Data-to-text generation can be conceptually divided into two parts: ordering and structuring the information (planning), and generating fluent language describing the information (realization). Modern neural generation systems conflate these two steps into a single end-to-end differentiable system. We propose to split the generation process into a symbolic text-planning stage that is faithful to the input, followed by a neural generation stage that focuses only on realization. For training a plan-to-text generator, we present a method for matching reference texts to their corresponding text plans. For inference time, we describe a method for selecting high-quality text plans for new inputs. We implement and evaluate our approach on the WebNLG benchmark. Our results demonstrate that decoupling text planning from neural realization indeed improves the system's reliability and adequacy while maintaining fluent output. We observe improvements both in BLEU scores and in manual evaluations. Another benefit of our approach is the ability to output diverse realizations of the same input, paving the way to explicit control over the generated text structure.
Long and Diverse Text Generation with Planning-based Hierarchical Variational Model
Institute for Artificial Intelligence(人工智能研究院);Beijing National Research Center for Information Science and Technology(北京信息科学与技术国家研究中心);Tsinghua University(清华大学);Baozun(宝尊)
Improving Quality and Efficiency in Plan-based Neural Data-to-Text Generation
Bar Ilan University(巴伊兰大学);Allen Institute for Artificial Intelligence(艾伦人工智能研究所)
我们遵循我们在Moryossef et al(2019)中提出的神经数据到文本生成的分步方法,其中生成过程分为文本规划阶段和规划实现阶段。我们对该框架提出了四个扩展:(1)我们引入了一个可训练的神经规划组件,它可以比原始规划器快几个数量级地生成有效的计划;(2)我们引入了输入提示,提高了模型处理不可见关系和实体的能力;(3)我们引入了一个通过重新排序进行验证的阶段,大大提高了结果文本的忠实性;(4)引入了一个简单有效的指称表达式生成模块。这些扩展导致生成过程更快、更流畅和更准确。
Learning to Plan and Realize Separately for Open-Ended Dialogue Systems
University of North Carolina Charlotte(北卡罗来纳大学夏洛特分校);Institute for Human Machine Cognition (IHMC)(人机认知研究所);State University of New York Albany(纽约州立大学奥尔巴尼分校);Rensselaer Polytechnic Institute(伦斯勒理工学院)
Can Neural Generators for Dialogue Learn Sentence Planning and Discourse Structuring?
University of California, Santa Cruz(加州大学圣克鲁兹分校)
任务型对话系统中的应答往往实现多个命题,其最终形式取决于句子规划和语篇构建操作的使用。例如,推荐可以由明确的评价话语组成,例如Chanpen Thai是最佳选项,以及由理由话语关系相关的内容,例如It has great food and service,其将多个命题组合成单个短语。虽然神经生成方法在一个端到端学习框架中集成了句子规划和表面实现,但以前的工作尚未表明神经生成器可以:(1)执行常见的句子规划和话语结构化操作;(2)决定是在单个句子中还是在多个句子中实现内容;(3)决定是否在一个句子中实现内容。(3)概括训练中所见之外的句子规划和语篇关系操作。我们系统地创建大型训练语料库,展示特定的句子规划操作,然后测试神经模型,看看它们学到了什么。我们比较了没有显式潜变量的句子规划模型与在训练过程中提供显式监督的模型。我们表明,只有额外的监督模型可以重现句子规划和话语操作,并推广到训练中看不到的情况。