Research · September 10, 2014

Sequence-to-sequence

Neural sequence transformation was applied to translation.

Why it matters

It demonstrated end-to-end translation with paired recurrent networks.

The reported scores are for the stated WMT task, not a general measure of language understanding.

Event date

September 10, 2014

Timeline date

Event date

Verification

Sources gathered automatically · September 11, 2026

Lines

ID

evt-0008

Builds on LSTM
The authors explicitly use multilayer LSTM networks to encode the input sequence and generate a translation.
Sequence to Sequence Learning
Related Attention for translation
Two papers from the same month on neural sequence translation.

Builds on GNMT: neural translation in Google Translate
The post explains GNMT as recurrent networks learning a mapping from an input sequence to an output sequence directly, citing Sutskever, Vinyals and Le 2014.
A Neural Network for Machine Translation, at Production Scale
Builds on Show and Tell: a caption for an image
The model is carried over from encoder-decoder machine translation (Sutskever, Vinyals, Le): an image is encoded in place of a source sentence.
Show and Tell: A Neural Image Caption Generator (arXiv:1411.4555v1)