This paper presents an approach to automatically extracting and representing narrative information from stories written in natural language. Specifically, we present our results in extracting story graphs, a formalism that captures the entities (e.g., characters, props, locations) and their interactions in a story. The long-term goal of this research is to automatically extract this narrative information in order to use it in computational narrative systems such as story generators or interactive fiction systems. Our approach combines narrative domain knowledge and off-the-shelf natural language processing (NLP) tools into a machine learning framework to build story graphs by automatically identifying entities, actions, and narrative roles. We report the performance of our fully automated system in a corpus of 21 stories and provide examples of the extracted story graphs and their uses in computational narrative systems.
My research aims to contribute to research in the narrative authoring domain by using cognitive models in narrative plan generation. These cognitive models determine how actions and events in narrative affect the audience. My research intends to leverage these models in narrative planning and use them to provide intelligent narrative plans that are structured to invoke specific responses from audiences when they experience the narrative. This sort of approach would greatly benefit the enrich growing set of variables of narrative planning. My research is in the nascent field of the computational modeling of narrative, work that seeks to enable computerassisted authoring of stories by modeling the cognitive processes of both author and audience. I intend to extend work on narrative generation that uses planning algorithms to create stories that are consistent and complete (Young 2007). Previous work in narrative planning has been effective at borrowing policy planning and state-space search algorithms from AI in order to generate plot (Riedl and Young 2014). However, the majority of this work focuses on structural properties of a story (e.g., causal consistency (Li et al. 2012), intentionality (Riedl and Young 2010), conflict between characters (Ware et al. 2014)) but does not address the impact that the story has on the cognitive and affective response of its audience (e.g., tension, suspense). The goal of my work is to leverage models of author and audience to address these types of limitations.
Computational Narrative has provided several examples of how to process narrations using semantical approaches. While many useful concepts for computational management of stories have been unveiled, a common barrier has hindered their development: semantic knowledge is still too complex to handle. In this paper, a focus shift based on narrative structure is proposed. Instead of digging deeper into the possibilities of semantic processing, analysing structural properties of stories and keeping the semantic load to a minimum can allow for a more efficient use of available narrative corpora, even without mimicking human behaviour.
Several existing systems have used reader models for story generation, but they have focused on either interactive contexts or pure discourse-level manipulation. I intend to buid a reader-model story generator that not only applies reader modelling to full plot generation, but which also draws on theories about intentionality and emotions put forward by Lisa Zunshine, Keith Oatley, and Raymond Mar. To evaluate the contributions of the reader model, I'll compare it with human-authored stories using measures of reader engagement. I'll also run the model on human-authored stories and compare the results to a human gold-standard analysis.
Kazakova, Vera A. (University of Central Florida) | Hastings, Lauren (University of Central Florida) | Posadas, Andres (University of Central Florida) | Gonzalez, Lucas C. (University of Central Florida) | Knauf, Rainer ( Technische Universität Ilmenau ) | Jantke, Klaus P. (ADICOM Software GmbH) | Gonzalez, Avelino J. (University of Central Florida)
In this work we present fAIble: a novel graph-based modular storytelling framework. fAIble is centered around a graph database, incorporates invariable elements of folktale structure, while accounting for thoughts and actions. Action outcomes are a product of probabilistic story generation. Probabilities are based on elements of common sense, invariable elements of folktale structure, high-level character roles, and a wide variety of other variables (e.g. characters' physical and psychological traits, context-based likelihood of encountering specific items and characters, etc.). A prototype implementation is tested through an anonymous questionnaire. Results demonstrate the ability of graph-based cognition to produce coherent story prototypes with sensible character actions, while maintaining output variability.