arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2204.09453cs.CL
收藏

面向开放式文本生成的事件转换规划

Event Transition Planning for Open-ended Text Generation

The University of Hong Kong(香港大学);Tencent AI Lab(腾讯人工智能实验室);Shandong University(山东大学);Shanghai Artificial Intelligence Laboratory(上海人工智能实验室)

Qintong Li, Piji Li, Wei Bi, Zhaochun Ren, Yuxuan Lai, Lingpeng Kong

摘要:

开放式文本生成任务,如对话生成和故事完成,需要模型在有限的先前上下文下生成连贯的延续。这些任务的开放性给当今的神经自回归文本生成器带来了新的挑战。尽管这些神经模型擅长产生类似人类的文本,但它们很难安排给定事实和可能的后续事件之间的因果关系和关系。为了弥合这一差距,我们提出了一种新的两阶段的方法,明确安排随后的事件在开放式文本生成。我们的方法可以被理解为一个专门训练的粗到细的算法,其中一个事件过渡规划器提供了一个“粗糙”的情节骨架和文本生成器在第二阶段细化骨架。在两个开放式文本生成任务上的实验表明,该方法有效地提高了生成文本的质量,特别是在一致性和多样性方面。代码可在以下网址获得:\url{https://github.com/qtli/EventPlanforTextGen}。

英文摘要:

Open-ended text generation tasks, such as dialogue generation and story completion, require models to generate a coherent continuation given limited preceding context. The open-ended nature of these tasks brings new challenges to the neural auto-regressive text generators nowadays. Despite these neural models are good at producing human-like text, it is difficult for them to arrange causalities and relations between given facts and possible ensuing events. To bridge this gap, we propose a novel two-stage method which explicitly arranges the ensuing events in open-ended text generation. Our approach can be understood as a specially-trained coarse-to-fine algorithm, where an event transition planner provides a "coarse" plot skeleton and a text generator in the second stage refines the skeleton. Experiments on two open-ended text generation tasks demonstrate that our proposed method effectively improves the quality of the generated text, especially in coherence and diversity. The code is available at: \url{https://github.com/qtli/EventPlanforTextGen}.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑