Modeling Protagonist Emotions for Emotion-Aware Storytelling
Preprint
Abstract
Emotions and their evolution play a central role in creating a captivating story. In this paper, we present the first study on modeling the emotional trajectory of the protagonist in neural storytelling. We design methods that generate stories that adhere to given story titles and desired emotion arcs for the protagonist. Our models include Emotion Supervision (EmoSup) and two Emotion-Reinforced (EmoRL) models. The EmoRL models use special rewards designed to regularize the story generation process through reinforcement learning. Our automatic and manual evaluations demonstrate that these models are significantly better at generating stories that follow the desired emotion arcs compared to baseline methods, without sacrificing story quality.Keywords
All Related Versions
This publication has 50 references indexed in Scilit:
- Self-Critical Sequence Training for Image Captioning2017
- On the Use of Character Affinities for Story Plot Generation2016
- Dynamic Facial Expressions of Emotion Transmit an Evolving Hierarchy of Signals over TimeCurrent Biology, 2014
- Employing emotions to drive plot generation in a computer-based storytellerCognitive Systems Research, 2007
- MEXICA: A computer model of a cognitive account of creative writingJournal of Experimental & Theoretical Artificial Intelligence, 2001
- Making sense of emotion in stories and social lifeCognition and Emotion, 1993
- Simple statistical gradient-following algorithms for connectionist reinforcement learningMachine Learning, 1992
- An argument for basic emotionsCognition and Emotion, 1992
- Do readers mentally represent characters' emotional states?Cognition and Emotion, 1992
- Prominent characters and events organize narrative understandingJournal of Memory and Language, 1985