Learning to Control the Fine-grained Sentiment for Story Ending Generation

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 6020–6026 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 6020 Learning to Control the Fine-grained Sentiment for Story Ending Generation Fuli Luo1⇤, Damai Dai1,3⇤, Pengcheng Yang1,2, Tianyu Liu1, Baobao Chang1,3, Zhifang Sui1,3, Xu Sun1,2 1Key Lab of Computational Linguistics, School of EECS, Peking University 2Deep Learning Lab, Beijing Institute of Big Data Research, Peking University 3Peng Cheng Laboratory, China {luofuli,daidamai,yang pc,tianyu0421,chbb,szf,xusun}@pku.edu.cn Abstract Automatic story ending generation is an inter- esting and challenging task in natural language generation. Previous studies are mainly lim- ited to generate coherent, reasonable and di- versified story endings, and few works focus on controlling the sentiment of story endings. This paper focuses on generating a story end- ing which meets the given fine-grained senti- ment intensity. There are two major challenges to this task. First is the lack of story corpus which has fine-grained sentiment labels. Sec- ond is the difficulty of explicitly controlling sentiment intensity when generating endings. Therefore, we propose a generic and novel framework which consists of a sentiment ana- lyzer and a sentimental generator, respectively addressing the two challenges. The sentiment analyzer adopts a series of methods to ac- quire sentiment intensities of the story dataset. The sentimental generator introduces the sen- timent intensity into decoder via a Gaussian Kernel Layer to control the sentiment of the output. To the best of our knowledge, this is the first endeavor to control the fine-grained sentiment for story ending generation with- out manually annotating sentiment labels. Ex- periments show that our proposed framework can generate story endings which are not only more coherent and fluent but also able to meet the given sentiment intensity better.1 1 Introduction Story ending generation aims at completing the plot and concluding a story given a story con- text. Previous works mainly study on how to gen- erate a coherent, reasonable and diversified story ending (Li et al., 2018; Guan et al., 2018; Xu et al., 2018). However, few of them focus on controllable story ending generation, especially ⇤Equal Contribution. 1Our code and data can be found at https://github. com/luofuli/sentimental-story-ending Target Sentiment Generated Story Endings 0.1 She still lost the game and was very upset. 0.3 She almost won the game, but eventually lost. 0.5 The game ended with a draw. 0.7 She eventually won the game. 0.9 She won the game and was very proud of her team. Story context: Sally really loves to play soccer. She joined a team with her friends and she plays everyday. Her coach and her teammates are all really fun. Sally practiced extra hard for her first match. Figure 1: An example of the input story context and output story endings for this task. All of the story end- ings are coherent with the story context but express dif- ferent sentiment intensities. controlling the sentiment for story ending gener- ation. Yao et al. (2018b) is the only work on controlling the sentiment for story …

Content truncated to 500 words