Modeling Human Mental States with an Entity-based Narrative Graph
arXiv:2104.07079 · doi:10.18653/v1/2021.naacl-main.391
Abstract
Understanding narrative text requires capturing characters' motivations, goals, and mental states. This paper proposes an Entity-based Narrative Graph (ENG) to model the internal-states of characters in a story. We explicitly model entities, their interactions and the context in which they appear, and learn rich representations for them. We experiment with different task-adaptive pre-training objectives, in-domain training, and symbolic inference to capture dependencies between different decisions in the output space. We evaluate our model on two narrative understanding tasks: predicting character mental states, and desire fulfillment, and conduct a qualitative analysis.
arXiv commentAccepted at NAACL 2021
References in corpus (7)
- Adam: A Method for Stochastic Optimization
- Semi-Supervised Classification with Graph Convolutional Networks
- Convolutional Neural Networks for Sentence Classification
- Tracking the World State with Recurrent Entity Networks
- A Corpus and Evaluation Framework for Deeper Understanding of Commonsense Stories
- COMET: Commonsense Transformers for Automatic Knowledge Graph Construction
- Ask, and shall you receive?: Understanding Desire Fulfillment in Natural Language Text