NeurIPS 2020poster77 citations
Prophet Attention: Predicting Attention with Future Attention
Fenglin Liu, Xuancheng Ren, Xian Wu, Shen Ge, Wei Fan, Yuexian Zou, Xu Sun
Abstract
Recently, attention based models have been used extensively in many sequence-to-sequence learning systems. Especially for image captioning, the attention based models are expected to ground correct image regions with proper generated words. However, for each time step in the decoding process, the attention based models usually use the hidden state of the current input to attend to the image regions. Under this setting, these attention models have a
BibTeX
@inproceedings{NEURIPS2020_13fe9d84,
author = {Liu, Fenglin and Ren, Xuancheng and Wu, Xian and Ge, Shen and Fan, Wei and Zou, Yuexian and Sun, Xu},
booktitle = {Advances in Neural Information Processing Systems},
editor = {H. Larochelle and M. Ranzato and R. Hadsell and M.F. Balcan and H. Lin},
pages = {1865--1876},
publisher = {Curran Associates, Inc.},
title = {Prophet Attention: Predicting Attention with Future Attention},
url = {https://proceedings.neurips.cc/paper_files/paper/2020/file/13fe9d84310e77f13a6d184dbf1232f3-Paper.pdf},
volume = {33},
year = {2020}
}