Sketch and Customize: A Counterfactual Story Generator

TL;DR

This paper introduces a sketch-and-customize model for counterfactual story rewriting, improving causal reasoning in text generation by explicitly modeling the relationship between conditions and story endings.

Contribution

The paper proposes a novel two-stage generation approach that enhances causal reasoning in story rewriting by explicitly separating skeleton extraction and customization.

Findings

Outperforms traditional sequence-to-sequence models in generating relevant and causally consistent story endings.

Effectively captures causal relations between conditions and story endings.

Produces more accurate counterfactual story rewrites.

Abstract

Recent text generation models are easy to generate relevant and fluent text for the given text, while lack of causal reasoning ability when we change some parts of the given text. Counterfactual story rewriting is a recently proposed task to test the causal reasoning ability for text generation models, which requires a model to predict the corresponding story ending when the condition is modified to a counterfactual one. Previous works have shown that the traditional sequence-to-sequence model cannot well handle this problem, as it often captures some spurious correlations between the original and counterfactual endings, instead of the causal relations between conditions and endings. To address this issue, we propose a sketch-and-customize generation model guided by the causality implicated in the conditions and endings. In the sketch stage, a skeleton is extracted by removing words which are conflict to the counterfactual condition, from the original ending. In the customize stage, a generation model is used to fill proper words in the skeleton under the guidance of the counterfactual condition. In this way, the obtained counterfactual ending is both relevant to the original ending and consistent with the counterfactual condition. Experimental results show that the proposed model generates much better endings, as compared with the traditional sequence-to-sequence model.