Show, Attend and Tell: Neural Image Caption Generation with Visual Attention

方法与贡献

为图像描述生成引入确定性软视觉注意力和随机硬视觉注意力。

随机注意力的早期参考工作，用于理解期望特征与采样选择之间的区别。

结合主题阅读

把注意力建模为隐变量：学习注意力的概率分布 →

引用

Kelvin Xu, Jimmy Lei Ba, Ryan Kiros, Kyunghyun Cho, Aaron Courville, Ruslan Salakhutdinov, Richard S. Zemel, Yoshua Bengio. Show, Attend and Tell: Neural Image Caption Generation with Visual Attention. ICML 2015.

@inproceedings{xu2015show,
  title = {{Show, Attend and Tell: Neural Image Caption Generation with Visual Attention}},
  author = {Xu, Kelvin and Ba, Jimmy Lei and Kiros, Ryan and Cho, Kyunghyun and Courville, Aaron and Salakhutdinov, Ruslan and Zemel, Richard S. and Bengio, Yoshua},
  year = {2015},
  booktitle = {ICML 2015},
  url = {https://proceedings.mlr.press/v37/xuc15.html},
  eprint = {1502.03044},
  archivePrefix = {arXiv}
}