Neural Foundations · 2014
Sequence to Sequence Learning with Neural Networks
Introduced a general end-to-end LSTM encoder-decoder that maps an input sequence to a fixed-length vector and decodes it into an output sequence, removing the need for fixed-alignment, feature-engineered pipelines in sequence transduction.
Editorial record
Plain-language summary
The paper used one multilayer LSTM to read an entire source sentence into a single vector and a second LSTM to generate the target sentence from that vector. A key practical trick was reversing the order of source words, which shortened the average distance between corresponding input and output tokens and made optimization easier, pushing WMT'14 English-to-French BLEU to competitive levels. It showed that a purely neural, general-purpose sequence model could rival phrase-based statistical MT and became the template for later encoder-decoder systems.
Knowledge graph
Relationships
Antecedents
ExtendsEvidence: Direct
Attention Is All You Need
Encoder-decoder transduction framing carried over from seq2seq
P-001 background
ExtendsEvidence: Direct
Neural Machine Translation by Jointly Learning to Align and Translate (Bahdanau attention)
Attention removes seq2seq bottleneck
A-007
Descendants
Applies toEvidence: Strongly supported
Long Short-Term Memory (LSTM)
Seq2seq built from LSTMs
A-006
CombinesEvidence: Strongly supported
Gated Recurrent Unit (GRU) / RNN Encoder-Decoder
GRU encoder-decoder sibling of seq2seq
A-006
Source record
Provenance
- Record ID
- A-006
- Record created
- 2026-07-13
- Last reviewed
- 2026-07-14
- Record version
- 2
- https://arxiv.org/abs/1409.3215
- arXiv:1409.3215
Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.