Neural Machine Translation by Jointly Learning to Align and Translate
Dzmitry Bahdanau, Kyunghyun Cho, and Yoshua Bengio submitted 'Neural Machine Translation by Jointly Learning to Align and Translate' on September 1, 2014. The paper proposed extending encoder-decoder neural machine translation by allowing the model to automatically soft-search for source-sentence parts relevant to predicting each target word rather than relying only on a single fixed-length source representation.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented neural-machine-translation research milestone for learned soft alignment/attention in sequence models, addressing the fixed-length-vector bottleneck in encoder-decoder translation.
People, organizations & technologies
PEOPLE
Dzmitry Bahdanau, Kyunghyun Cho, Yoshua Bengio
TECHNOLOGIES
Neural machine translation, Soft alignment, Attention mechanism, Encoder-decoder
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-10T04:36:34+08:00
The original arXiv record establishes the September 1, 2014 first-submission date, authors, and the proposed soft-search/alignment mechanism for neural machine translation. The official ICLR 2015 archive independently corroborates the paper title, authors, and conference presentation. This record uses the arXiv first-submission date strictly as a research-publication milestone. It does not infer an exact software-release date, geography, institutional affiliation, ancestry, downstream influence, causality, or relationships.
Sources
Neural Machine Translation by Jointly Learning to Align and Translate
arXiv
PUBLISHED
2014-09-01
ICLR 2015
International Conference on Learning Representations
Trace the chronology
Record identity
LINEAiGE IDbahdanau-attention-2014