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	ORGANIZER	JIJIen-ja	2018/08/14 11:46:13	1904	16.63	17.01	16.36	-	-	-	-	-	0.00	0.00	NMT	No	NMT with Attention
2	sarah	JIJIen-ja	2019/07/22 16:44:31	2814	22.65	23.48	22.76	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
3	sarah	JIJIen-ja	2019/07/22 16:47:09	2815	21.80	22.67	21.91	-	-	-	-	-	-	-	NMT	No	Transformer, single model
4	NHK-NES	JIJIen-ja	2019/07/24 14:20:51	2885	28.25	29.76	28.63	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 with other resources
5	NHK-NES	JIJIen-ja	2019/07/24 14:23:33	2886	27.22	28.75	27.63	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 using another domain-tag
6	ORGANIZER	JIJIen-ja	2017/07/19 12:47:25	1393	15.77	16.65	15.76	-	-	-	-	0.00	0.00	0.00	SMT	No	Phrase-based SMT
7	ORGANIZER	JIJIen-ja	2017/07/19 13:01:12	1395	16.22	16.95	16.22	-	-	-	-	0.00	0.00	0.00	SMT	No	Hierarchical Phrase-based SMT
8	ORGANIZER	JIJIen-ja	2017/07/19 13:34:40	1397	14.95	15.38	14.79	-	-	-	-	0.00	0.00	0.00	SMT	No	Tree-to-String SMT
9	XMUNLP	JIJIen-ja	2017/07/24 09:11:36	1429	17.48	18.57	17.92	-	-	-	-	0.00	0.00	0.00	NMT	No	single nmt model
10	XMUNLP	JIJIen-ja	2017/07/25 10:33:31	1443	19.61	20.72	20.14	-	-	-	-	0.00	0.00	0.00	NMT	No	ensemble of 4 nmt models
11	ORGANIZER	JIJIen-ja	2017/07/28 15:04:14	1514	5.31	6.68	5.69	-	-	-	-	0.00	0.00	0.00	RBMT	No	RBMT-A
12	ORGANIZER	JIJIen-ja	2017/07/28 15:06:23	1515	4.72	5.98	4.97	-	-	-	-	0.00	0.00	0.00	RBMT	No	RBMT-B
13	ORGANIZER	JIJIen-ja	2017/07/28 15:13:37	1518	11.29	13.12	11.84	-	-	-	-	0.00	0.00	0.00	NMT	No	ONLINE-A
14	ORGANIZER	JIJIen-ja	2017/07/28 15:15:12	1519	7.46	8.73	7.63	-	-	-	-	0.00	0.00	0.00	NMT	No	ONLINE-B
15	NTT	JIJIen-ja	2017/07/30 10:41:00	1603	19.13	20.47	19.41	-	-	-	-	0.00	0.00	0.00	NMT	No	Single Model: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
16	NTT	JIJIen-ja	2017/08/01 04:17:58	1679	20.37	21.82	20.68	-	-	-	-	0.00	0.00	0.00	NMT	No	Ensemble 8 Models: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking

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	ORGANIZER	JIJIen-ja	2018/08/14 11:46:13	1904	0.653766	0.657279	0.658830	-	-	-	-	-	0.000000	0.000000	NMT	No	NMT with Attention
2	sarah	JIJIen-ja	2019/07/22 16:44:31	2814	0.705044	0.707209	0.710111	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
3	sarah	JIJIen-ja	2019/07/22 16:47:09	2815	0.698609	0.699981	0.701739	-	-	-	-	-	-	-	NMT	No	Transformer, single model
4	NHK-NES	JIJIen-ja	2019/07/24 14:20:51	2885	0.739036	0.741164	0.743977	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 with other resources
5	NHK-NES	JIJIen-ja	2019/07/24 14:23:33	2886	0.730816	0.731653	0.734618	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 using another domain-tag
6	ORGANIZER	JIJIen-ja	2017/07/19 12:47:25	1393	0.580284	0.584892	0.585437	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT
7	ORGANIZER	JIJIen-ja	2017/07/19 13:01:12	1395	0.594923	0.601505	0.602516	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Hierarchical Phrase-based SMT
8	ORGANIZER	JIJIen-ja	2017/07/19 13:34:40	1397	0.594072	0.597791	0.599530	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Tree-to-String SMT
9	XMUNLP	JIJIen-ja	2017/07/24 09:11:36	1429	0.659372	0.665578	0.668759	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	single nmt model
10	XMUNLP	JIJIen-ja	2017/07/25 10:33:31	1443	0.684120	0.688497	0.691056	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ensemble of 4 nmt models
11	ORGANIZER	JIJIen-ja	2017/07/28 15:04:14	1514	0.505227	0.515050	0.513580	-	-	-	-	0.000000	0.000000	0.000000	RBMT	No	RBMT-A
12	ORGANIZER	JIJIen-ja	2017/07/28 15:06:23	1515	0.518416	0.531603	0.532079	-	-	-	-	0.000000	0.000000	0.000000	RBMT	No	RBMT-B
13	ORGANIZER	JIJIen-ja	2017/07/28 15:13:37	1518	0.597473	0.605532	0.603374	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ONLINE-A
14	ORGANIZER	JIJIen-ja	2017/07/28 15:15:12	1519	0.540558	0.545430	0.545178	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ONLINE-B
15	NTT	JIJIen-ja	2017/07/30 10:41:00	1603	0.668517	0.670920	0.676594	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Single Model: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
16	NTT	JIJIen-ja	2017/08/01 04:17:58	1679	0.680598	0.684048	0.688863	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Ensemble 8 Models: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking

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	ORGANIZER	JIJIen-ja	2018/08/14 11:46:13	1904	0.514530	0.514530	0.514530	-	-	-	-	-	0.000000	0.000000	NMT	No	NMT with Attention
2	sarah	JIJIen-ja	2019/07/22 16:44:31	2814	0.614840	0.614840	0.614840	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
3	sarah	JIJIen-ja	2019/07/22 16:47:09	2815	0.613770	0.613770	0.613770	-	-	-	-	-	-	-	NMT	No	Transformer, single model
4	NHK-NES	JIJIen-ja	2019/07/24 14:20:51	2885	0.654260	0.654260	0.654260	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 with other resources
5	NHK-NES	JIJIen-ja	2019/07/24 14:23:33	2886	0.660570	0.660570	0.660570	-	-	-	-	-	-	-	NMT	Yes	ensemble-6, beamwidth-30 using another domain-tag
6	ORGANIZER	JIJIen-ja	2017/07/19 12:47:25	1393	0.545240	0.545240	0.545240	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT
7	ORGANIZER	JIJIen-ja	2017/07/19 13:01:12	1395	0.550260	0.550260	0.550260	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Hierarchical Phrase-based SMT
8	ORGANIZER	JIJIen-ja	2017/07/19 13:34:40	1397	0.530370	0.530370	0.530370	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Tree-to-String SMT
9	XMUNLP	JIJIen-ja	2017/07/24 09:11:36	1429	0.531470	0.531470	0.531470	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	single nmt model
10	XMUNLP	JIJIen-ja	2017/07/25 10:33:31	1443	0.546360	0.546360	0.546360	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ensemble of 4 nmt models
11	ORGANIZER	JIJIen-ja	2017/07/28 15:04:14	1514	0.473940	0.473940	0.473940	-	-	-	-	0.000000	0.000000	0.000000	RBMT	No	RBMT-A
12	ORGANIZER	JIJIen-ja	2017/07/28 15:06:23	1515	0.487320	0.487320	0.487320	-	-	-	-	0.000000	0.000000	0.000000	RBMT	No	RBMT-B
13	ORGANIZER	JIJIen-ja	2017/07/28 15:13:37	1518	0.533120	0.533120	0.533120	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ONLINE-A
14	ORGANIZER	JIJIen-ja	2017/07/28 15:15:12	1519	0.503210	0.503210	0.503210	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	ONLINE-B
15	NTT	JIJIen-ja	2017/07/30 10:41:00	1603	0.536970	0.536970	0.536970	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Single Model: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking
16	NTT	JIJIen-ja	2017/08/01 04:17:58	1679	0.537800	0.537800	0.537800	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Ensemble 8 Models: joint BPE 16k, BiLSTM Encoder 51222, LtoR LSTM Decoder 512*2, Beam Search 20 w/ length-based reranking

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