Modelling Protagonist Goals and Desires in First-Person Narrative

TL;DR

This paper introduces DesireDB, a dataset for modeling protagonist goals and desires in narratives, and evaluates methods for tracking desire fulfillment, with the LSTM Skip-Thought model achieving an F-measure of 0.7.

Contribution

The paper presents DesireDB, a new dataset with annotations for desires and their fulfillment, and evaluates computational models for desire tracking in narratives.

Findings

LSTM Skip-Thought model achieves 0.7 F-measure on desire fulfillment.

DesireDB provides gold-standard labels for desire statements and evidence.

Different methods for desire tracking are systematically evaluated.

Abstract

Many genres of natural language text are narratively structured, a testament to our predilection for organizing our experiences as narratives. There is broad consensus that understanding a narrative requires identifying and tracking the goals and desires of the characters and their narrative outcomes. However, to date, there has been limited work on computational models for this problem. We introduce a new dataset, DesireDB, which includes gold-standard labels for identifying statements of desire, textual evidence for desire fulfillment, and annotations for whether the stated desire is fulfilled given the evidence in the narrative context. We report experiments on tracking desire fulfillment using different methods, and show that LSTM Skip-Thought model achieves F-measure of 0.7 on our corpus.