This website requires JavaScript.

GAE-ISumm: Unsupervised Graph-Based Summarization of Indian Languages

Lakshmi Sireesha VakadaAnudeep ChMounika MarreddySubba Reddy OotaRadhika Mamidi
Dec 2022
摘要
Document summarization aims to create a precise and coherent summary of atext document. Many deep learning summarization models are developed mainly forEnglish, often requiring a large training corpus and efficient pre-trainedlanguage models and tools. However, English summarization models forlow-resource Indian languages are often limited by rich morphologicalvariation, syntax, and semantic differences. In this paper, we proposeGAE-ISumm, an unsupervised Indic summarization model that extracts summariesfrom text documents. In particular, our proposed model, GAE-ISumm uses GraphAutoencoder (GAE) to learn text representations and a document summary jointly.We also provide a manually-annotated Telugu summarization dataset TELSUM, toexperiment with our model GAE-ISumm. Further, we experiment with the mostpublicly available Indian language summarization datasets to investigate theeffectiveness of GAE-ISumm on other Indian languages. Our experiments ofGAE-ISumm in seven languages make the following observations: (i) it iscompetitive or better than state-of-the-art results on all datasets, (ii) itreports benchmark results on TELSUM, and (iii) the inclusion of positional andcluster information in the proposed model improved the performance ofsummaries.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?
0
被引用
笔记
问答