While serious games related to conflict exist at present, our work represents the first attempt to teach conflict resolution skills through a game in a manner informed by sociological and psychological theories of conflict and current best practice for conflict resolution. In this paper, we present a computational approach to conflict generation and resolution. We describe the five phases involved in our conflict modeling process: conflict situation creation, conflict detection, player modeling and conflict strategy prediction, conflict management, and conflict resolution, and discuss the three major elements of our player model: assertiveness, cooperativeness, and relationship. Finally, we overview a simple resource management game we have developed in which we have begun experimenting with our conflict model concepts.