nlp领域英文文献中的常见词汇与结构1

本文介绍了一种基于RNN的序列模型SummaRuNNer,它在文档摘要任务上展现出超越或与现有最优技术相当的性能。模型具有高度可解释性,且通过抽象训练方法仅依赖人类生成的参考摘要。研究聚焦于单一文档的句子级提取,对比了抽取式和生成式摘要技术,并指出大部分文献集中在抽取式方法。我们的工作在技术贡献上包括新颖的训练方式和大规模神经网络应用的实验基础。
摘要由CSDN通过智能技术生成
  • 文献一:

Summarunner: A recurrent neural network based sequence model for extractive summarization of documents

Ramesh Nallapati, Feifei Zhai, and Bowen Zhou. Summarunner: A recurrent 
neural network based sequence model for extractive summarization of 
documents. In Thirty-First AAAI Conference on Artificial Intelligence,
 2017
A Recurrent Neural Network based Sequence Model for Extractive 
Summarization of Documents

结构提取:A … based … Model for… 基于…的…模型

 We present SummaRuNNer, a Recurrent Neural Network(RNN) based
 sequence model for extractive summarization of documents and 
 show that it achieves performance better than or comparable to 
 state-of-the-art.

结构提取:we present…(model)…and show that it achieves performance better than or comparable to state-of-the-art 我们提出了一种…模型,并证明其性能优于或可与目前的顶级技术水平相媲美。
(注意state-of-the-art的评价是要有确切指标支撑的,不要轻易在自己的论文中使用,show that 后的语句可进行更换)

 Our model has the additional advantage of being very interpretable, 
 since it allows visualization of its predictions broken up by 
 abstract features such as information content, salience and novelty.

结构提取: Our model has the additional advantage of being very… , since it…我们的模型还有一个额外的优势,那就是…,因为它…
interpretable: 可解释的;可翻译的。在文中的语义条件下可解释为名词:具有很好的解释性;可解释性很强
visualization: n.可视化
prediction:(模型)预测结果
abstract features: 抽象特点

Another novel contribution of our work is abstractive training of
our extractive model that can train on human generated reference
summaries alone, eliminating the need for sentence-level ex-tractive 
labels.

结构提取:Another novel contribution of our work is…(that从句 or 伴随状态)我们工作的另一个新贡献是…,该模型可以…

 Summarization techniques are mainly classified into two categories:
 extractive and abstractive. Ex-tractive methods aim to select salient
 snippets, sentences or passages from documents, while abstractive 
 summarization techniques aim to concisely paraphrase the information
 con-tent in the documents

结构提取:…techniques are mainly classified into two categories:
… and … one aim to …, while the other one aim to 某某技术主要分为两类:a 和 b。a方法旨在…,而b技术旨在…

A vast majority of the literature on document summarization is 
devoted to extractive summarization

结构提取:A vast majority of the literature on…is
devoted to…绝大多数文献都致力于…可以用在介绍之前学术进展的部分

Recently, neural network based approaches have become popular for 
extractive summarization.

结构提取:Recently, … based approaches have become popular for…近年来,基于…的…方法越来越受欢迎。可以用在介绍最新学术进展的部分并引入自己思路

Our main contributions are as follows: (a).....

我们的贡献如下:(罗列创新点)

Like (Cheng and Lapata 2016), our work also focuses only 
onsententialextractive summarization of single documents
using neural networks. We use the same corpus used by
 (Nallapati et al. 2016) and (Cheng and Lapata 2016) for
our experiments, since its large size makes it attractive for
training deep neural networks such as ours, with several
thousands of parameters.

结构提取:Like…, our work also focuses only
on … using … We use the same corpus used by
… for our experiments, since its large size makes it attractive for
training deep neural networks such as ours, with several
thousands of parameters.

评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值