NICT_LOGO.JPG KYOTO-U_LOGO.JPG

WAT

The Workshop on Asian Translation
Evaluation Results

[EVALUATION RESULTS TOP] | [BLEU] | [RIBES] | [AMFM] | [HUMAN (WAT2022)] | [HUMAN (WAT2021)] | [HUMAN (WAT2020)] | [HUMAN (WAT2019)] | [HUMAN (WAT2018)] | [HUMAN (WAT2017)] | [HUMAN (WAT2016)] | [HUMAN (WAT2015)] | [HUMAN (WAT2014)] | [EVALUATION RESULTS USAGE POLICY]

BLEU


# Team Task Date/Time DataID BLEU
Method
Other
Resources
System
Description
juman kytea mecab moses-
tokenizer
stanford-
segmenter-
ctb
stanford-
segmenter-
pku
indic-
tokenizer
unuse myseg kmseg
1ORGANIZERJPCja-zh2016/07/13 15:11:50966-30.60--32.0331.25- 0.00 0.00 0.00SMTNoPhrase-based SMT
2ORGANIZERJPCja-zh2016/07/13 15:16:28967-30.26--31.5730.91- 0.00 0.00 0.00SMTNoHierarchical Phrase-based SMT
3ORGANIZERJPCja-zh2016/07/13 15:31:58968-31.05--32.3531.70- 0.00 0.00 0.00SMTNoString-to-Tree SMT
4ORGANIZERJPCja-zh2016/07/26 10:35:571038-23.02--23.5723.29- 0.00 0.00 0.00OtherYesOnline A (2016)
5ORGANIZERJPCja-zh2016/08/01 18:33:201069- 9.42-- 9.59 8.79- 0.00 0.00 0.00OtherYesOnline B (2016)
6NICT-2JPCja-zh2016/08/04 17:37:251081-33.35--34.6433.81- 0.00 0.00 0.00SMTNoPhrase-based SMT with Preordering + Domain Adaptation
7NICT-2JPCja-zh2016/08/05 18:13:161106-33.40--34.6433.83- 0.00 0.00 0.00SMTYesPhrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC)
8ORGANIZERJPCja-zh2016/08/08 17:57:511118-12.35--13.7213.17- 0.00 0.00 0.00OtherYesRBMT C (2016)
9bjtu_nlpJPCja-zh2016/08/16 12:45:011150-31.49--32.7932.51- 0.00 0.00 0.00NMTNoRNN Encoder-Decoder with attention mechanism, single model
10SenseJPCja-zh2016/08/29 01:08:541282-29.87--32.1130.75- 0.00 0.00 0.00SMTNoClustercat-C10-PBMT
11SenseJPCja-zh2016/08/29 09:52:571283-29.59--31.8430.46- 0.00 0.00 0.00SMTNoBaseline-C10-PBMT
12SenseJPCja-zh2016/08/29 23:10:281293-29.59--31.8830.44- 0.00 0.00 0.00SMTNoBaseline-C50-PBMT
13SenseJPCja-zh2016/08/30 08:15:191295-29.44--31.7130.36- 0.00 0.00 0.00SMTNoClustercat-C50-PBMT
14ORGANIZERJPCja-zh2016/11/16 11:16:251340-33.04--33.9233.34- 0.00 0.00 0.00NMTYesOnline A (2016/11/14)
15u-tkbJPCja-zh2017/07/26 12:33:411465-31.80--33.1932.83- 0.00 0.00 0.00NMTNoNMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
16ORGANIZERJPCja-zh2018/08/15 18:14:011960-39.07--40.3239.75-- 0.00 0.00NMTNoNMT with Attention
17USTCJPCja-zh2018/08/31 16:54:462202-39.71--40.5440.05-- 0.00 0.00NMTNotensor2tensor, 4 model average, r2l rerank
18ryanJPCja-zh2019/07/25 22:04:252950-41.22--42.3741.99----NMTNoBase Transformer
19sarahJPCja-zh2019/07/26 11:39:522982-42.13--43.2742.98----NMTNoTransformer, ensemble of 4 models
20KNU_HyundaiJPCja-zh2019/07/27 08:37:113162-43.19--44.5643.86----NMTYesTransformer(base) + *Used ASPEC corpus* with relative position, bt, r2l rerank, 4-model ensemble(9-check point)
21goku20JPCja-zh2020/09/21 11:55:294088-39.49--40.6140.14----NMTNomBART pre-training transformer, single model
22goku20JPCja-zh2020/09/22 00:06:064106-40.47--41.6841.22----NMTNomBART pre-training transformer, ensemble of 3 models
23tpt_watJPCja-zh2021/04/27 01:49:085694-37.83--39.1738.89----NMTNoBase Transformer model with separated vocabs 8k size
24Bering LabJPCja-zh2021/04/30 14:56:115853-43.48--44.6744.24----NMTYesTransformer Ensemble with additional crawled parallel corpus

Notice:

Back to top

RIBES


# Team Task Date/Time DataID RIBES
Method
Other
Resources
System
Description
juman kytea mecab moses-
tokenizer
stanford-
segmenter-
ctb
stanford-
segmenter-
pku
indic-
tokenizer
unuse myseg kmseg
1ORGANIZERJPCja-zh2016/07/13 15:11:50966-0.787321--0.7978880.794388-0.0000000.0000000.000000SMTNoPhrase-based SMT
2ORGANIZERJPCja-zh2016/07/13 15:16:28967-0.788415--0.7991180.796685-0.0000000.0000000.000000SMTNoHierarchical Phrase-based SMT
3ORGANIZERJPCja-zh2016/07/13 15:31:58968-0.793846--0.8028050.800848-0.0000000.0000000.000000SMTNoString-to-Tree SMT
4ORGANIZERJPCja-zh2016/07/26 10:35:571038-0.754241--0.7606720.760148-0.0000000.0000000.000000OtherYesOnline A (2016)
5ORGANIZERJPCja-zh2016/08/01 18:33:201069-0.642026--0.6510700.643520-0.0000000.0000000.000000OtherYesOnline B (2016)
6NICT-2JPCja-zh2016/08/04 17:37:251081-0.808513--0.8179960.815322-0.0000000.0000000.000000SMTNoPhrase-based SMT with Preordering + Domain Adaptation
7NICT-2JPCja-zh2016/08/05 18:13:161106-0.811788--0.8203200.818701-0.0000000.0000000.000000SMTYesPhrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC)
8ORGANIZERJPCja-zh2016/08/08 17:57:511118-0.688240--0.7086810.700210-0.0000000.0000000.000000OtherYesRBMT C (2016)
9bjtu_nlpJPCja-zh2016/08/16 12:45:011150-0.816577--0.8229780.820820-0.0000000.0000000.000000NMTNoRNN Encoder-Decoder with attention mechanism, single model
10SenseJPCja-zh2016/08/29 01:08:541282-0.791760--0.8032880.798268-0.0000000.0000000.000000SMTNoClustercat-C10-PBMT
11SenseJPCja-zh2016/08/29 09:52:571283-0.790983--0.8034000.797822-0.0000000.0000000.000000SMTNoBaseline-C10-PBMT
12SenseJPCja-zh2016/08/29 23:10:281293-0.788582--0.8013820.796637-0.0000000.0000000.000000SMTNoBaseline-C50-PBMT
13SenseJPCja-zh2016/08/30 08:15:191295-0.789417--0.8021230.796346-0.0000000.0000000.000000SMTNoClustercat-C50-PBMT
14ORGANIZERJPCja-zh2016/11/16 11:16:251340-0.824829--0.8291220.829067-0.0000000.0000000.000000NMTYesOnline A (2016/11/14)
15u-tkbJPCja-zh2017/07/26 12:33:411465-0.819791--0.8260550.825025-0.0000000.0000000.000000NMTNoNMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
16ORGANIZERJPCja-zh2018/08/15 18:14:011960-0.847112--0.8508510.850913--0.0000000.000000NMTNoNMT with Attention
17USTCJPCja-zh2018/08/31 16:54:462202-0.846472--0.8507500.849818--0.0000000.000000NMTNotensor2tensor, 4 model average, r2l rerank
18ryanJPCja-zh2019/07/25 22:04:252950-0.852969--0.8582040.857643----NMTNoBase Transformer
19sarahJPCja-zh2019/07/26 11:39:522982-0.853579--0.8578050.857191----NMTNoTransformer, ensemble of 4 models
20KNU_HyundaiJPCja-zh2019/07/27 08:37:113162-0.860783--0.8658520.864763----NMTYesTransformer(base) + *Used ASPEC corpus* with relative position, bt, r2l rerank, 4-model ensemble(9-check point)
21goku20JPCja-zh2020/09/21 11:55:294088-0.845571--0.8511730.850139----NMTNomBART pre-training transformer, single model
22goku20JPCja-zh2020/09/22 00:06:064106-0.848852--0.8550670.854644----NMTNomBART pre-training transformer, ensemble of 3 models
23tpt_watJPCja-zh2021/04/27 01:49:085694-0.841632--0.8453600.844727----NMTNoBase Transformer model with separated vocabs 8k size
24Bering LabJPCja-zh2021/04/30 14:56:115853-0.866184--0.8710920.869849----NMTYesTransformer Ensemble with additional crawled parallel corpus

Notice:

Back to top

AMFM


# Team Task Date/Time DataID AMFM
Method
Other
Resources
System
Description
unuse unuse unuse unuse unuse unuse unuse unuse unuse unuse
1ORGANIZERJPCja-zh2016/07/13 15:11:50966-0.710940--0.7109400.710940-0.0000000.0000000.000000SMTNoPhrase-based SMT
2ORGANIZERJPCja-zh2016/07/13 15:16:28967-0.718360--0.7183600.718360-0.0000000.0000000.000000SMTNoHierarchical Phrase-based SMT
3ORGANIZERJPCja-zh2016/07/13 15:31:58968-0.720030--0.7200300.720030-0.0000000.0000000.000000SMTNoString-to-Tree SMT
4ORGANIZERJPCja-zh2016/07/26 10:35:571038-0.702350--0.7023500.702350-0.0000000.0000000.000000OtherYesOnline A (2016)
5ORGANIZERJPCja-zh2016/08/01 18:33:201069-0.527180--0.5271800.527180-0.0000000.0000000.000000OtherYesOnline B (2016)
6NICT-2JPCja-zh2016/08/04 17:37:251081-0.723270--0.7232700.723270-0.0000000.0000000.000000SMTNoPhrase-based SMT with Preordering + Domain Adaptation
7NICT-2JPCja-zh2016/08/05 18:13:161106-0.731520--0.7315200.731520-0.0000000.0000000.000000SMTYesPhrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC)
8ORGANIZERJPCja-zh2016/08/08 17:57:511118-0.475430--0.4754300.475430-0.0000000.0000000.000000OtherYesRBMT C (2016)
9bjtu_nlpJPCja-zh2016/08/16 12:45:011150-0.701490--0.7014900.701490-0.0000000.0000000.000000NMTNoRNN Encoder-Decoder with attention mechanism, single model
10SenseJPCja-zh2016/08/29 01:08:541282-0.709390--0.7093900.709390-0.0000000.0000000.000000SMTNoClustercat-C10-PBMT
11SenseJPCja-zh2016/08/29 09:52:571283-0.707590--0.7075900.707590-0.0000000.0000000.000000SMTNoBaseline-C10-PBMT
12SenseJPCja-zh2016/08/29 23:10:281293-0.710030--0.7100300.710030-0.0000000.0000000.000000SMTNoBaseline-C50-PBMT
13SenseJPCja-zh2016/08/30 08:15:191295-0.710150--0.7101500.710150-0.0000000.0000000.000000SMTNoClustercat-C50-PBMT
14ORGANIZERJPCja-zh2016/11/16 11:16:251340-0.735470--0.7354700.735470-0.0000000.0000000.000000NMTYesOnline A (2016/11/14)
15u-tkbJPCja-zh2017/07/26 12:33:411465-0.706720--0.7067200.706720-0.0000000.0000000.000000NMTNoNMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
16ORGANIZERJPCja-zh2018/08/15 18:14:011960-0.752360--0.7523600.752360--0.0000000.000000NMTNoNMT with Attention
17USTCJPCja-zh2018/08/31 16:54:462202-0.757690--0.7576900.757690--0.0000000.000000NMTNotensor2tensor, 4 model average, r2l rerank
18ryanJPCja-zh2019/07/25 22:04:252950-0.000000--0.0000000.000000----NMTNoBase Transformer
19sarahJPCja-zh2019/07/26 11:39:522982-0.000000--0.0000000.000000----NMTNoTransformer, ensemble of 4 models
20KNU_HyundaiJPCja-zh2019/07/27 08:37:113162-0.000000--0.0000000.000000----NMTYesTransformer(base) + *Used ASPEC corpus* with relative position, bt, r2l rerank, 4-model ensemble(9-check point)
21goku20JPCja-zh2020/09/21 11:55:294088-0.000000--0.0000000.000000----NMTNomBART pre-training transformer, single model
22goku20JPCja-zh2020/09/22 00:06:064106-0.000000--0.0000000.000000----NMTNomBART pre-training transformer, ensemble of 3 models
23tpt_watJPCja-zh2021/04/27 01:49:085694-0.885218--0.8852180.885218----NMTNoBase Transformer model with separated vocabs 8k size
24Bering LabJPCja-zh2021/04/30 14:56:115853-0.899221--0.8992210.899221----NMTYesTransformer Ensemble with additional crawled parallel corpus

Notice:

Back to top

HUMAN (WAT2022)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

HUMAN (WAT2021)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

HUMAN (WAT2020)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

HUMAN (WAT2019)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description
1sarahJPCja-zh2019/07/26 11:39:522982UnderwayNMTNoTransformer, ensemble of 4 models
2KNU_HyundaiJPCja-zh2019/07/27 08:37:113162UnderwayNMTYesTransformer(base) + *Used ASPEC corpus* with relative position, bt, r2l rerank, 4-model ensemble(9-check point)

Notice:
Back to top

HUMAN (WAT2018)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

HUMAN (WAT2017)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description
1u-tkbJPCja-zh2017/07/26 12:33:41146521.750NMTNoNMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)

Notice:
Back to top

HUMAN (WAT2016)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description
1ORGANIZERJPCja-zh2016/11/16 11:16:25134032.500NMTYesOnline A (2016/11/14)
2NICT-2JPCja-zh2016/08/05 18:13:16110614.000SMTYesPhrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC)
3ORGANIZERJPCja-zh2016/07/13 15:16:289674.750SMTNoHierarchical Phrase-based SMT
4ORGANIZERJPCja-zh2016/07/13 15:31:589684.250SMTNoString-to-Tree SMT
5bjtu_nlpJPCja-zh2016/08/16 12:45:011150-1.000NMTNoRNN Encoder-Decoder with attention mechanism, single model
6NICT-2JPCja-zh2016/08/04 17:37:251081-11.000SMTNoPhrase-based SMT with Preordering + Domain Adaptation
7ORGANIZERJPCja-zh2016/07/26 10:35:571038-23.000OtherYesOnline A (2016)
8ORGANIZERJPCja-zh2016/08/08 17:57:511118-41.250OtherYesRBMT C (2016)

Notice:
Back to top

HUMAN (WAT2015)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

HUMAN (WAT2014)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description

Notice:
Back to top

EVALUATION RESULTS USAGE POLICY

When you use the WAT evaluation results for any purpose such as:
- writing technical papers,
- making presentations about your system,
- advertising your MT system to the customers,
you can use the information about translation directions, scores (including both automatic and human evaluations) and ranks of your system among others. You can also use the scores of the other systems, but you MUST anonymize the other system's names. In addition, you can show the links (URLs) to the WAT evaluation result pages.

NICT (National Institute of Information and Communications Technology)
Kyoto University
Last Modified: 2018-08-02