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	JPCen-ja	2016/07/13 16:47:25	973	32.36	34.26	32.52	-	-	-	-	0.00	0.00	0.00	SMT	No	Phrase-based SMT
2	ORGANIZER	JPCen-ja	2016/07/13 16:50:03	974	34.57	36.61	34.79	-	-	-	-	0.00	0.00	0.00	SMT	No	Hierarchical Phrase-based SMT
3	ORGANIZER	JPCen-ja	2016/07/13 16:52:07	975	35.60	37.65	35.82	-	-	-	-	0.00	0.00	0.00	SMT	No	Tree-to-String SMT
4	Kyoto-U	JPCen-ja	2016/07/13 17:38:24	986	36.04	38.14	36.30	-	-	-	-	0.00	0.00	0.00	EBMT	No	Baseline
5	ORGANIZER	JPCen-ja	2016/07/26 10:20:53	1036	36.88	37.89	36.83	-	-	-	-	0.00	0.00	0.00	Other	Yes	Online A (2016)
6	ORGANIZER	JPCen-ja	2016/08/02 10:26:14	1073	21.57	22.62	21.65	-	-	-	-	0.00	0.00	0.00	Other	Yes	Online B (2016)
7	NICT-2	JPCen-ja	2016/08/04 17:23:27	1078	39.03	40.74	38.98	-	-	-	-	0.00	0.00	0.00	SMT	No	Phrase-based SMT with Preordering + Domain Adaptation
8	ORGANIZER	JPCen-ja	2016/08/05 14:36:45	1085	23.02	24.90	23.45	-	-	-	-	0.00	0.00	0.00	Other	Yes	RBMT D (2016)
9	ORGANIZER	JPCen-ja	2016/08/05 14:38:28	1086	26.64	28.48	26.84	-	-	-	-	0.00	0.00	0.00	Other	Yes	RBMT F (2016)
10	ORGANIZER	JPCen-ja	2016/08/05 14:41:32	1087	21.35	23.17	21.53	-	-	-	-	0.00	0.00	0.00	Other	Yes	RBMT E (2016)
11	NICT-2	JPCen-ja	2016/08/05 17:56:31	1098	40.90	42.51	40.66	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC) + Google 5-gram LM
12	bjtu_nlp	JPCen-ja	2016/08/08 12:26:18	1112	39.46	41.16	39.45	-	-	-	-	0.00	0.00	0.00	NMT	No	RNN Encoder-Decoder with attention mechanism, single model
13	JAPIO	JPCen-ja	2016/08/15 15:56:40	1141	45.57	46.40	45.74	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus
14	JAPIO	JPCen-ja	2016/08/17 11:36:23	1156	47.79	48.57	47.92	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + JPC/JAPIO corpora
15	JAPIO	JPCen-ja	2016/08/17 11:39:13	1157	50.28	51.08	50.53	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus including some sentences in testset
16	JAPIO	JPCen-ja	2016/10/27 13:05:35	1330	38.59	40.00	38.65	-	-	-	-	0.00	0.00	0.00	SMT	No	Phrase-based SMT with Preordering
17	EHR	JPCen-ja	2017/07/19 18:48:50	1406	44.44	45.59	44.15	-	-	-	-	0.00	0.00	0.00	NMT	No	SMT reranked NMT (sub word based, by Moses and OpenNMT)
18	EHR	JPCen-ja	2017/07/19 19:08:34	1407	44.63	45.94	44.53	-	-	-	-	0.00	0.00	0.00	NMT	No	Simple NMT (sub word based, by OpenNMT)
19	JAPIO	JPCen-ja	2017/07/25 12:13:33	1445	55.55	56.03	55.40	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus including some sentences in testset
20	JAPIO	JPCen-ja	2017/07/25 18:02:12	1451	44.69	46.01	44.53	-	-	-	-	0.00	0.00	0.00	NMT	No	OpenNMT(dbrnn)
21	JAPIO	JPCen-ja	2017/07/25 18:10:48	1453	48.39	49.34	48.29	-	-	-	-	0.00	0.00	0.00	NMT	Yes	OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor
22	JAPIO	JPCen-ja	2017/07/25 18:14:57	1454	50.27	51.23	50.17	-	-	-	-	0.00	0.00	0.00	NMT	Yes	Combination of 4 NMT systems (OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor)
23	JAPIO	JPCen-ja	2017/07/26 10:16:13	1462	51.79	52.23	51.75	-	-	-	-	0.00	0.00	0.00	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus
24	u-tkb	JPCen-ja	2017/07/26 12:46:57	1470	38.91	41.12	39.11	-	-	-	-	0.00	0.00	0.00	NMT	No	NMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
25	ORGANIZER	JPCen-ja	2018/08/15 18:35:10	1964	43.84	45.28	43.70	-	-	-	-	-	0.00	0.00	NMT	No	NMT with Attention
26	EHR	JPCen-ja	2018/09/08 20:31:56	2248	47.36	48.66	47.22	-	-	-	-	-	0.00	0.00	NMT	No	SMT reranked NMT
27	EHR	JPCen-ja	2018/09/13 12:50:53	2284	47.57	48.68	47.40	-	-	-	-	-	0.00	0.00	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 13)
28	EHR	JPCen-ja	2018/09/15 15:50:14	2395	48.01	49.14	47.83	-	-	-	-	-	0.00	0.00	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 18)
29	EHR	JPCen-ja	2018/09/16 15:09:19	2476	48.03	49.24	47.86	-	-	-	-	-	0.00	0.00	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 20)
30	sarah	JPCen-ja	2019/07/26 11:26:05	2973	47.67	49.06	47.51	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
31	KNU_Hyundai	JPCen-ja	2019/07/27 12:06:18	3185	49.04	50.43	49.08	-	-	-	-	-	-	-	NMT	No	Transformer Base, relative position, BT, r2l reranking, checkpoint ensemble
32	KNU_Hyundai	JPCen-ja	2019/07/27 12:51:20	3193	48.99	50.17	48.96	-	-	-	-	-	-	-	NMT	Yes	Transformer Base (+ASPEC data), relative position, BT, r2l reranking, checkpoint ensemble
33	goku20	JPCen-ja	2020/09/21 12:14:57	4093	46.08	47.67	45.93	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, single model
34	goku20	JPCen-ja	2020/09/22 00:08:01	4107	46.31	48.01	46.21	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, ensemble of 3 models
35	tpt_wat	JPCen-ja	2021/04/27 02:26:24	5706	49.37	50.64	49.42	-	-	-	-	-	-	-	NMT	No	Base Transformer model with joint vocab, size 8k
36	Bering Lab	JPCen-ja	2021/05/04 20:52:15	6387	48.83	50.07	48.77	-	-	-	-	-	-	-	NMT	Yes	Transformer Ensemble with additional crawled parallel corpus
37	sakura	JPCen-ja	2024/08/08 19:51:21	7276	49.90	51.10	50.00	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh) - Best
38	sakura	JPCen-ja	2024/08/09 00:14:13	7292	49.60	50.70	49.60	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh)

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	JPCen-ja	2016/07/13 16:47:25	973	0.728539	0.728281	0.729077	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT
2	ORGANIZER	JPCen-ja	2016/07/13 16:50:03	974	0.777759	0.778657	0.779049	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Hierarchical Phrase-based SMT
3	ORGANIZER	JPCen-ja	2016/07/13 16:52:07	975	0.797353	0.796783	0.798025	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Tree-to-String SMT
4	Kyoto-U	JPCen-ja	2016/07/13 17:38:24	986	0.808999	0.807972	0.809610	-	-	-	-	0.000000	0.000000	0.000000	EBMT	No	Baseline
5	ORGANIZER	JPCen-ja	2016/07/26 10:20:53	1036	0.798168	0.792471	0.796308	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	Online A (2016)
6	ORGANIZER	JPCen-ja	2016/08/02 10:26:14	1073	0.743083	0.735203	0.740962	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	Online B (2016)
7	NICT-2	JPCen-ja	2016/08/04 17:23:27	1078	0.826228	0.823582	0.824428	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT with Preordering + Domain Adaptation
8	ORGANIZER	JPCen-ja	2016/08/05 14:36:45	1085	0.761224	0.757341	0.760325	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT D (2016)
9	ORGANIZER	JPCen-ja	2016/08/05 14:38:28	1086	0.773673	0.769244	0.773344	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT F (2016)
10	ORGANIZER	JPCen-ja	2016/08/05 14:41:32	1087	0.743484	0.741985	0.742300	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT E (2016)
11	NICT-2	JPCen-ja	2016/08/05 17:56:31	1098	0.836556	0.832401	0.832622	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC) + Google 5-gram LM
12	bjtu_nlp	JPCen-ja	2016/08/08 12:26:18	1112	0.842762	0.840148	0.842669	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	RNN Encoder-Decoder with attention mechanism, single model
13	JAPIO	JPCen-ja	2016/08/15 15:56:40	1141	0.851376	0.848580	0.849513	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus
14	JAPIO	JPCen-ja	2016/08/17 11:36:23	1156	0.859139	0.856392	0.857422	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JPC/JAPIO corpora
15	JAPIO	JPCen-ja	2016/08/17 11:39:13	1157	0.859957	0.857655	0.858750	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus including some sentences in testset
16	JAPIO	JPCen-ja	2016/10/27 13:05:35	1330	0.839141	0.835888	0.838096	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT with Preordering
17	EHR	JPCen-ja	2017/07/19 18:48:50	1406	0.860998	0.858466	0.860659	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	SMT reranked NMT (sub word based, by Moses and OpenNMT)
18	EHR	JPCen-ja	2017/07/19 19:08:34	1407	0.866722	0.864256	0.866205	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Simple NMT (sub word based, by OpenNMT)
19	JAPIO	JPCen-ja	2017/07/25 12:13:33	1445	0.875667	0.873486	0.874423	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus including some sentences in testset
20	JAPIO	JPCen-ja	2017/07/25 18:02:12	1451	0.864568	0.862781	0.864251	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	OpenNMT(dbrnn)
21	JAPIO	JPCen-ja	2017/07/25 18:10:48	1453	0.880215	0.877980	0.879319	-	-	-	-	0.000000	0.000000	0.000000	NMT	Yes	OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor
22	JAPIO	JPCen-ja	2017/07/25 18:14:57	1454	0.886403	0.883481	0.885747	-	-	-	-	0.000000	0.000000	0.000000	NMT	Yes	Combination of 4 NMT systems (OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor)
23	JAPIO	JPCen-ja	2017/07/26 10:16:13	1462	0.864038	0.861596	0.862200	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus
24	u-tkb	JPCen-ja	2017/07/26 12:46:57	1470	0.845815	0.846888	0.845551	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	NMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
25	ORGANIZER	JPCen-ja	2018/08/15 18:35:10	1964	0.860702	0.857422	0.859818	-	-	-	-	-	0.000000	0.000000	NMT	No	NMT with Attention
26	EHR	JPCen-ja	2018/09/08 20:31:56	2248	0.871439	0.868538	0.871104	-	-	-	-	-	0.000000	0.000000	NMT	No	SMT reranked NMT
27	EHR	JPCen-ja	2018/09/13 12:50:53	2284	0.870730	0.867947	0.870616	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 13)
28	EHR	JPCen-ja	2018/09/15 15:50:14	2395	0.873426	0.870306	0.873007	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 18)
29	EHR	JPCen-ja	2018/09/16 15:09:19	2476	0.872828	0.870332	0.872442	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 20)
30	sarah	JPCen-ja	2019/07/26 11:26:05	2973	0.871185	0.870264	0.871307	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
31	KNU_Hyundai	JPCen-ja	2019/07/27 12:06:18	3185	0.878567	0.876831	0.878260	-	-	-	-	-	-	-	NMT	No	Transformer Base, relative position, BT, r2l reranking, checkpoint ensemble
32	KNU_Hyundai	JPCen-ja	2019/07/27 12:51:20	3193	0.881651	0.879163	0.881114	-	-	-	-	-	-	-	NMT	Yes	Transformer Base (+ASPEC data), relative position, BT, r2l reranking, checkpoint ensemble
33	goku20	JPCen-ja	2020/09/21 12:14:57	4093	0.868699	0.866600	0.868532	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, single model
34	goku20	JPCen-ja	2020/09/22 00:08:01	4107	0.870328	0.868370	0.870081	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, ensemble of 3 models
35	tpt_wat	JPCen-ja	2021/04/27 02:26:24	5706	0.879917	0.877040	0.878975	-	-	-	-	-	-	-	NMT	No	Base Transformer model with joint vocab, size 8k
36	Bering Lab	JPCen-ja	2021/05/04 20:52:15	6387	0.880505	0.878333	0.880066	-	-	-	-	-	-	-	NMT	Yes	Transformer Ensemble with additional crawled parallel corpus
37	sakura	JPCen-ja	2024/08/08 19:51:21	7276	0.883106	0.880884	0.883261	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh) - Best
38	sakura	JPCen-ja	2024/08/09 00:14:13	7292	0.881716	0.879446	0.882026	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh)

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	JPCen-ja	2016/07/13 16:47:25	973	0.711900	0.711900	0.711900	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT
2	ORGANIZER	JPCen-ja	2016/07/13 16:50:03	974	0.715300	0.715300	0.715300	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Hierarchical Phrase-based SMT
3	ORGANIZER	JPCen-ja	2016/07/13 16:52:07	975	0.717030	0.717030	0.717030	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Tree-to-String SMT
4	Kyoto-U	JPCen-ja	2016/07/13 17:38:24	986	0.716990	0.716990	0.716990	-	-	-	-	0.000000	0.000000	0.000000	EBMT	No	Baseline
5	ORGANIZER	JPCen-ja	2016/07/26 10:20:53	1036	0.719110	0.719110	0.719110	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	Online A (2016)
6	ORGANIZER	JPCen-ja	2016/08/02 10:26:14	1073	0.659950	0.659950	0.659950	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	Online B (2016)
7	NICT-2	JPCen-ja	2016/08/04 17:23:27	1078	0.725540	0.725540	0.725540	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT with Preordering + Domain Adaptation
8	ORGANIZER	JPCen-ja	2016/08/05 14:36:45	1085	0.647730	0.647730	0.647730	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT D (2016)
9	ORGANIZER	JPCen-ja	2016/08/05 14:38:28	1086	0.675470	0.675470	0.675470	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT F (2016)
10	ORGANIZER	JPCen-ja	2016/08/05 14:41:32	1087	0.646930	0.646930	0.646930	-	-	-	-	0.000000	0.000000	0.000000	Other	Yes	RBMT E (2016)
11	NICT-2	JPCen-ja	2016/08/05 17:56:31	1098	0.738630	0.738630	0.738630	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC) + Google 5-gram LM
12	bjtu_nlp	JPCen-ja	2016/08/08 12:26:18	1112	0.722560	0.722560	0.722560	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	RNN Encoder-Decoder with attention mechanism, single model
13	JAPIO	JPCen-ja	2016/08/15 15:56:40	1141	0.747910	0.747910	0.747910	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus
14	JAPIO	JPCen-ja	2016/08/17 11:36:23	1156	0.762850	0.762850	0.762850	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JPC/JAPIO corpora
15	JAPIO	JPCen-ja	2016/08/17 11:39:13	1157	0.768690	0.768690	0.768690	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus including some sentences in testset
16	JAPIO	JPCen-ja	2016/10/27 13:05:35	1330	0.733020	0.733020	0.733020	-	-	-	-	0.000000	0.000000	0.000000	SMT	No	Phrase-based SMT with Preordering
17	EHR	JPCen-ja	2017/07/19 18:48:50	1406	0.747050	0.747050	0.747050	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	SMT reranked NMT (sub word based, by Moses and OpenNMT)
18	EHR	JPCen-ja	2017/07/19 19:08:34	1407	0.747770	0.747770	0.747770	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	Simple NMT (sub word based, by OpenNMT)
19	JAPIO	JPCen-ja	2017/07/25 12:13:33	1445	0.802260	0.802260	0.802260	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus including some sentences in testset
20	JAPIO	JPCen-ja	2017/07/25 18:02:12	1451	0.746720	0.746720	0.746720	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	OpenNMT(dbrnn)
21	JAPIO	JPCen-ja	2017/07/25 18:10:48	1453	0.767720	0.767720	0.767720	-	-	-	-	0.000000	0.000000	0.000000	NMT	Yes	OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor
22	JAPIO	JPCen-ja	2017/07/25 18:14:57	1454	0.776790	0.776790	0.776790	-	-	-	-	0.000000	0.000000	0.000000	NMT	Yes	Combination of 4 NMT systems (OpenNMT(dbrnn) + JPC/Japio corpora + NMT/rule-based posteditor)
23	JAPIO	JPCen-ja	2017/07/26 10:16:13	1462	0.781150	0.781150	0.781150	-	-	-	-	0.000000	0.000000	0.000000	SMT	Yes	Phrase-based SMT with Preordering + Japio corpus
24	u-tkb	JPCen-ja	2017/07/26 12:46:57	1470	0.734010	0.734010	0.734010	-	-	-	-	0.000000	0.000000	0.000000	NMT	No	NMT with SMT phrase translation (phrase extraction with branching entropy; attention over bidirectional LSTMs; by Harvard NMT)
25	ORGANIZER	JPCen-ja	2018/08/15 18:35:10	1964	0.744270	0.744270	0.744270	-	-	-	-	-	0.000000	0.000000	NMT	No	NMT with Attention
26	EHR	JPCen-ja	2018/09/08 20:31:56	2248	0.757510	0.757510	0.757510	-	-	-	-	-	0.000000	0.000000	NMT	No	SMT reranked NMT
27	EHR	JPCen-ja	2018/09/13 12:50:53	2284	0.759140	0.759140	0.759140	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 13)
28	EHR	JPCen-ja	2018/09/15 15:50:14	2395	0.761200	0.761200	0.761200	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 18)
29	EHR	JPCen-ja	2018/09/16 15:09:19	2476	0.759120	0.759120	0.759120	-	-	-	-	-	0.000000	0.000000	NMT	Yes	SMT reranked NMT (4M traning data from WAT and NTCIR, Epoch 20)
30	sarah	JPCen-ja	2019/07/26 11:26:05	2973	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	Transformer, ensemble of 4 models
31	KNU_Hyundai	JPCen-ja	2019/07/27 12:06:18	3185	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	Transformer Base, relative position, BT, r2l reranking, checkpoint ensemble
32	KNU_Hyundai	JPCen-ja	2019/07/27 12:51:20	3193	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	Yes	Transformer Base (+ASPEC data), relative position, BT, r2l reranking, checkpoint ensemble
33	goku20	JPCen-ja	2020/09/21 12:14:57	4093	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, single model
34	goku20	JPCen-ja	2020/09/22 00:08:01	4107	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	mBART pre-training transformer, ensemble of 3 models
35	tpt_wat	JPCen-ja	2021/04/27 02:26:24	5706	0.885762	0.885762	0.885762	-	-	-	-	-	-	-	NMT	No	Base Transformer model with joint vocab, size 8k
36	Bering Lab	JPCen-ja	2021/05/04 20:52:15	6387	0.885435	0.885435	0.885435	-	-	-	-	-	-	-	NMT	Yes	Transformer Ensemble with additional crawled parallel corpus
37	sakura	JPCen-ja	2024/08/08 19:51:21	7276	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh) - Best
38	sakura	JPCen-ja	2024/08/09 00:14:13	7292	0.000000	0.000000	0.000000	-	-	-	-	-	-	-	NMT	No	LLM: Rakuten/RakutenAI-7B-chat Fine-Tuned with JPC Corpus in six direction (En-Ja, Ja-En, Ko-Ja, Ja-Ko, Zh-Ja, Ja-Zh)

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)

#	Team	Task	Date/Time	DataID	HUMAN	Method	Other Resources	System Description
1	bjtu_nlp	JPCen-ja	2016/08/08 12:26:18	1112	39.500	NMT	No	RNN Encoder-Decoder with attention mechanism, single model
2	NICT-2	JPCen-ja	2016/08/05 17:56:31	1098	37.750	SMT	Yes	Phrase-based SMT with Preordering + Domain Adaptation (JPC and ASPEC) + Google 5-gram LM
3	ORGANIZER	JPCen-ja	2016/07/13 16:52:07	975	30.750	SMT	No	Tree-to-String SMT
4	NICT-2	JPCen-ja	2016/08/04 17:23:27	1078	30.750	SMT	No	Phrase-based SMT with Preordering + Domain Adaptation
5	JAPIO	JPCen-ja	2016/08/17 11:36:23	1156	26.750	SMT	Yes	Phrase-based SMT with Preordering + JPC/JAPIO corpora
6	ORGANIZER	JPCen-ja	2016/07/13 16:50:03	974	21.000	SMT	No	Hierarchical Phrase-based SMT
7	ORGANIZER	JPCen-ja	2016/07/26 10:20:53	1036	20.000	Other	Yes	Online A (2016)
8	JAPIO	JPCen-ja	2016/08/15 15:56:40	1141	17.750	SMT	Yes	Phrase-based SMT with Preordering + JAPIO corpus
9	ORGANIZER	JPCen-ja	2016/08/05 14:38:28	1086	12.750	Other	Yes	RBMT F (2016)

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