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
1CNLP-NITS-PPMMCHMM24en-bn2022/07/11 12:54:096744------28.70---NMTNoTransliteration-based phrase pairs augmentation and visual features in training using BRNN encoder and doubly-attentive-rnn decoder.
2SILO_NLPMMCHMM24en-bn2022/07/14 21:41:046940------28.70---NMTNoObject Tags (Image) + Finetune mBART
3ODIAGENMMCHMM24en-bn2023/07/06 04:05:407108------30.50---NMTNoImage features extracted as Object tags appended with text and MBART fine-tuning
4BITS-PMMCHMM24en-bn2023/07/08 13:38:187122------48.70---NMTYesNLLB model finetuned on captions + object tags of original & synthetic images using DETR model
51128MMCHMM24en-bn2024/07/23 15:47:137138------30.50---NMTNoinitial model
600-7MMCHMM24en-bn2024/08/05 15:16:587192------45.30---NMTYesTest
7v036MMCHMM24en-bn2024/08/11 12:35:307318------20.70---NMTNoNMT based system using both image descriptors and text description. A multistage LLM pipeline used for extracting image data descriptions and translation. Fine tuning done in few cases Models Used:
8239233MMCHMM24en-bn2024/08/13 05:00:557375------21.70---NMTYesOne-shot prompt for synthetic QA description from captions; translate QA using IndicTrans2; generate caption from QA as context
9v036MMCHMM24en-bn2024/08/15 16:27:537401------ 6.10---NMTNo
10v036MMCHMM24en-bn2024/08/15 19:41:507414------33.90---NMTNo
11IITP-AI-NLP-MLMMCHMM24en-bn2025/10/22 21:56:317462------47.00---NMTYesUsed Selective Attention Architecture with IndicTrans as the base model and CLIP ViT-B/16 model to extract image features. We extract a) Full image feats, and b) cropped image feats and pick the one w

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
1CNLP-NITS-PPMMCHMM24en-bn2022/07/11 12:54:096744------0.688931---NMTNoTransliteration-based phrase pairs augmentation and visual features in training using BRNN encoder and doubly-attentive-rnn decoder.
2SILO_NLPMMCHMM24en-bn2022/07/14 21:41:046940------0.666817---NMTNoObject Tags (Image) + Finetune mBART
3ODIAGENMMCHMM24en-bn2023/07/06 04:05:407108------0.690706---NMTNoImage features extracted as Object tags appended with text and MBART fine-tuning
4BITS-PMMCHMM24en-bn2023/07/08 13:38:187122------0.831946---NMTYesNLLB model finetuned on captions + object tags of original & synthetic images using DETR model
51128MMCHMM24en-bn2024/07/23 15:47:137138------0.701175---NMTNoinitial model
600-7MMCHMM24en-bn2024/08/05 15:16:587192------0.796451---NMTYesTest
7v036MMCHMM24en-bn2024/08/11 12:35:307318------0.625485---NMTNoNMT based system using both image descriptors and text description. A multistage LLM pipeline used for extracting image data descriptions and translation. Fine tuning done in few cases Models Used:
8239233MMCHMM24en-bn2024/08/13 05:00:557375------0.644341---NMTYesOne-shot prompt for synthetic QA description from captions; translate QA using IndicTrans2; generate caption from QA as context
9v036MMCHMM24en-bn2024/08/15 16:27:537401------0.558068---NMTNo
10v036MMCHMM24en-bn2024/08/15 19:41:507414------0.736029---NMTNo
11IITP-AI-NLP-MLMMCHMM24en-bn2025/10/22 21:56:317462------0.815367---NMTYesUsed Selective Attention Architecture with IndicTrans as the base model and CLIP ViT-B/16 model to extract image features. We extract a) Full image feats, and b) cropped image feats and pick the one w

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
1CNLP-NITS-PPMMCHMM24en-bn2022/07/11 12:54:096744------0.000000---NMTNoTransliteration-based phrase pairs augmentation and visual features in training using BRNN encoder and doubly-attentive-rnn decoder.
2SILO_NLPMMCHMM24en-bn2022/07/14 21:41:046940------0.000000---NMTNoObject Tags (Image) + Finetune mBART
3ODIAGENMMCHMM24en-bn2023/07/06 04:05:407108------0.000000---NMTNoImage features extracted as Object tags appended with text and MBART fine-tuning
4BITS-PMMCHMM24en-bn2023/07/08 13:38:187122------0.000000---NMTYesNLLB model finetuned on captions + object tags of original & synthetic images using DETR model
51128MMCHMM24en-bn2024/07/23 15:47:137138------0.000000---NMTNoinitial model
600-7MMCHMM24en-bn2024/08/05 15:16:587192------0.000000---NMTYesTest
7v036MMCHMM24en-bn2024/08/11 12:35:307318------0.000000---NMTNoNMT based system using both image descriptors and text description. A multistage LLM pipeline used for extracting image data descriptions and translation. Fine tuning done in few cases Models Used:
8239233MMCHMM24en-bn2024/08/13 05:00:557375------0.000000---NMTYesOne-shot prompt for synthetic QA description from captions; translate QA using IndicTrans2; generate caption from QA as context
9v036MMCHMM24en-bn2024/08/15 16:27:537401------0.000000---NMTNo
10v036MMCHMM24en-bn2024/08/15 19:41:507414------0.000000---NMTNo
11IITP-AI-NLP-MLMMCHMM24en-bn2025/10/22 21:56:317462------0.000000---NMTYesUsed Selective Attention Architecture with IndicTrans as the base model and CLIP ViT-B/16 model to extract image features. We extract a) Full image feats, and b) cropped image feats and pick the one w

Notice:

Back to top

HUMAN (WAT2022)


# Team Task Date/Time DataID HUMAN
Method
 Other
Resources
System
Description
1CNLP-NITS-PPMMCHMM24en-bn2022/07/11 12:54:096744UnderwayNMTNoTransliteration-based phrase pairs augmentation and visual features in training using BRNN encoder and doubly-attentive-rnn decoder.
2SILO_NLPMMCHMM24en-bn2022/07/14 21:41:046940UnderwayNMTNoObject Tags (Image) + Finetune mBART

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

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