NICT_LOGO.JPG KYOTO-U_LOGO.JPG

WAT

The Workshop on Asian Translation
Evaluation Results

[EVALUATION RESULTS TOP] | [BLEU] | [RIBES] | [AMFM] | [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
1NHK-NESJIJIen-ja2019/07/24 14:20:51288528.2529.7628.63-------NMTYesensemble-6, beamwidth-30 with other resources
2NHK-NESJIJIen-ja2019/07/24 14:23:33288627.2228.7527.63-------NMTYesensemble-6, beamwidth-30 using another domain-tag
3sarahJIJIen-ja2019/07/22 16:44:31281422.6523.4822.76-------NMTNoTransformer, ensemble of 4 models
4sarahJIJIen-ja2019/07/22 16:47:09281521.8022.6721.91-------NMTNoTransformer, single model
5NTTJIJIen-ja2017/08/01 04:17:58167920.3721.8220.68---- 0.00 0.00 0.00NMTNoEnsemble 8 Models: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
6XMUNLPJIJIen-ja2017/07/25 10:33:31144319.6120.7220.14---- 0.00 0.00 0.00NMTNoensemble of 4 nmt models
7NTTJIJIen-ja2017/07/30 10:41:00160319.1320.4719.41---- 0.00 0.00 0.00NMTNoSingle Model: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
8XMUNLPJIJIen-ja2017/07/24 09:11:36142917.4818.5717.92---- 0.00 0.00 0.00NMTNosingle nmt model
9ORGANIZERJIJIen-ja2018/08/14 11:46:13190416.6317.0116.36----- 0.00 0.00NMTNoNMT with Attention
10ORGANIZERJIJIen-ja2017/07/19 13:01:12139516.2216.9516.22---- 0.00 0.00 0.00SMTNoHierarchical Phrase-based SMT
11ORGANIZERJIJIen-ja2017/07/19 12:47:25139315.7716.6515.76---- 0.00 0.00 0.00SMTNoPhrase-based SMT
12ORGANIZERJIJIen-ja2017/07/19 13:34:40139714.9515.3814.79---- 0.00 0.00 0.00SMTNoTree-to-String SMT
13ORGANIZERJIJIen-ja2017/07/28 15:13:37151811.2913.1211.84---- 0.00 0.00 0.00NMTNoONLINE-A
14ORGANIZERJIJIen-ja2017/07/28 15:15:121519 7.46 8.73 7.63---- 0.00 0.00 0.00NMTNoONLINE-B
15ORGANIZERJIJIen-ja2017/07/28 15:04:141514 5.31 6.68 5.69---- 0.00 0.00 0.00RBMTNoRBMT-A
16ORGANIZERJIJIen-ja2017/07/28 15:06:231515 4.72 5.98 4.97---- 0.00 0.00 0.00RBMTNoRBMT-B

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
1NHK-NESJIJIen-ja2019/07/24 14:20:5128850.7390360.7411640.743977-------NMTYesensemble-6, beamwidth-30 with other resources
2NHK-NESJIJIen-ja2019/07/24 14:23:3328860.7308160.7316530.734618-------NMTYesensemble-6, beamwidth-30 using another domain-tag
3sarahJIJIen-ja2019/07/22 16:44:3128140.7050440.7072090.710111-------NMTNoTransformer, ensemble of 4 models
4sarahJIJIen-ja2019/07/22 16:47:0928150.6986090.6999810.701739-------NMTNoTransformer, single model
5XMUNLPJIJIen-ja2017/07/25 10:33:3114430.6841200.6884970.691056----0.0000000.0000000.000000NMTNoensemble of 4 nmt models
6NTTJIJIen-ja2017/08/01 04:17:5816790.6805980.6840480.688863----0.0000000.0000000.000000NMTNoEnsemble 8 Models: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
7NTTJIJIen-ja2017/07/30 10:41:0016030.6685170.6709200.676594----0.0000000.0000000.000000NMTNoSingle Model: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
8XMUNLPJIJIen-ja2017/07/24 09:11:3614290.6593720.6655780.668759----0.0000000.0000000.000000NMTNosingle nmt model
9ORGANIZERJIJIen-ja2018/08/14 11:46:1319040.6537660.6572790.658830-----0.0000000.000000NMTNoNMT with Attention
10ORGANIZERJIJIen-ja2017/07/28 15:13:3715180.5974730.6055320.603374----0.0000000.0000000.000000NMTNoONLINE-A
11ORGANIZERJIJIen-ja2017/07/19 13:01:1213950.5949230.6015050.602516----0.0000000.0000000.000000SMTNoHierarchical Phrase-based SMT
12ORGANIZERJIJIen-ja2017/07/19 13:34:4013970.5940720.5977910.599530----0.0000000.0000000.000000SMTNoTree-to-String SMT
13ORGANIZERJIJIen-ja2017/07/19 12:47:2513930.5802840.5848920.585437----0.0000000.0000000.000000SMTNoPhrase-based SMT
14ORGANIZERJIJIen-ja2017/07/28 15:15:1215190.5405580.5454300.545178----0.0000000.0000000.000000NMTNoONLINE-B
15ORGANIZERJIJIen-ja2017/07/28 15:06:2315150.5184160.5316030.532079----0.0000000.0000000.000000RBMTNoRBMT-B
16ORGANIZERJIJIen-ja2017/07/28 15:04:1415140.5052270.5150500.513580----0.0000000.0000000.000000RBMTNoRBMT-A

Notice:

Back to top

AMFM


# Team Task Date/Time DataID AMFM
Method
Other
Resources
System
Description
juman kytea mecab moses-
tokenizer
stanford-
segmenter-
ctb
stanford-
segmenter-
pku
indic-
tokenizer
unuse myseg kmseg
1NHK-NESJIJIen-ja2019/07/24 14:23:3328860.6605700.6605700.660570-------NMTYesensemble-6, beamwidth-30 using another domain-tag
2NHK-NESJIJIen-ja2019/07/24 14:20:5128850.6542600.6542600.654260-------NMTYesensemble-6, beamwidth-30 with other resources
3sarahJIJIen-ja2019/07/22 16:44:3128140.6148400.6148400.614840-------NMTNoTransformer, ensemble of 4 models
4sarahJIJIen-ja2019/07/22 16:47:0928150.6137700.6137700.613770-------NMTNoTransformer, single model
5ORGANIZERJIJIen-ja2017/07/19 13:01:1213950.5502600.5502600.550260----0.0000000.0000000.000000SMTNoHierarchical Phrase-based SMT
6XMUNLPJIJIen-ja2017/07/25 10:33:3114430.5463600.5463600.546360----0.0000000.0000000.000000NMTNoensemble of 4 nmt models
7ORGANIZERJIJIen-ja2017/07/19 12:47:2513930.5452400.5452400.545240----0.0000000.0000000.000000SMTNoPhrase-based SMT
8NTTJIJIen-ja2017/08/01 04:17:5816790.5378000.5378000.537800----0.0000000.0000000.000000NMTNoEnsemble 8 Models: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
9NTTJIJIen-ja2017/07/30 10:41:0016030.5369700.5369700.536970----0.0000000.0000000.000000NMTNoSingle Model: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
10ORGANIZERJIJIen-ja2017/07/28 15:13:3715180.5331200.5331200.533120----0.0000000.0000000.000000NMTNoONLINE-A
11XMUNLPJIJIen-ja2017/07/24 09:11:3614290.5314700.5314700.531470----0.0000000.0000000.000000NMTNosingle nmt model
12ORGANIZERJIJIen-ja2017/07/19 13:34:4013970.5303700.5303700.530370----0.0000000.0000000.000000SMTNoTree-to-String SMT
13ORGANIZERJIJIen-ja2018/08/14 11:46:1319040.5145300.5145300.514530-----0.0000000.000000NMTNoNMT with Attention
14ORGANIZERJIJIen-ja2017/07/28 15:15:1215190.5032100.5032100.503210----0.0000000.0000000.000000NMTNoONLINE-B
15ORGANIZERJIJIen-ja2017/07/28 15:06:2315150.4873200.4873200.487320----0.0000000.0000000.000000RBMTNoRBMT-B
16ORGANIZERJIJIen-ja2017/07/28 15:04:1415140.4739400.4739400.473940----0.0000000.0000000.000000RBMTNoRBMT-A

Notice:

Back to top

HUMAN (WAT2019)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description
1NHK-NESJIJIen-ja2019/07/24 14:23:33288687.750NMTYesensemble-6, beamwidth-30 using another domain-tag
2NHK-NESJIJIen-ja2019/07/24 14:20:51288581.250NMTYesensemble-6, beamwidth-30 with other resources
3sarahJIJIen-ja2019/07/22 16:44:31281463.250NMTNoTransformer, ensemble of 4 models
4sarahJIJIen-ja2019/07/22 16:47:09281555.250NMTNoTransformer, single model

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
1ORGANIZERJIJIen-ja2017/07/28 15:13:37151869.750NMTNoONLINE-A
2ORGANIZERJIJIen-ja2017/07/28 15:04:14151431.250RBMTNoRBMT-A
3NTTJIJIen-ja2017/08/01 04:17:58167917.750NMTNoEnsemble 8 Models: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
4NTTJIJIen-ja2017/07/30 10:41:00160314.500NMTNoSingle Model: joint BPE 16k, BiLSTM Encoder 512*2*2, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
5XMUNLPJIJIen-ja2017/07/25 10:33:31144311.750NMTNoensemble of 4 nmt models
6ORGANIZERJIJIen-ja2017/07/19 13:01:12139510.250SMTNoHierarchical Phrase-based SMT

Notice:
Back to top

HUMAN (WAT2016)


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

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