Interested in Contributing?

Scored Systems

System Submitter System Notes Constraint Run Notes BLEU BLEU-cased TER BEER 2.0 CharactTER
uedin-nmt  (Details) barry
University of Edinburgh
Nematus deep model, layer norm, bpe, ldc synthetic. yes Ensemble of 4 l-r, reranked with ensemble of 4 r-l.

36.3

36.3

1.334

0.351

0.601

xmunmt  (Details) Boli Wang
Xiamen University
BPE, CWMT + UN + Xinhua synthetic yes ensemble of 4 models

35.8

35.8

1.365

0.350

0.610

test123  (Details) serenade
Peking University
for test yes

35.1

35.1

0.999

0.347

0.623

toy-nmt  (Details) serenade
Peking University
Single Model yes

35.1

35.1

0.999

0.347

0.623

test123  (Details) serenade
Peking University
for test yes

35.0

35.0

0.999

0.347

0.625

SogouKnowing-nmt  (Details) Sogou Knowing
Sogou
Ensemble of 5 model, rerank with NMT variants and NgramLM feature, BPE yes Ensemble of 5 model, rerank with NMT variants and NgramLM feature, BPE

34.9

34.9

1.402

0.345

0.607

test123  (Details) serenade
Peking University
for test yes

34.9

34.8

0.999

0.346

0.606

uedin-nmt  (Details) barry
University of Edinburgh
Nematus deep model, layer norm, bpe, ldc synthetic. yes Single best

34.6

34.6

1.246

0.343

0.623

test123  (Details) serenade
Peking University
for test yes

34.5

34.5

failed

0.345

0.612

xmunmt  (Details) Boli Wang
Xiamen University
BPE, CWMT + UN + Xinhua synthetic yes single model

34.3

34.3

1.389

0.343

0.627

test-sky  (Details) gtang
Uppsala University
yes

34.0

34.0

28.968

0.182

0.781

test123  (Details) serenade
Peking University
for test yes

33.5

33.5

0.998

0.340

0.622

bit_1  (Details) bit-nmt
first yes

33.1

33.1

failed

0.338

0.635

test123  (Details) serenade
Peking University
for test yes

32.2

32.2

1.019

0.334

0.628

SogouKnowing-nmt  (Details) Sogou Knowing
Sogou
Single deep NMT model yes Single deep NMT model, reranked with NMT variants and other featues, BPE, etc.

31.7

31.7

1.859

0.330

0.652

jingjingz-en-zh  (Details) jingjingz
no transformer-23000

31.5

31.5

17.648

0.220

0.757

jhu-nmt  (Details) sding
Johns Hopkins University
Nematus, BPE, CWMT + UN, 4 ensemble, subsample of 2m back-translated data from xmu-monolingual corpus yes

31.1

31.1

failed

0.328

0.659

CASICT-DCU_NMT  (Details) peerachet
Institute Of Computing Technology, Chinese Academy Of Sciences
Neural Machine Translation (RNNSearch) Training Data : CWMT+WMT (In cooperation with School of Computing, Dublin City University) yes NMT RNNSearch (CWMT+WMT) (Preserve Chinese punctuation)

30.5

30.5

0.999

0.325

0.697

test123  (Details) serenade
Peking University
for test yes

30.4

30.4

failed

0.320

0.758

xmunmt  (Details) Boli Wang
Xiamen University
BPE, CWMT + UN yes single model

30.4

30.4

2.513

0.321

0.660

test123  (Details) serenade
Peking University
for test yes

30.3

30.3

0.999

0.324

0.664

test123  (Details) serenade
Peking University
for test yes

29.8

29.8

1.000

0.321

0.660

test123  (Details) serenade
Peking University
for test yes

27.8

27.8

0.999

0.310

0.693

test123  (Details) serenade
Peking University
for test yes

27.2

27.2

0.999

0.306

0.710

Transformer_UNOnly  (Details) peerachet
Institute Of Computing Technology, Chinese Academy Of Sciences
yes

27.2

27.2

1.000

0.305

0.677

NMT359  (Details) Frank
Zhenyi
yes

26.9

26.8

failed

0.275

0.868

Oregon State University S  (Details) cosmmb
Oregon State University
two layers NMT, bpe, decoding search yes

25.9

25.9

5.622

failed

failed

Oregon State University R  (Details) cosmmb
Oregon State University
two layers NMT, bpe, decoding search yes

25.5

25.5

6.757

failed

failed

UU-HNMT  (Details) gtang
Uppsala University
Helsinki Neural Machine Translation system yes HNMT with hybrid encoder and character-level decoder, ensemble of 3 models.

23.9

23.9

1.000

failed

failed

PROMT SMT  (Details) Alex Molchanov
PROMT LLC
no

23.5

23.5

2.292

0.285

0.788

ParFDA  (Details) bicici
ParFDA Moses Phrase-based SMT system yes en-zh (zh segmented with Stanford segmenter, 2.8 million training sentences)

20.9

20.9

3.101

failed

failed

ParFDA  (Details) bicici
ParFDA Moses Phrase-based SMT system yes en-zh (zh segmented with Stanford segmenter)

19.6

19.6

2.888

failed

failed

test-mt  (Details) lyonlys
None
no
Testing system  (Details) Myl
Pnt ut
nematus, 3models, NE, bpe yes nematus, 3models, NE, bpe

failed

failed

failed

0.173

0.803

test-mt  (Details) lyonlys
None
no

failed

failed

0.000

0.820

0.000

test-mt  (Details) lyonlys
None
no
shannonai  (Details) zhujinyi
Peking University
no

failed

failed

1.232

0.104

failed

test-mt  (Details) lyonlys
None
no