Unsupervised Abstractive Opinion Summarization by Generating Sentences with Tree-Structured Topic Guidance
Unsupervised Abstractive Opinion Summarization by Generating Sentences with Tree-Structured Topic Guidance
复制标题
DOI：
10.1162/tacl_a_00406
复制
发表时间：
2021-06
影响因子：
10.9
通讯作者：
Masaru Isonuma;Junichiro Mori;D. Bollegala;I. Sakata
中科院分区：
文献类型：
--
作者：
Masaru Isonuma;Junichiro Mori;D. Bollegala;I. Sakata
摘要本文提出了一种新的无监督摘要方法。虽然基本的变分自动编码器为基础的模型假设一个单峰高斯先验的潜在代码的句子，我们交替它与递归高斯混合，其中每个混合成分对应于一个主题句的潜在代码，并混合由树结构的主题分布。通过解码每个高斯分量，我们生成具有树形结构的主题指导的句子，其中根句子传达通用内容，叶句子描述特定主题。实验结果表明，所生成的主题句适合作为固执己见的文本的摘要，其比最近的无监督摘要模型（Bražinskas等人，2020年）。此外，我们证明了潜在高斯的方差代表了句子的粒度，类似于高斯词嵌入（Vilnis和McCallum，2015）。
Abstract This paper presents a novel unsupervised abstractive summarization method for opinionated texts. While the basic variational autoencoder-based models assume a unimodal Gaussian prior for the latent code of sentences, we alternate it with a recursive Gaussian mixture, where each mixture component corresponds to the latent code of a topic sentence and is mixed by a tree-structured topic distribution. By decoding each Gaussian component, we generate sentences with tree-structured topic guidance, where the root sentence conveys generic content, and the leaf sentences describe specific topics. Experimental results demonstrate that the generated topic sentences are appropriate as a summary of opinionated texts, which are more informative and cover more input contents than those generated by the recent unsupervised summarization model (Bražinskas et al., 2020). Furthermore, we demonstrate that the variance of latent Gaussians represents the granularity of sentences, analogous to Gaussian word embedding (Vilnis and McCallum, 2015).