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
1sakuraMMTja-en2021/05/04 15:36:486253---55.00------NMTYesText-only mBART25 finetuning
2HW-TSCMMTja-en2020/09/15 12:18:413725---52.28------NMTYestext_only_baseline_pretrained_transformer_big
3sakuraMMTja-en2021/05/04 19:42:266349---52.20------NMTNoThe ensemble of two transformer-based MNMT models with universal visual representation
4TMUMMTja-en2020/09/14 16:31:373692---48.38------NMTNoThe Ensemble of 3 baselines. Baseline: Multimodal NMT (BiGRU + ResNet50; decinit)
5TMUMMTja-en2020/09/14 18:05:093706---48.33------NMTNoThe Ensemble of 3 models. model: Multimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
6TMUMMTja-en2020/09/18 04:49:373871---47.86------NMTNoThe Ensemble of top6 models.
7TMUMMTja-en2020/09/18 19:46:043969---46.26------NMTNoMultimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
8TMUMMTja-en2020/09/15 15:57:053745---46.19------NMTNoBaseline: Multimodal NMT (BiGRU + ResNet50; decinit).
9ORGANIZERMMTja-en2020/08/26 22:41:073574---42.35------NMTNoBaseline 4: multimodal NMT (Transformer + ResNet50 w/ cross-lingual attention)
10ORGANIZERMMTja-en2020/08/26 21:56:253568---42.24------NMTNoBaseline 1: text only NMT (Transformer)
11ORGANIZERMMTja-en2020/08/26 22:39:593573---42.20------NMTNoBaseline 2: multimodal NMT (Transformer + ResNet50)
12ORGANIZERMMTja-en2020/08/26 22:42:163575---42.07------NMTNoBaseline 3: multimodal NMT (Transformer + ResNet50 w/ visual attention)
13ORGANIZERMMTja-en2020/08/26 22:02:213569---41.49------NMTNoBaseline 5: multimodal NMT (Transformer + ResNet50 w/ visual attention and cross-lingual attention)

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
1sakuraMMTja-en2021/05/04 15:36:486253---0.917155------NMTYesText-only mBART25 finetuning
2HW-TSCMMTja-en2020/09/15 12:18:413725---0.911053------NMTYestext_only_baseline_pretrained_transformer_big
3sakuraMMTja-en2021/05/04 19:42:266349---0.909991------NMTNoThe ensemble of two transformer-based MNMT models with universal visual representation
4TMUMMTja-en2020/09/14 18:05:093706---0.900684------NMTNoThe Ensemble of 3 models. model: Multimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
5TMUMMTja-en2020/09/18 04:49:373871---0.899967------NMTNoThe Ensemble of top6 models.
6TMUMMTja-en2020/09/14 16:31:373692---0.899610------NMTNoThe Ensemble of 3 baselines. Baseline: Multimodal NMT (BiGRU + ResNet50; decinit)
7TMUMMTja-en2020/09/18 19:46:043969---0.895877------NMTNoMultimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
8TMUMMTja-en2020/09/15 15:57:053745---0.895075------NMTNoBaseline: Multimodal NMT (BiGRU + ResNet50; decinit).
9ORGANIZERMMTja-en2020/08/26 22:39:593573---0.870853------NMTNoBaseline 2: multimodal NMT (Transformer + ResNet50)
10ORGANIZERMMTja-en2020/08/26 22:41:073574---0.869016------NMTNoBaseline 4: multimodal NMT (Transformer + ResNet50 w/ cross-lingual attention)
11ORGANIZERMMTja-en2020/08/26 22:02:213569---0.868410------NMTNoBaseline 5: multimodal NMT (Transformer + ResNet50 w/ visual attention and cross-lingual attention)
12ORGANIZERMMTja-en2020/08/26 22:42:163575---0.867956------NMTNoBaseline 3: multimodal NMT (Transformer + ResNet50 w/ visual attention)
13ORGANIZERMMTja-en2020/08/26 21:56:253568---0.866101------NMTNoBaseline 1: text only NMT (Transformer)

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
1ORGANIZERMMTja-en2020/08/26 22:02:213569---0.617225------NMTNoBaseline 5: multimodal NMT (Transformer + ResNet50 w/ visual attention and cross-lingual attention)
2ORGANIZERMMTja-en2020/08/26 22:42:163575---0.616765------NMTNoBaseline 3: multimodal NMT (Transformer + ResNet50 w/ visual attention)
3ORGANIZERMMTja-en2020/08/26 21:56:253568---0.616433------NMTNoBaseline 1: text only NMT (Transformer)
4ORGANIZERMMTja-en2020/08/26 22:39:593573---0.615417------NMTNoBaseline 2: multimodal NMT (Transformer + ResNet50)
5ORGANIZERMMTja-en2020/08/26 22:41:073574---0.615256------NMTNoBaseline 4: multimodal NMT (Transformer + ResNet50 w/ cross-lingual attention)
6sakuraMMTja-en2021/05/04 15:36:486253---0.579986------NMTYesText-only mBART25 finetuning
7sakuraMMTja-en2021/05/04 19:42:266349---0.577316------NMTNoThe ensemble of two transformer-based MNMT models with universal visual representation
8TMUMMTja-en2020/09/14 16:31:373692---0.000000------NMTNoThe Ensemble of 3 baselines. Baseline: Multimodal NMT (BiGRU + ResNet50; decinit)
9TMUMMTja-en2020/09/14 18:05:093706---0.000000------NMTNoThe Ensemble of 3 models. model: Multimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
10HW-TSCMMTja-en2020/09/15 12:18:413725---0.000000------NMTYestext_only_baseline_pretrained_transformer_big
11TMUMMTja-en2020/09/15 15:57:053745---0.000000------NMTNoBaseline: Multimodal NMT (BiGRU + ResNet50; decinit).
12TMUMMTja-en2020/09/18 04:49:373871---0.000000------NMTNoThe Ensemble of top6 models.
13TMUMMTja-en2020/09/18 19:46:043969---0.000000------NMTNoMultimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.

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
1sakuraMMTja-en2021/05/04 19:42:266349UnderwayNMTNoThe ensemble of two transformer-based MNMT models with universal visual representation

Notice:
Back to top

HUMAN (WAT2020)


# Team Task Date/Time DataID HUMAN
Method
Other
Resources
System
Description
1HW-TSCMMTja-en2020/09/15 12:18:4137254.798NMTYestext_only_baseline_pretrained_transformer_big
2TMUMMTja-en2020/09/14 18:05:0937064.375NMTNoThe Ensemble of 3 models. model: Multimodal NMT (BiGRU + ResNet50; double attention) w/ DropNet.
3TMUMMTja-en2020/09/14 16:31:3736924.317NMTNoThe Ensemble of 3 baselines. Baseline: Multimodal NMT (BiGRU + ResNet50; decinit)

Notice:
Back to top

HUMAN (WAT2019)


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

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

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