Shaping the Narrative Arc: An Information-Theoretic Approach to Collaborative Dialogue

TL;DR

The work tackles the challenge of generating engaging collaborative narratives by introducing an information-theoretic narrative-arc framework that jointly models a base conversation and a discrete set of universes. A universe model, updated via recursive belief propagation, tracks how each utterance shifts belief over universes, forming a narrative arc whose shape is regulated by entropy-based metrics and an alpha-controlled revealing-concealing balance. Empirical results show that tuning alpha to favor revealing improves next-line prediction accuracy and yields more engaging interactions in expert-improviser studies, while concealing tends to dilute specificity. The framework is designed to be agnostic to the particular conversation or universe model and is applicable across genres, including theatre, film, and scripted dialogue, with extensive supplementary material detailing implementation.

Abstract

We consider the problem of designing an artificial agent capable of interacting with humans in collaborative dialogue to produce creative, engaging narratives. In this task, the goal is to establish universe details, and to collaborate on an interesting story in that universe, through a series of natural dialogue exchanges. Our model can augment any probabilistic conversational agent by allowing it to reason about universe information established and what potential next utterances might reveal. Ideally, with each utterance, agents would reveal just enough information to add specificity and reduce ambiguity without limiting the conversation. We empirically show that our model allows control over the rate at which the agent reveals information and that doing so significantly improves accuracy in predicting the next line of dialogues from movies. We close with a case-study with four professional theatre performers, who preferred interactions with our model-augmented agent over an unaugmented agent.

Figures (9)

• Figure 1: The narrative arcs of a synthetic dialogue (a), using the Newsgroups universe model (b) and Movies universe model (c). This dialogue is likely SCIENCE or TALK under the Newsgroups model, and DRAMA or COMEDY under the Movie genres model.
• Figure 2: First 20 lines of Romeo and Juliet modeled with Newsgroups (top), Movies (middle), and DeepMoji (bottom) universe models.
• Figure 3: Narrative arcs over 10 utterances at increasing $\alpha$ values: concealing (top), neutral (mid), revealing (bottom). On the right are utterances generated by each model after priming (bold). Dotted red line indicates the start of narrative arc shaping.
• Figure 4: Revealing and Concealing across Universe Models. Dialogue generated to be (a) revealing ($\alpha = 20$) under the objective model Newsgroups is revealing under the testing Movies universe. The same is true for (b) concealing ($\alpha = -25$) dialogue. Data shown are means and standard deviation (shaded) over 20 runs of random conversation model.
• Figure 5: Information revelation region as $\alpha$ varies for (left) top-$3$ accuracy and (right) MRR in universe model modulated prediction task

Theorems & Definitions (2)

• Proposition 1
• proof