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	FBAI	ALT2en-my	2019/07/27 14:38:04	3203	-	-	-	-	-	-	-	-	39.25	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
2	FBAI	ALT2en-my	2019/07/27 08:10:07	3149	-	-	-	-	-	-	-	-	36.77	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
3	NICT-4	ALT2en-my	2019/07/26 11:33:55	2979	-	-	-	-	-	-	-	-	31.33	-	Other	Yes	Same as last year but with cleaner monolingual data
4	NICT	ALT2en-my	2019/07/22 17:15:42	2818	-	-	-	-	-	-	-	-	30.84	-	NMT	Yes	Single model+language model pre-training
5	NICT-4	ALT2en-my	2019/07/23 14:09:56	2855	-	-	-	-	-	-	-	-	30.79	-	NMT	Yes
6	sakura	ALT2en-my	2021/05/03 21:39:24	6031	-	-	-	-	-	-	-	-	29.62	-	NMT	No	Marian NMT with Attention
7	sakura	ALT2en-my	2021/04/24 17:04:17	5464	-	-	-	-	-	-	-	-	29.25	-	NMT	No	Marian NMT with Attention
8	UCSMNLP	ALT2en-my	2019/07/26 17:00:18	3021	-	-	-	-	-	-	-	-	28.20	-	SMT	Yes
9	NICT	ALT2en-my	2019/07/22 17:16:44	2819	-	-	-	-	-	-	-	-	25.88	-	NMT	No	Single model+language model pre-training
10	YCC-MT1	ALT2en-my	2021/05/25 11:19:17	6420	-	-	-	-	-	-	-	-	25.48	-	SMT	No	UCSY+ALT, Hybrid Hiero, -xml-input exclusive result, en-my
11	YCC-MT1	ALT2en-my	2021/05/25 08:52:37	6408	-	-	-	-	-	-	-	-	25.11	-	SMT	No	UCSY+ALT, Hybrid SMT, -xml-input inclusive result
12	ORGANIZER	ALT2en-my	2019/07/22 19:12:36	2827	-	-	-	-	-	-	-	-	22.53	-	NMT	No	NMT with Attention
13	YCC-MT1	ALT2en-my	2021/05/05 21:00:09	6394	-	-	-	-	-	-	-	-	21.02	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input exclusive res
14	YCC-MT1	ALT2en-my	2021/05/05 21:07:48	6395	-	-	-	-	-	-	-	-	21.02	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input inclusive
15	YCC-MT1	ALT2en-my	2021/05/04 10:06:25	6195	-	-	-	-	-	-	-	-	20.88	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid PBSMT with XML markup technique, -xml-input inclusive result
16	UCSYNLP	ALT2en-my	2019/07/23 15:28:07	2858	-	-	-	-	-	-	-	-	20.86	-	NMT	No	NMT with attention
17	ORGANIZER	ALT2en-my	2019/07/24 16:18:02	2900	-	-	-	-	-	-	-	-	20.63	-	Other	Yes	Online A
18	YCC-MT1	ALT2en-my	2021/05/04 11:08:40	6201	-	-	-	-	-	-	-	-	20.13	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid OSM with XML markup technique, -xml-input inclusive result
19	sarah	ALT2en-my	2019/07/27 08:01:01	3146	-	-	-	-	-	-	-	-	19.94	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 0 + 1
20	YCC-MT2	ALT2en-my	2021/06/10 16:31:31	6446	-	-	-	-	-	-	-	-	19.92	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
21	sarah	ALT2en-my	2019/07/27 08:05:39	3147	-	-	-	-	-	-	-	-	19.75	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 1
22	YCC-MT2	ALT2en-my	2021/06/10 16:27:49	6445	-	-	-	-	-	-	-	-	19.75	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
23	sarah	ALT2en-my	2019/07/27 07:28:01	3137	-	-	-	-	-	-	-	-	19.64	-	NMT	No	Single model. Transformer. 5 layers, 2 heads, random 0
24	YCC-MT2	ALT2en-my	2021/06/10 16:36:46	6447	-	-	-	-	-	-	-	-	19.57	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.6 0.4
25	YCC-MT2	ALT2en-my	2021/06/10 15:51:14	6438	-	-	-	-	-	-	-	-	18.71	-	NMT	No	UCSY+ALT, transformer, en-my, word-syl
26	YCC-MT2	ALT2en-my	2021/06/10 15:40:57	6437	-	-	-	-	-	-	-	-	17.24	-	NMT	No	UCSY+ALT, s2s, en-my, word-syl
27	YCC-MT2	ALT2en-my	2021/06/10 17:00:39	6452	-	-	-	-	-	-	-	-	16.46	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.5 0.5
28	YCC-MT2	ALT2en-my	2021/06/10 16:56:45	6451	-	-	-	-	-	-	-	-	16.41	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.4 0.6
29	YCC-MT2	ALT2en-my	2021/06/10 17:07:11	6454	-	-	-	-	-	-	-	-	16.23	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.6 0.4
30	YCC-MT2	ALT2en-my	2021/06/10 16:07:54	6441	-	-	-	-	-	-	-	-	15.38	-	NMT	No	UCSY+ALT, s2s, en-my, word-word
31	YCC-MT2	ALT2en-my	2021/05/04 08:06:57	6175	-	-	-	-	-	-	-	-	14.82	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), English-Myanmar (Syllable Unit)
32	YCC-MT2	ALT2en-my	2021/06/10 16:11:44	6442	-	-	-	-	-	-	-	-	14.66	-	NMT	No	UCSY+ALT, transformer, en-my, word-word
33	YCC-MT2	ALT2en-my	2021/05/04 08:27:40	6178	-	-	-	-	-	-	-	-	14.02	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.4 0.6 result (i.e. best BLEU score), English-tree to Myanmar-Syllable Translation
34	NECTEC	ALT2en-my	2021/05/04 08:54:38	6186	-	-	-	-	-	-	-	-	13.90	-	NMT	No	Shared-multi-transformer Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
35	NECTEC	ALT2en-my	2021/05/04 08:50:27	6182	-	-	-	-	-	-	-	-	12.94	-	NMT	No	Multi-source Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
36	NECTEC	ALT2en-my	2021/05/04 09:06:13	6189	-	-	-	-	-	-	-	-	12.82	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:tree, target:string, used 2 GPUs
37	NECTEC	ALT2en-my	2021/05/04 08:52:28	6183	-	-	-	-	-	-	-	-	12.72	-	NMT	No	Transformer tree2string, used NLTK tree parser for English, used 2 GPUs

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	FBAI	ALT2en-my	2019/07/27 14:38:04	3203	-	-	-	-	-	-	-	-	0.782472	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
2	FBAI	ALT2en-my	2019/07/27 08:10:07	3149	-	-	-	-	-	-	-	-	0.769902	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
3	NICT	ALT2en-my	2019/07/22 17:15:42	2818	-	-	-	-	-	-	-	-	0.760247	-	NMT	Yes	Single model+language model pre-training
4	sakura	ALT2en-my	2021/04/24 17:04:17	5464	-	-	-	-	-	-	-	-	0.751248	-	NMT	No	Marian NMT with Attention
5	NICT-4	ALT2en-my	2019/07/23 14:09:56	2855	-	-	-	-	-	-	-	-	0.750353	-	NMT	Yes
6	sakura	ALT2en-my	2021/05/03 21:39:24	6031	-	-	-	-	-	-	-	-	0.739320	-	NMT	No	Marian NMT with Attention
7	NICT-4	ALT2en-my	2019/07/26 11:33:55	2979	-	-	-	-	-	-	-	-	0.732998	-	Other	Yes	Same as last year but with cleaner monolingual data
8	NICT	ALT2en-my	2019/07/22 17:16:44	2819	-	-	-	-	-	-	-	-	0.727299	-	NMT	No	Single model+language model pre-training
9	sarah	ALT2en-my	2019/07/27 08:05:39	3147	-	-	-	-	-	-	-	-	0.719219	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 1
10	sarah	ALT2en-my	2019/07/27 08:01:01	3146	-	-	-	-	-	-	-	-	0.716243	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 0 + 1
11	sarah	ALT2en-my	2019/07/27 07:28:01	3137	-	-	-	-	-	-	-	-	0.711348	-	NMT	No	Single model. Transformer. 5 layers, 2 heads, random 0
12	YCC-MT2	ALT2en-my	2021/06/10 16:36:46	6447	-	-	-	-	-	-	-	-	0.702318	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.6 0.4
13	YCC-MT2	ALT2en-my	2021/06/10 16:31:31	6446	-	-	-	-	-	-	-	-	0.701542	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
14	UCSYNLP	ALT2en-my	2019/07/23 15:28:07	2858	-	-	-	-	-	-	-	-	0.698507	-	NMT	No	NMT with attention
15	YCC-MT2	ALT2en-my	2021/06/10 16:27:49	6445	-	-	-	-	-	-	-	-	0.698334	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
16	YCC-MT2	ALT2en-my	2021/06/10 16:56:45	6451	-	-	-	-	-	-	-	-	0.687596	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.4 0.6
17	YCC-MT2	ALT2en-my	2021/06/10 17:00:39	6452	-	-	-	-	-	-	-	-	0.685076	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.5 0.5
18	ORGANIZER	ALT2en-my	2019/07/24 16:18:02	2900	-	-	-	-	-	-	-	-	0.680398	-	Other	Yes	Online A
19	YCC-MT2	ALT2en-my	2021/06/10 15:51:14	6438	-	-	-	-	-	-	-	-	0.680358	-	NMT	No	UCSY+ALT, transformer, en-my, word-syl
20	YCC-MT2	ALT2en-my	2021/06/10 17:07:11	6454	-	-	-	-	-	-	-	-	0.679639	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.6 0.4
21	YCC-MT2	ALT2en-my	2021/06/10 15:40:57	6437	-	-	-	-	-	-	-	-	0.675465	-	NMT	No	UCSY+ALT, s2s, en-my, word-syl
22	YCC-MT2	ALT2en-my	2021/06/10 16:11:44	6442	-	-	-	-	-	-	-	-	0.674845	-	NMT	No	UCSY+ALT, transformer, en-my, word-word
23	ORGANIZER	ALT2en-my	2019/07/22 19:12:36	2827	-	-	-	-	-	-	-	-	0.670375	-	NMT	No	NMT with Attention
24	YCC-MT2	ALT2en-my	2021/05/04 08:06:57	6175	-	-	-	-	-	-	-	-	0.659582	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), English-Myanmar (Syllable Unit)
25	YCC-MT2	ALT2en-my	2021/06/10 16:07:54	6441	-	-	-	-	-	-	-	-	0.659550	-	NMT	No	UCSY+ALT, s2s, en-my, word-word
26	YCC-MT2	ALT2en-my	2021/05/04 08:27:40	6178	-	-	-	-	-	-	-	-	0.639593	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.4 0.6 result (i.e. best BLEU score), English-tree to Myanmar-Syllable Translation
27	NECTEC	ALT2en-my	2021/05/04 09:06:13	6189	-	-	-	-	-	-	-	-	0.625476	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:tree, target:string, used 2 GPUs
28	NECTEC	ALT2en-my	2021/05/04 08:52:28	6183	-	-	-	-	-	-	-	-	0.610951	-	NMT	No	Transformer tree2string, used NLTK tree parser for English, used 2 GPUs
29	NECTEC	ALT2en-my	2021/05/04 08:54:38	6186	-	-	-	-	-	-	-	-	0.608810	-	NMT	No	Shared-multi-transformer Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
30	YCC-MT1	ALT2en-my	2021/05/25 11:19:17	6420	-	-	-	-	-	-	-	-	0.607339	-	SMT	No	UCSY+ALT, Hybrid Hiero, -xml-input exclusive result, en-my
31	NECTEC	ALT2en-my	2021/05/04 08:50:27	6182	-	-	-	-	-	-	-	-	0.598012	-	NMT	No	Multi-source Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
32	UCSMNLP	ALT2en-my	2019/07/26 17:00:18	3021	-	-	-	-	-	-	-	-	0.596811	-	SMT	Yes
33	YCC-MT1	ALT2en-my	2021/05/05 21:00:09	6394	-	-	-	-	-	-	-	-	0.588198	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input exclusive res
34	YCC-MT1	ALT2en-my	2021/05/05 21:07:48	6395	-	-	-	-	-	-	-	-	0.588198	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input inclusive
35	YCC-MT1	ALT2en-my	2021/05/25 08:52:37	6408	-	-	-	-	-	-	-	-	0.567187	-	SMT	No	UCSY+ALT, Hybrid SMT, -xml-input inclusive result
36	YCC-MT1	ALT2en-my	2021/05/04 10:06:25	6195	-	-	-	-	-	-	-	-	0.553319	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid PBSMT with XML markup technique, -xml-input inclusive result
37	YCC-MT1	ALT2en-my	2021/05/04 11:08:40	6201	-	-	-	-	-	-	-	-	0.545962	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid OSM with XML markup technique, -xml-input inclusive result

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	FBAI	ALT2en-my	2019/07/27 14:38:04	3203	-	-	-	-	-	-	-	-	0.789410	-	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel
2	FBAI	ALT2en-my	2019/07/27 08:10:07	3149	-	-	-	-	-	-	-	-	0.782200	-	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
3	sakura	ALT2en-my	2021/04/24 17:04:17	5464	-	-	-	-	-	-	-	-	0.757860	-	NMT	No	Marian NMT with Attention
4	NICT	ALT2en-my	2019/07/22 17:15:42	2818	-	-	-	-	-	-	-	-	0.757590	-	NMT	Yes	Single model+language model pre-training
5	NICT-4	ALT2en-my	2019/07/26 11:33:55	2979	-	-	-	-	-	-	-	-	0.756290	-	Other	Yes	Same as last year but with cleaner monolingual data
6	NICT-4	ALT2en-my	2019/07/23 14:09:56	2855	-	-	-	-	-	-	-	-	0.753200	-	NMT	Yes
7	sakura	ALT2en-my	2021/05/03 21:39:24	6031	-	-	-	-	-	-	-	-	0.752340	-	NMT	No	Marian NMT with Attention
8	NICT	ALT2en-my	2019/07/22 17:16:44	2819	-	-	-	-	-	-	-	-	0.746480	-	NMT	No	Single model+language model pre-training
9	sarah	ALT2en-my	2019/07/27 08:01:01	3146	-	-	-	-	-	-	-	-	0.714740	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 0 + 1
10	sarah	ALT2en-my	2019/07/27 07:28:01	3137	-	-	-	-	-	-	-	-	0.713620	-	NMT	No	Single model. Transformer. 5 layers, 2 heads, random 0
11	sarah	ALT2en-my	2019/07/27 08:05:39	3147	-	-	-	-	-	-	-	-	0.712580	-	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 1
12	UCSMNLP	ALT2en-my	2019/07/26 17:00:18	3021	-	-	-	-	-	-	-	-	0.693420	-	SMT	Yes
13	YCC-MT1	ALT2en-my	2021/05/25 08:52:37	6408	-	-	-	-	-	-	-	-	0.689400	-	SMT	No	UCSY+ALT, Hybrid SMT, -xml-input inclusive result
14	YCC-MT2	ALT2en-my	2021/06/10 17:00:39	6452	-	-	-	-	-	-	-	-	0.688820	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.5 0.5
15	YCC-MT2	ALT2en-my	2021/06/10 16:31:31	6446	-	-	-	-	-	-	-	-	0.688340	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
16	YCC-MT2	ALT2en-my	2021/06/10 16:56:45	6451	-	-	-	-	-	-	-	-	0.688240	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.4 0.6
17	YCC-MT2	ALT2en-my	2021/06/10 16:36:46	6447	-	-	-	-	-	-	-	-	0.687470	-	NMT	No	UCSY+ALT, s2s-plus-transformer, en-my, word-syl, weight: 0.6 0.4
18	UCSYNLP	ALT2en-my	2019/07/23 15:28:07	2858	-	-	-	-	-	-	-	-	0.687260	-	NMT	No	NMT with attention
19	YCC-MT1	ALT2en-my	2021/05/25 11:19:17	6420	-	-	-	-	-	-	-	-	0.684110	-	SMT	No	UCSY+ALT, Hybrid Hiero, -xml-input exclusive result, en-my
20	YCC-MT2	ALT2en-my	2021/06/10 17:07:11	6454	-	-	-	-	-	-	-	-	0.681180	-	NMT	No	UCSY+ALt, s2s-plus-transformer, en-my, word-word, weight: 0.6 0.4
21	YCC-MT2	ALT2en-my	2021/06/10 16:27:49	6445	-	-	-	-	-	-	-	-	0.681020	-	NMT	No	UCSY+ALT, 2 model ensembling, s2s-plus-transformer, en-my, word-syl, weight: 0.5 0.5
22	YCC-MT2	ALT2en-my	2021/06/10 16:11:44	6442	-	-	-	-	-	-	-	-	0.679630	-	NMT	No	UCSY+ALT, transformer, en-my, word-word
23	YCC-MT2	ALT2en-my	2021/06/10 15:51:14	6438	-	-	-	-	-	-	-	-	0.678260	-	NMT	No	UCSY+ALT, transformer, en-my, word-syl
24	YCC-MT2	ALT2en-my	2021/06/10 15:40:57	6437	-	-	-	-	-	-	-	-	0.674100	-	NMT	No	UCSY+ALT, s2s, en-my, word-syl
25	YCC-MT2	ALT2en-my	2021/06/10 16:07:54	6441	-	-	-	-	-	-	-	-	0.672950	-	NMT	No	UCSY+ALT, s2s, en-my, word-word
26	YCC-MT2	ALT2en-my	2021/05/04 08:06:57	6175	-	-	-	-	-	-	-	-	0.663840	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), English-Myanmar (Syllable Unit)
27	ORGANIZER	ALT2en-my	2019/07/22 19:12:36	2827	-	-	-	-	-	-	-	-	0.661920	-	NMT	No	NMT with Attention
28	YCC-MT1	ALT2en-my	2021/05/04 10:06:25	6195	-	-	-	-	-	-	-	-	0.655310	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid PBSMT with XML markup technique, -xml-input inclusive result
29	YCC-MT1	ALT2en-my	2021/05/04 11:08:40	6201	-	-	-	-	-	-	-	-	0.654820	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid OSM with XML markup technique, -xml-input inclusive result
30	NECTEC	ALT2en-my	2021/05/04 08:50:27	6182	-	-	-	-	-	-	-	-	0.654780	-	NMT	No	Multi-source Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
31	YCC-MT1	ALT2en-my	2021/05/05 21:00:09	6394	-	-	-	-	-	-	-	-	0.653840	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input exclusive res
32	YCC-MT1	ALT2en-my	2021/05/05 21:07:48	6395	-	-	-	-	-	-	-	-	0.653840	-	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid Hiero with XML markup technique, -xml-input inclusive
33	NECTEC	ALT2en-my	2021/05/04 08:52:28	6183	-	-	-	-	-	-	-	-	0.645760	-	NMT	No	Transformer tree2string, used NLTK tree parser for English, used 2 GPUs
34	YCC-MT2	ALT2en-my	2021/05/04 08:27:40	6178	-	-	-	-	-	-	-	-	0.645470	-	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.4 0.6 result (i.e. best BLEU score), English-tree to Myanmar-Syllable Translation
35	NECTEC	ALT2en-my	2021/05/04 08:54:38	6186	-	-	-	-	-	-	-	-	0.645260	-	NMT	No	Shared-multi-transformer Archi: Transformer, source1:string, source2:tree, target:string, used 2 GPUs
36	NECTEC	ALT2en-my	2021/05/04 09:06:13	6189	-	-	-	-	-	-	-	-	0.638870	-	NMT	No	Multi-source Archi: s2s, source1:string, source2:tree, target:string, used 2 GPUs
37	ORGANIZER	ALT2en-my	2019/07/24 16:18:02	2900	-	-	-	-	-	-	-	-	0.622430	-	Other	Yes	Online A

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)

#	Team	Task	Date/Time	DataID	HUMAN	Method	Other Resources	System Description
1	sakura	ALT2en-my	2021/05/03 21:39:24	6031	Underway	NMT	No	Marian NMT with Attention
2	YCC-MT2	ALT2en-my	2021/05/04 08:06:57	6175	Underway	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.5 0.5 result (i.e. best BLEU score), English-Myanmar (Syllable Unit)
3	YCC-MT2	ALT2en-my	2021/05/04 08:27:40	6178	Underway	NMT	No	2 Model Ensemble (RNN Attention + Transformer), --weights 0.4 0.6 result (i.e. best BLEU score), English-tree to Myanmar-Syllable Translation
4	YCC-MT1	ALT2en-my	2021/05/04 10:06:25	6195	Underway	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid PBSMT with XML markup technique, -xml-input inclusive result
5	YCC-MT1	ALT2en-my	2021/05/04 11:08:40	6201	Underway	SMT	No	UCSY only, Manually extracted transliterated Burmese word (including combined-words) and English word pairs from the whole ALT Myanmar corpus, run Hybrid OSM with XML markup technique, -xml-input inclusive result

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)

#	Team	Task	Date/Time	DataID	HUMAN	Method	Other Resources	System Description
1	NICT	ALT2en-my	2019/07/22 17:15:42	2818	Underway	NMT	Yes	Single model+language model pre-training
2	UCSYNLP	ALT2en-my	2019/07/23 15:28:07	2858	Underway	NMT	No	NMT with attention
3	NICT-4	ALT2en-my	2019/07/26 11:33:55	2979	Underway	Other	Yes	Same as last year but with cleaner monolingual data
4	UCSMNLP	ALT2en-my	2019/07/26 17:00:18	3021	Underway	SMT	Yes
5	sarah	ALT2en-my	2019/07/27 07:28:01	3137	Underway	NMT	No	Single model. Transformer. 5 layers, 2 heads, random 0
6	sarah	ALT2en-my	2019/07/27 08:01:01	3146	Underway	NMT	No	Ensemble. Transformer. 5 layers, 2 heads, random 0 + 0 + 1
7	FBAI	ALT2en-my	2019/07/27 08:10:07	3149	Underway	NMT	No	ensemble 5 models, (BT iter1 + Self-Training)
8	FBAI	ALT2en-my	2019/07/27 14:38:04	3203	Underway	NMT	Yes	5 model ensemble, (BT iter1 + Self-Training) + (BT+ST) iter2 + fine tuning + noisy channel

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