arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.12086cs.CL
收藏

稳定样式转换器Transformer:基于编码器-解码器的文本样式转换删除生成方法

Stable Style Transformer: Delete and Generate Approach with Encoder-Decoder for Text Style Transfer

Kakao Enterprise Corp.(Kakao企业公司)

Joosung Lee

摘要:

文本风格转换是指在保持输入句子内容的基础上,转换文本的风格,从而生成句子的任务。由于并行数据集的有限性和难以构建性,现有的研究大多是在非并行数据集上进行的。在这项工作中,我们介绍了一种方法,在非并行数据集的两个阶段。第一阶段是通过分类器直接删除句子的属性标记。第二阶段是通过组合内容令牌和目标样式来生成转移的句子。我们在两个基准数据集上进行实验,并评估上下文,风格,流畅性和语义。仅使用这些自动度量很难选择最佳系统,但可以选择稳定的系统。我们认为,在所有的自动评估指标中,只有鲁棒系统才是可以在实际应用中使用的最低条件。许多以前的系统很难在某些情况下使用,因为性能在几个评估指标中显着较低。然而,我们的系统在所有自动评估指标中都是稳定的,并且具有与其他模型相当的结果。此外,我们比较了我们的系统和不稳定的系统,通过人的评价的性能结果。我们的代码和数据可以在链接(https://github.com/rungjoo/Stable-Style-Transformer)上找到。

英文摘要:

Text style transfer is the task that generates a sentence by preserving the content of the input sentence and transferring the style. Most existing studies are progressing on non-parallel datasets because parallel datasets are limited and hard to construct. In this work, we introduce a method that follows two stages in non-parallel datasets. The first stage is to delete attribute markers of a sentence directly through a classifier. The second stage is to generate a transferred sentence by combining the content tokens and the target style. We experiment on two benchmark datasets and evaluate context, style, fluency, and semantic. It is difficult to select the best system using only these automatic metrics, but it is possible to select stable systems. We consider only robust systems in all automatic evaluation metrics to be the minimum conditions that can be used in real applications. Many previous systems are difficult to use in certain situations because performance is significantly lower in several evaluation metrics. However, our system is stable in all automatic evaluation metrics and has results comparable to other models. Also, we compare the performance results of our system and the unstable system through human evaluation. Our code and data are available at the link (https://github.com/rungjoo/Stable-Style-Transformer).

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。

↑