English
Agreement on Target-Bidirectional LSTMs for Sequence-to-Sequence Learning
文件大小：
972k
资源说明：Recurrent neural networks, particularly the long short-term
memory networks, are extremely appealing for sequence-tosequence
learning tasks. Despite their great success, they typically
suffer from a fundamental shortcoming: they are prone
to generate unbalanced targets with good prefixes but bad
suffixes, and thus performance suffers when dealing with
long sequences. We propose a simple yet effective approach
to overcome this shortcoming. Our approach relies on the
agreement between a pair of target-directional LSTMs, which
generates more balanced targets. In addition, we develop
two efficient approximate search methods for agreement that
are empirically shown to be almost optimal in terms of
sequence-level losses. Extensive experiments were performed
on two standard sequence-to-sequence transduction tasks:
machine transliteration and grapheme-to-phoneme transformation.
The results show that the proposed approach achieves
consistent and substantial improvements, compared to six
state-of-the-art systems. In particular, our approach outperforms
the best reported error rates by a margin (up to 9% relative
gains) on the grapheme-to-phoneme task. Our toolkit is
publicly available on https://github.com/lemaoliu/Agtarbidir.
本源码包内暂不包含可直接显示的源代码文件，请下载源码包。
