WAT

The Workshop on Asian Translation

Evaluation Results

BLEU

#	Team	Task	Date/Time	DataID	BLEU										Method	Other Resources	System Description
#	Team	Task	Date/Time	DataID	juman	kytea	mecab	moses- tokenizer	stanford- segmenter- ctb	stanford- segmenter- pku	indic- tokenizer	unuse	myseg	kmseg	Method	Other Resources	System Description
1	NICT	ALT2my-en	2019/07/22 17:06:59	2816	-	-	-	30.15	-	-	-	-	-	-	NMT	Yes	Single model+language model pre-training+back-translation
2	ORGANIZER	ALT2my-en	2019/07/22 19:08:03	2826	-	-	-	14.85	-	-	-	-	-	-	NMT	No	NMT with Attention
3	NICT	ALT2my-en	2019/07/23 13:40:55	2854	-	-	-	18.51	-	-	-	-	-	-	NMT	No	Single model+language model pre-training
4	ORGANIZER	ALT2my-en	2019/07/24 16:17:22	2899	-	-	-	14.59	-	-	-	-	-	-	Other	Yes	Online A
5	NICT-4	ALT2my-en	2019/07/26 11:31:52	2977	-	-	-	24.75	-	-	-	-	-	-	Other	Yes	Same as last year but with cleaner monolingual data
6	UCSMNLP	ALT2my-en	2019/07/26 17:01:16	3022	-	-	-	10.70	-	-	-	-	-	-	SMT	Yes
7	FBAI	ALT2my-en	2019/07/27 08:05:36	3148	-	-	-	26.75	-	-	-	-	-	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
8	FBAI	ALT2my-en	2019/07/27 14:36:58	3201	-	-	-	38.59	-	-	-	-	-	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
9	UCSYNLP	ALT2my-en	2019/07/29 13:51:40	3252	-	-	-	19.64	-	-	-	-	-	-	NMT	No	NMT with Attention
10	sakura	ALT2my-en	2021/04/20 23:35:56	5230	-	-	-	19.75	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
11	sakura	ALT2my-en	2021/05/03 11:22:26	5990	-	-	-	18.70	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
12	YCC-MT2	ALT2my-en	2021/05/04 08:44:59	6181	-	-	-	8.04	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.6 0.4 result (i.e. best BLEU score), Myanmar-Syllable to English Translation
13	NECTEC	ALT2my-en	2021/05/04 09:01:27	6188	-	-	-	6.24	-	-	-	-	-	-	NMT	No	Transformer pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
14	NECTEC	ALT2my-en	2021/05/04 09:31:57	6192	-	-	-	4.62	-	-	-	-	-	-	NMT	No	Shared-multi-source Transformer, archi: Transformer, source1:string, source2:POS, target:string, used 2 GPUs
15	YCC-MT2	ALT2my-en	2021/05/06 20:47:22	6398	-	-	-	7.92	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), Myanmar Word to English Translation, used in-house myWord Segmenter (release soon)
16	NECTEC	ALT2my-en	2021/05/22 19:29:10	6400	-	-	-	6.72	-	-	-	-	-	-	NMT	No	s2s or RNN-based Attention, pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
17	NECTEC	ALT2my-en	2021/05/22 19:37:14	6401	-	-	-	4.73	-	-	-	-	-	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
18	NECTEC	ALT2my-en	2021/05/22 19:45:14	6402	-	-	-	6.13	-	-	-	-	-	-	NMT	No	Shared-multi-s2s, Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
19	YCC-MT2	ALT2my-en	2021/06/10 15:56:14	6439	-	-	-	11.86	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, syl-word
20	YCC-MT2	ALT2my-en	2021/06/10 16:03:28	6440	-	-	-	10.80	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, syl-word
21	YCC-MT2	ALT2my-en	2021/06/10 16:15:48	6443	-	-	-	11.50	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, word-word
22	YCC-MT2	ALT2my-en	2021/06/10 16:20:49	6444	-	-	-	10.37	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, word-word
23	YCC-MT2	ALT2my-en	2021/06/10 16:42:51	6448	-	-	-	12.48	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, syl-word
24	YCC-MT2	ALT2my-en	2021/06/10 16:48:01	6449	-	-	-	12.72	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.5 0.5
25	YCC-MT2	ALT2my-en	2021/06/10 16:51:57	6450	-	-	-	12.85	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.6 0.4
26	YCC-MT2	ALT2my-en	2021/06/10 17:10:38	6455	-	-	-	12.77	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word
27	YCC-MT2	ALT2my-en	2021/06/10 17:15:49	6456	-	-	-	12.80	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.5, 0.5
28	YCC-MT2	ALT2my-en	2021/06/10 17:29:43	6457	-	-	-	13.01	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.6, 0.4

Notice:

This table is sorted by the leftmost segmenters. You can change the segmenter used to sort by clicking each segmenter link.

RIBES

#	Team	Task	Date/Time	DataID	RIBES										Method	Other Resources	System Description
#	Team	Task	Date/Time	DataID	juman	kytea	mecab	moses- tokenizer	stanford- segmenter- ctb	stanford- segmenter- pku	indic- tokenizer	unuse	myseg	kmseg	Method	Other Resources	System Description
1	NICT	ALT2my-en	2019/07/22 17:06:59	2816	-	-	-	0.791705	-	-	-	-	-	-	NMT	Yes	Single model+language model pre-training+back-translation
2	ORGANIZER	ALT2my-en	2019/07/22 19:08:03	2826	-	-	-	0.700166	-	-	-	-	-	-	NMT	No	NMT with Attention
3	NICT	ALT2my-en	2019/07/23 13:40:55	2854	-	-	-	0.744808	-	-	-	-	-	-	NMT	No	Single model+language model pre-training
4	ORGANIZER	ALT2my-en	2019/07/24 16:17:22	2899	-	-	-	0.602267	-	-	-	-	-	-	Other	Yes	Online A
5	NICT-4	ALT2my-en	2019/07/26 11:31:52	2977	-	-	-	0.760394	-	-	-	-	-	-	Other	Yes	Same as last year but with cleaner monolingual data
6	UCSMNLP	ALT2my-en	2019/07/26 17:01:16	3022	-	-	-	0.570835	-	-	-	-	-	-	SMT	Yes
7	FBAI	ALT2my-en	2019/07/27 08:05:36	3148	-	-	-	0.783571	-	-	-	-	-	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
8	FBAI	ALT2my-en	2019/07/27 14:36:58	3201	-	-	-	0.840001	-	-	-	-	-	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
9	UCSYNLP	ALT2my-en	2019/07/29 13:51:40	3252	-	-	-	0.707789	-	-	-	-	-	-	NMT	No	NMT with Attention
10	sakura	ALT2my-en	2021/04/20 23:35:56	5230	-	-	-	0.742698	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
11	sakura	ALT2my-en	2021/05/03 11:22:26	5990	-	-	-	0.736523	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
12	YCC-MT2	ALT2my-en	2021/05/04 08:44:59	6181	-	-	-	0.630357	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.6 0.4 result (i.e. best BLEU score), Myanmar-Syllable to English Translation
13	NECTEC	ALT2my-en	2021/05/04 09:01:27	6188	-	-	-	0.620840	-	-	-	-	-	-	NMT	No	Transformer pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
14	NECTEC	ALT2my-en	2021/05/04 09:31:57	6192	-	-	-	0.587155	-	-	-	-	-	-	NMT	No	Shared-multi-source Transformer, archi: Transformer, source1:string, source2:POS, target:string, used 2 GPUs
15	YCC-MT2	ALT2my-en	2021/05/06 20:47:22	6398	-	-	-	0.629755	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), Myanmar Word to English Translation, used in-house myWord Segmenter (release soon)
16	NECTEC	ALT2my-en	2021/05/22 19:29:10	6400	-	-	-	0.616469	-	-	-	-	-	-	NMT	No	s2s or RNN-based Attention, pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
17	NECTEC	ALT2my-en	2021/05/22 19:37:14	6401	-	-	-	0.578146	-	-	-	-	-	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
18	NECTEC	ALT2my-en	2021/05/22 19:45:14	6402	-	-	-	0.609560	-	-	-	-	-	-	NMT	No	Shared-multi-s2s, Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
19	YCC-MT2	ALT2my-en	2021/06/10 15:56:14	6439	-	-	-	0.673532	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, syl-word
20	YCC-MT2	ALT2my-en	2021/06/10 16:03:28	6440	-	-	-	0.673755	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, syl-word
21	YCC-MT2	ALT2my-en	2021/06/10 16:15:48	6443	-	-	-	0.670478	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, word-word
22	YCC-MT2	ALT2my-en	2021/06/10 16:20:49	6444	-	-	-	0.664105	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, word-word
23	YCC-MT2	ALT2my-en	2021/06/10 16:42:51	6448	-	-	-	0.692376	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, syl-word
24	YCC-MT2	ALT2my-en	2021/06/10 16:48:01	6449	-	-	-	0.691281	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.5 0.5
25	YCC-MT2	ALT2my-en	2021/06/10 16:51:57	6450	-	-	-	0.689418	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.6 0.4
26	YCC-MT2	ALT2my-en	2021/06/10 17:10:38	6455	-	-	-	0.685502	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word
27	YCC-MT2	ALT2my-en	2021/06/10 17:15:49	6456	-	-	-	0.688797	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.5, 0.5
28	YCC-MT2	ALT2my-en	2021/06/10 17:29:43	6457	-	-	-	0.686109	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.6, 0.4

Notice:

This table is sorted by the leftmost segmenters. You can change the segmenter used to sort by clicking each segmenter link.

AMFM

#	Team	Task	Date/Time	DataID	AMFM										Method	Other Resources	System Description
#	Team	Task	Date/Time	DataID	unuse	unuse	unuse	unuse	unuse	unuse	unuse	unuse	unuse	unuse	Method	Other Resources	System Description
1	NICT	ALT2my-en	2019/07/22 17:06:59	2816	-	-	-	0.649050	-	-	-	-	-	-	NMT	Yes	Single model+language model pre-training+back-translation
2	ORGANIZER	ALT2my-en	2019/07/22 19:08:03	2826	-	-	-	0.467460	-	-	-	-	-	-	NMT	No	NMT with Attention
3	NICT	ALT2my-en	2019/07/23 13:40:55	2854	-	-	-	0.565430	-	-	-	-	-	-	NMT	No	Single model+language model pre-training
4	ORGANIZER	ALT2my-en	2019/07/24 16:17:22	2899	-	-	-	0.549380	-	-	-	-	-	-	Other	Yes	Online A
5	NICT-4	ALT2my-en	2019/07/26 11:31:52	2977	-	-	-	0.579570	-	-	-	-	-	-	Other	Yes	Same as last year but with cleaner monolingual data
6	UCSMNLP	ALT2my-en	2019/07/26 17:01:16	3022	-	-	-	0.538280	-	-	-	-	-	-	SMT	Yes
7	FBAI	ALT2my-en	2019/07/27 08:05:36	3148	-	-	-	0.627530	-	-	-	-	-	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
8	FBAI	ALT2my-en	2019/07/27 14:36:58	3201	-	-	-	0.685200	-	-	-	-	-	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
9	UCSYNLP	ALT2my-en	2019/07/29 13:51:40	3252	-	-	-	0.532640	-	-	-	-	-	-	NMT	No	NMT with Attention
10	sakura	ALT2my-en	2021/04/20 23:35:56	5230	-	-	-	0.562680	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
11	sakura	ALT2my-en	2021/05/03 11:22:26	5990	-	-	-	0.550430	-	-	-	-	-	-	NMT	No	Marian NMT with Attention
12	YCC-MT2	ALT2my-en	2021/05/04 08:44:59	6181	-	-	-	0.407880	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.6 0.4 result (i.e. best BLEU score), Myanmar-Syllable to English Translation
13	NECTEC	ALT2my-en	2021/05/04 09:01:27	6188	-	-	-	0.424640	-	-	-	-	-	-	NMT	No	Transformer pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
14	NECTEC	ALT2my-en	2021/05/04 09:31:57	6192	-	-	-	0.391710	-	-	-	-	-	-	NMT	No	Shared-multi-source Transformer, archi: Transformer, source1:string, source2:POS, target:string, used 2 GPUs
15	YCC-MT2	ALT2my-en	2021/05/06 20:47:22	6398	-	-	-	0.420200	-	-	-	-	-	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), Myanmar Word to English Translation, used in-house myWord Segmenter (release soon)
16	NECTEC	ALT2my-en	2021/05/22 19:29:10	6400	-	-	-	0.395310	-	-	-	-	-	-	NMT	No	s2s or RNN-based Attention, pos2string Myanmar-English, used myPOS ver 2.0 POS Tagger, used 2 GPUs
17	NECTEC	ALT2my-en	2021/05/22 19:37:14	6401	-	-	-	0.357150	-	-	-	-	-	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
18	NECTEC	ALT2my-en	2021/05/22 19:45:14	6402	-	-	-	0.376140	-	-	-	-	-	-	NMT	No	Shared-multi-s2s, Archi: s2s, source1:string, source2:pos, target:string, used 2 GPUs
19	YCC-MT2	ALT2my-en	2021/06/10 15:56:14	6439	-	-	-	0.430120	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, syl-word
20	YCC-MT2	ALT2my-en	2021/06/10 16:03:28	6440	-	-	-	0.462440	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, syl-word
21	YCC-MT2	ALT2my-en	2021/06/10 16:15:48	6443	-	-	-	0.425310	-	-	-	-	-	-	NMT	No	UCSY+ALT, s2s, my-en, word-word
22	YCC-MT2	ALT2my-en	2021/06/10 16:20:49	6444	-	-	-	0.461980	-	-	-	-	-	-	NMT	No	UCSY+ALT, transformer, my-en, word-word
23	YCC-MT2	ALT2my-en	2021/06/10 16:42:51	6448	-	-	-	0.437760	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, syl-word
24	YCC-MT2	ALT2my-en	2021/06/10 16:48:01	6449	-	-	-	0.438520	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.5 0.5
25	YCC-MT2	ALT2my-en	2021/06/10 16:51:57	6450	-	-	-	0.434960	-	-	-	-	-	-	NMT	No	UCSY+ALt, s2s-plus-transformer, my-en, syl-word, weight: 0.6 0.4
26	YCC-MT2	ALT2my-en	2021/06/10 17:10:38	6455	-	-	-	0.439350	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word
27	YCC-MT2	ALT2my-en	2021/06/10 17:15:49	6456	-	-	-	0.437300	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.5, 0.5
28	YCC-MT2	ALT2my-en	2021/06/10 17:29:43	6457	-	-	-	0.432530	-	-	-	-	-	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, my-en, word-word, weight: 0.6, 0.4

Notice:

This table is sorted by the leftmost segmenters. You can change the segmenter used to sort by clicking each segmenter link.
Adequacy-Fluency Metrics (AMFM) is a two-dimensional automatic evaluation metric for machine translation, designed to operate at the sentence level. It is based on adequacy and fluency, to decouple semantic and syntactic components of the translation process to provide a balanced view on translation quality.
AMFM is calculated without tokenizers.
The detail of AMFM is shown on the following paper: "Adequacy–Fluency Metrics: Evaluating MT in the Continuous Space Model Framework" [pdf]. Invited Talk in WAT2015 also helps understanding [slide].

HUMAN (WAT2022)

Notice:

HUMAN (WAT2022) is the result of the Pairwise Crowdsourcing Evaluation on WAT2022.
HUMAN (WAT2022) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2021)

Notice:

HUMAN (WAT2021) is the result of the Pairwise Crowdsourcing Evaluation on WAT2021.
HUMAN (WAT2021) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2020)

Notice:

HUMAN (WAT2020) is the result of the Pairwise Crowdsourcing Evaluation on WAT2020.
HUMAN (WAT2020) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2019)

Notice:

HUMAN (WAT2019) is the result of the Pairwise Crowdsourcing Evaluation on WAT2019.
HUMAN (WAT2019) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2018)

Notice:

HUMAN (WAT2018) is the result of the Pairwise Crowdsourcing Evaluation on WAT2018.
HUMAN (WAT2018) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2017)

Notice:

HUMAN (WAT2017) is the result of the Pairwise Crowdsourcing Evaluation on WAT2017.
HUMAN (WAT2017) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2016)

Notice:

HUMAN (WAT2016) is the result of the Pairwise Crowdsourcing Evaluation on WAT2016.
HUMAN (WAT2016) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.

HUMAN (WAT2015)

Notice:

HUMAN (WAT2015) is the result of the Pairwise Crowdsourcing Evaluation on WAT2015.
HUMAN (WAT2015) was evaluated by 5 different workers and the final decision is made by the voting of the judgements.
The detail of the evaluation can be found in the pdf document (PDF file).

HUMAN (WAT2014)

Notice:

HUMAN (WAT2014) is the result of the Pairwise Crowdsourcing Evaluation on WAT2014.
HUMAN (WAT2014) was evaluated by 3 different workers and the final decision is made by the voting of the judgements.
The detail of the evaluation can be found in the pdf document (PDF file).

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

WAT The Workshop on Asian Translation Evaluation Results

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

WAT

The Workshop on Asian Translation

Evaluation Results