10
Brief History of Neural Machine Translation Jamie Seol Jamie Seol

Brief History of Neural Machine Translation

Embed Size (px)

Citation preview

Page 1: Brief History of Neural Machine Translation

Brief History of Neural Machine Translation

Jamie Seol

Jamie Seol

Page 2: Brief History of Neural Machine Translation

Jamie Seol

Machine Translation

Lee, et al. "Naver machine translation system for WAT 2015." Proc. WAT 2015

Page 3: Brief History of Neural Machine Translation

Jamie Seol

Traditional Statistical Machine Translation

심광섭, and 김영택. "[기술해설] 기계 번역 시스템." 정보과학회지 12.8 (1994): 17-23.

Page 4: Brief History of Neural Machine Translation

Jamie Seol

Traditional Statistical Machine Translation

Nguyen, et al. "A tree-to-string phrase-based model for statistical machine translation." Proc. CoNLL 2008.

Page 5: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• pros• no need domain knowledge• training multiple features jointly

• cons• training and decoding is quite slow• out-of-vocabulary problem

Lee, et al. "Naver machine translation system for WAT 2015." Proc. WAT 2015

Page 6: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• sequence-to-sequence (Sutskever et al. NIPS 2014)

Sutskever, et al. "Sequence to sequence learning with neural networks." NIPS 2014.

Page 7: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• encoder-decoder (Cho et al. EMNLP 2014)

Cho, et al. "Learning phrase representations using RNN encoder-decoder for statistical machine translation." EMNLP 2014.

Page 8: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• alignment problem!

Tamura, et al. "Recurrent Neural Networks for Word Alignment Model." ACL 2014.

Yang, et al. "Word Alignment Modeling with Context Dependent Deep Neural Network." ACL 2013.

Page 9: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• jointly learning to align (Bahdanau et al., ICLR 2015)

Bahdanau, et al. "Neural machine translation by jointly learning to align and translate." ICLR 2015. (arXiv 2014)

Page 10: Brief History of Neural Machine Translation

Jamie Seol

Neural Machine Translation

• most of recent works use encoder-decoder model• Lu, Liang, et al. "A study of the recurrent neural network encoder-decoder for large

vocabulary speech recognition." Proc. Interspeech 2015. • Luong, et al. "Effective approaches to attention-based neural machine translation."

EMNLP 2015. • Eriguchi, et al. "Tree-to-Sequence Attentional Neural Machine Translation." ACL

2016. • Cho, et al. "On the properties of neural machine translation: Encoder-decoder

approaches." SSST 2014. • Lee, et al. "Naver machine translation system for WAT 2015." Proc. WAT 2015