动态协同网络用于问答

Nov, 2016

Dynamic Coattention Networks For Question Answering

Caiming Xiong, Victor Zhong, Richard Socher

TL;DR我们提出了动态协同注意网络（DCN）来解决深度学习模型在问题回答中可能会遇到局部极大值的问题，并在Stanford问题回答数据集上将准确率从71.0% F1提高到75.9％ F1，使用DCN合奏能够获得80.4％ F1。

Abstract

Several deep learning models have been proposed for question answering. However, due to their single-pass nature, they have no way to recover from →