A Corpus for Understanding and Generating Moral Stories

TL;DR

This paper introduces STORAL, a new dataset of Chinese and English moral stories, and proposes tasks and algorithms to improve machine understanding and generation of moral narratives by bridging story plots and implied morals.

Contribution

It presents novel understanding and generation tasks for moral stories, along with a retrieval-augmented algorithm to enhance model performance on these tasks.

Findings

Models struggle with abstract moral concepts.

Retrieval-augmented methods improve task performance.

STORAL dataset enables evaluation of moral story understanding.

Abstract

Teaching morals is one of the most important purposes of storytelling. An essential ability for understanding and writing moral stories is bridging story plots and implied morals. Its challenges mainly lie in: (1) grasping knowledge about abstract concepts in morals, (2) capturing inter-event discourse relations in stories, and (3) aligning value preferences of stories and morals concerning good or bad behavior. In this paper, we propose two understanding tasks and two generation tasks to assess these abilities of machines. We present STORAL, a new dataset of Chinese and English human-written moral stories. We show the difficulty of the proposed tasks by testing various models with automatic and manual evaluation on STORAL. Furthermore, we present a retrieval-augmented algorithm that effectively exploits related concepts or events in training sets as additional guidance to improve performance on these tasks.