论文地址：https://arxiv.org/abs/1704.04368
Sequence2Sequence Attention的两个问题：
（1）难以准确复述原文的事实细节、无法处理原文中的未登录词(OOV)；
（2）生成的摘要中存在重复的片段
提出了：
（1）融合了seq2seq模型和pointer network的pointer-generator network
（2）覆盖率机制(coverage mechanism)
相关文章
- 02-13『论文笔记』MoCo：Momentum Contrast for Unsupervised Visual Representation Learning
- 02-13论文笔记系列-iCaRL： Incremental Classifier and Representation Learning
- 02-13『论文笔记』Decoupling Representation and Classifier
- 02-13GRAIL Efficient Time Series Representation Learning论文阅读笔记（二）
- 02-13（CoRL2020）DIRL: Domain-Invariant Representation Learning Approach for Sim-to-Real Transfer 论文笔记
- 02-13【论文笔记】Face Anonymization by Manipulating Decoupled Identity Representation
- 02-13GraphSAGE：Inductive Representation Learning on Large Graphs——论文笔记
- 02-13论文笔记29 -- （Vehicle ReID）Going Beyond Real Data: A Robust Visual Representation for Vehicle Re-id
- 02-13论文阅读笔记（六十九）【CVPR2021】：BiCnet-TKS: Learning Efficient Spatial-Temporal Representation for Video Perso
- 02-13深度学习论文笔记-Deep Learning Face Representation from Predicting 10,000 Classes