Consistency and Coherency Enhanced Story Generation
Wei Wang,Piji Li,Hai-Tao Zheng
DOI: https://doi.org/10.1007/978-3-030-72113-8_46
2021-01-01
Abstract:Story generation is a challenging task, which demands to maintain consistency of the plots and characters throughout the story. Previous works have shown that GPT2, a large-scale language model, has achieved advanced performance on story generation. However, we observe that several serious issues still exist in the stories generated by GPT2, which can be categorized into two folds: consistency and coherency. In terms of consistency, on the one hand, GPT2 cannot guarantee the consistency of the plots explicitly. On the other hand, the generated stories usually contain coreference errors. In terms of coherency, GPT2 does not take account of the discourse relations between sentences of stories directly. To enhance the consistency and coherency of the generated stories, we propose a two-stage generation framework, where the first stage is to organize the story outline which depicts the story plots and events, and the second stage is to expand the outline into a complete story. Therefore, the consistency of the plots can be controlled and guaranteed explicitly. In addition, coreference supervision signals are incorporated to reduce coreference errors and improve coreference consistency. Moreover, we design an auxiliary task of discourse relation modeling to improve the coherency of the generated stories. Experimental results on a story dataset show that our model outperforms baseline approaches in terms of both automatic metrics and human evaluation.