Neural Machine Translation by Jointly Learning to Align and Translate
GPTKB entity
Statements (48)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
research paper
|
| gptkbp:application |
Language translation
|
| gptkbp:author |
gptkb:Yoshua_Bengio
gptkb:Dzmitry_Bahdanau Kyunghyun Cho |
| gptkbp:availableIn |
arXiv
|
| gptkbp:citedBy |
over 5000 times
|
| gptkbp:completed |
Attention improves translation
Joint learning is effective |
| gptkbp:contributedTo |
Natural Language Processing
|
| gptkbp:dataUsage |
IWSLT dataset
WMT dataset |
| gptkbp:description |
Alignment between source and target languages
Joint learning framework |
| gptkbp:discusses |
Future directions in research
Challenges in machine translation |
| gptkbp:evaluates |
ROUGE score
BLEU_score |
| gptkbp:fellowship |
10.1109/ICLR.2015.00112
|
| gptkbp:focusesOn |
Neural networks
Machine translation |
| gptkbp:futurePlans |
Improve computational efficiency
Enhance alignment techniques Expand to low-resource languages Explore unsupervised learning Investigate multilingual translation |
| https://www.w3.org/2000/01/rdf-schema#label |
Neural Machine Translation by Jointly Learning to Align and Translate
|
| gptkbp:impact |
Improved translation quality
|
| gptkbp:influencedBy |
Neural networks
Statistical machine translation |
| gptkbp:introduced |
Attention mechanism
|
| gptkbp:keywords |
Deep learning
Alignment Attention mechanism Neural machine translation Sequence-to-sequence learning |
| gptkbp:language |
English
|
| gptkbp:model |
Encoder-decoder architecture
|
| gptkbp:provides |
Theoretical insights
Empirical results |
| gptkbp:publishedIn |
gptkb:Proceedings_of_the_International_Conference_on_Learning_Representations_(ICLR)
|
| gptkbp:relatedTo |
Deep learning
Word embeddings Recurrent neural networks Sequence-to-sequence models |
| gptkbp:technique |
Attention-based model
End-to-end learning |
| gptkbp:year |
2015
|