Archetype-Driven Character Dialogue Generation for Interactive Narrative

Abstract

Abstract. Recent years have seen a growing interest in creating virtual agents to populate the cast of characters for interactive narrative. A key challenge posed by interactive characters for narrative environments is devising expressive dialogue generators. To be effective, character dialogue generators must be able to simultaneously take into account multiple sources of information that bear on dialogue, including character attributes, plot development, and communicative goals. Building on the narrative theory of character archetypes, we propose an archetype-driven character dialogue generator that uses a probabilistic unification framework to generate dialogue motivated by character personality and narrative history to achieve communicative goals. The generator's behavior is illustrated with character dialogue generation in a narrative-centered learning environment, CRYSTAL ISLAND.

Generating Sentence Planning Variations for Story Telling

“…In future work, we aim to explore interactions between a number of our novel narratological parameters. We expect to do this both with a rule-based approach, as well as by building on recent work on statistical models for expressive generation (Rieser and Lemon, 2011;Paiva and Evans, 2004;Langkilde, 1998;Rowe et al, 2008;Mairesse and Walker, 2011). This should allow us to train a narrative generator to achieve particular narrative effects, such as engagement or empathy with particular characters.…”

“…But oh, if you could have seen the look on his startled face and how he jumped back each time he caught his reflection in the bowl! Table 1: The Startled Squirrel Weblog Story limitation greatly restricts the range of applications; it also means that it is impossible to take advantage of recent work in expressive and statistical language generation that can dynamically and automatically produce a large number of variations of given content (Rieser and Lemon, 2011;Paiva and Evans, 2004;Langkilde, 1998;Rowe et al, 2008;Mairesse and Walker, 2011). Such variations are important for expressive purposes, we well as for user adaptation and personalization (Zukerman and Litman, 2001;Wang et al, 2005;McQuiggan et al, 2008).…”

Generating Different Story Tellings from Semantic Representations of Narrative

“…Prior research on nlg for story generation has primarily focused on using planning mechanisms in order to automatically generate story event structure, with limited work on the problems involved with automatically mapping the semantic representations of a story and its event and dialogue structure to the syntactic structures that allow the story to be told in natural language [3,4,5]. Recent research focuses on generating story dialogue on a turn by turn basis and scaling up text planners to produce larger text prose [6,7,8,9], but has not addressed the problem of bridging between the semantic representation of story structure and the nlg engine [5,7,6,10,11]. An example of this work is the storybook system [3] which explicitly focused on the ability to generate many versions of a single story, much in a spirit of our own work.…”

Towards Generating Stylistic Dialogues for Narratives Using Data-Driven Approaches

“…Stories can be represented through plot structures [20,21,24], and are often equated to a sequence of narrative actions [22], normally represented as objects, subjects, actions and other features. Narrative actions can be presented through discourse in different ways: visual [14], audio, and linguistic [10].…”